Reproducing ten years of road ageing –
accelerated carbonation and leaching of EAF steel slag
Pascal Suer*1, Jan-Erik Lindqvist2, Maria Arm1, Paul Frogner-Kockum1.
1 Swedish Geotechnical Institute, Linköping, Sweden.
2 Swedish Cement and Concrete Research Institute, Borås, Sweden.
*Corresponding author: pascal.suer@swedgeo.se; tel: +4613201889, fax +4613201912
Preprint of the final paper published in Science of the Total Environment, 2009, vol 407, no 18, pp 5110-5118.
http://dx.doi.org/10.1016/j.scitotenv.2009.05.039
Abstract
Reuse of industrial aggregates is still hindered by concern for their long-term properties. This paper proposes a laboratory method for accelerated ageing of steel slag, to predict environmental and technical properties, starting from fresh slag. Ageing processes in a 10-year old asphalt road with steel slag of electric arc furnace (EAF) type in the subbase were identified by scanning electron microscopy (SEM) and leaching tests. Samples from the road centre and the pavement edge were compared with each other and with samples of fresh slag. It was found that slag from the pavement edge showed traces of carbonation and leaching processes, whereas the road centre material was nearly identical to fresh slag, in spite of an accessible particle structure. Batches of moisturized road centre material exposed to oxygen, nitrogen or carbon dioxide (CO2) were used for accelerated ageing. Time (7–14 days), temperature (20–40 °C) and initial slag moisture content (8–20 %) were varied to achieve the carbonation (decrease in pH) and leaching that was observed in the pavement edge material. After ageing, water was added to assess leaching of metals and macroelements. 12 % moisture, CO2 and seven days at 40 °C gave the lowest pH value. This also reproduced the observed ageing effect for Ca, Cu, Ba, Fe, Mn, Pb, Ca (decreased leaching) and for V, Si, and Al (increased leaching). However, ageing effects on SO4, DOC and Cr were not reproduced.
Keywords: EAF steel slag, Highways and roads, Ageing, Carbonation, Laboratory tests, Leaching, Heavy metals
1 Introduction
According to the European Commission “Towards a thematic strategy on the urban environment”
(COM, 2004(60)) buildings and the built environment use 50 weight percent of materials taken from the Earth’s crust. In Sweden, about 70.2 million tons of sand, gravel, till and crushed rock were extracted in 2003, over half of which consisted of crushed rock. About half of the aggregates were consumed in road construction projects (SGU, 2005).
In order to obtain a sustainable use of resources, industrial by-products such as steel slag, crushed
that influence the material under its lifetime take longer than a few months. Knowledge of properties of similar but aged materials helps to assess the long-term technical and environmental properties. In addition, accelerating the ageing processes observed in constructions may give insight into the future behaviour of materials that at present are in storage.
This paper aims to (i) Identify crucial processes of ageing related to the usefulness of steel slag in roads; and (ii) Propose a method for accelerated ageing to predict the long-term properties.
Accelerated ageing methods have been researched for incinerator ash, usually with a purpose of either predicting the leaching of metals from the ash after short term storage, or treating the ash for better leaching characteristics. Carbonation, moisture exposure, repeated wetting and drying cycles, aeration and elevated temperatures reproduce short term storage ageing or improved leaching characteristics to a variable extent (Carter et al., 2003; Mostbauer et al., 2003; Polettini and Pomi, 2004; Van Gerven et al., 2005). Cappai et al. used accelerated carbonation as a possible treatment method for Waelz slag by application of a moistened CO2 stream (Cappai et al., 2006).
A similar procedure may be applicable to the long-term change in properties of electric arc furnace (EAF) steel slag and other by-products. A 10-year old road with an EAF-slag subbase layer was used to identify the key processes responsible for ageing, and apply these processes to accelerate the ageing of unaffected slag. The ageing method was verified by comparison of leaching characteristics of fresh, road aged and accelerated aged materials.
2 Materials and methods
2.1 Materials and sampling
The steel slag used in this study was electric arc furnace slag from steel production based on scrap metal. Slag samples were taken from the factory storage (one day old), from a small pile (seven months old) and from a road subbase (ten years old). All slag was produced in the same factory.
One day old slag with particle size 0–40 mm was sampled at the factory storage. After removing the surface layer of a storage pile, 0.5 m3 slag was top filled into buckets that were sealed. One subsample named “produced 2006” was subjected to leaching test within one week. The remainder was stored outdoors in a small (½ m high) pile and sampled after seven months. This material was named “7- months pile”.
Slag (0-300 mm) from the same factory had been used in the subbase of an asphalt paved road in 1996. From top to bottom this road consisted of a 100 mm thick asphalt bound wearing course and base, about 50 mm thick unbound base of crushed rock and about 450 mm thick unbound subbase of slag. Samples from this road were taken on 27–28 Sep 2006 from 1 m wide trenches dug down to the geotextile that separated the slag from the subgrade. One trench was dug perpendicular to the road for measurement of a pH profile. Two other trenches were dug parallel to the road – at the road centre and at the edge of the pavement – to provide larger samples for further characterization in the laboratory.
Trench depths were 0.7–1 m.
pH was measured in the field on samples taken in a grid on the perpendicular trench wall. The following sampling depths were chosen: bottom of the base layer, 5 cm from top of the slag layer, centre of the slag layer and finally bottom of the slag layer about 7 cm above the geotextile. In addition, two samples were taken in the subgrade. Each sample was sieved to < 2 mm, 5 ml material was placed in a flask and 25 ml deionizer water was added. The flask was shaken for five minutes and allowed to equilibrate for at least two hours before pH was measured with a pH-electrode (ISO 10390). pH was not measured on an unsieved sample since the particle size was up to 63 mm (
). Sieving to < 2 mm and using 5 ml material corresponded to the standard test for measurement of pH
Fig 1: Particle size distribution of “fresh” steel slag and steel slag from the road. (Dotted lines are the Swedish limit curves for subbase materials in paved roads.)
grov mellan
fin grov
mellan
fin 0,2 Sand 0,6 2 6 Grus 20 60
0,06
1 0,5 0,25 0,125
0,063 2 4 5,6 8 11,2 16 31,5 45 63 90 200
0,075 0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Kornstorle k, m m
Passerande mängd
Stålslagg Färsk Stålslagg Uppgävt ur väg.
Förstärkningslager, VÄG94
Slag from the two parallel trenches was sieved to < 40 mm and for each trench four 10-L buckets were top filled with slag and sealed with lock and tape. These samples were designated “road centre” and
“pavement edge”. Samples for SEM analyses were also taken. These were either stored in a vacuum exicator for later analysis of particle surfaces or cast in epoxy on site for later analysis of cross sections through the slag particles.
Slag from the same factory has been studied previously. In 1992, stored slag was characterized in the laboratory and used in an unpaved test road and a 15 m3 lysimeter (Flyhammar and Bendz, 2004;
Fällman, 1997b; Fällman and Hartlén, 1994; Fällman and Hartlén, 1996). The characterization in 1992 included a percolation test and data from this test is named “produced 1992” in this paper.
2.2 Total content
Samples were crushed to particle size <125 µm. For most elements the samples were dried at 105 °C, melted with LiBO2 followed by dissolution in diluted nitric acid (ASTM D3682). For analysis of As, Cd, Cu, Co, Hg, Ni, Pb, Sb, Se and S samples were dried at 50 °C and digested with nitric acid in microwave bomb (modified ASTM D 3683) for the road samples, produced 2006 and 7-months pile, and by nitric acid in PTFE bomb for As, Cd, Co, Cu, Hg, Ni, Pb, Zn in samples produced 1992 (SS028183). Measured contents were adjusted to dry matter at 105 °C. Analysis was performed by means of ICP-AES or ICP-MS.
2.3 SEM analyses
Scanning Electron Microscopy (SEM) was used to reveal the microstructure of samples from the road
7 months sample Road centre
electron micrographs (BSE) were taken of representative areas at magnifications of 50 times and 1500 times. EDXA with point calculation was also performed.
2.4 Leaching tests
Leaching properties of the sampled slag materials – road centre, pavement edge, produced 2006 and 7-months pile – were assessed. The method used was an up-flow percolation procedure according to CEN/TS 14405 but with a larger column to accommodate the larger particle size. The following procedure was used: First, material with particle size 0–40 mm was packed into 200 mm diameter columns. Particle size 0–40 mm corresponded to 90 weight percent of the material. The particle size was chosen to retain the standardised particle size/column diameter ratio for percolation test CEN/TS 14405. Second, the columns were saturated with deionized water and allowed to equilibrate for five days, after which pumping was resumed. Then, leachate samples were taken at liquid/solid ratios of 0.1, 0.3, 0.5, 1 and 2 L/kg (cumulative). Finally, each water sample was filtered at 0.45 µm and analyzed for pH, Eh, electrical conductivity, metals, macrocations, Cl, F, SO4 and DOC (dissolved organic carbon). Analyses were done with ICP-AES (EPA 200.7), ICP-SFMS (EPA 200.8), AFS (EN 13506), ISO 10359-1, ISO 10304-1&2 and EN 1484. All laboratory work was done by accredited laboratories.
This procedure differed from NT ENVIR 002 which Fällman and Hartlén used on material produced 1992. NT ENVIR 002 is performed on material crushed to particle size < 20 mm, in a 200 mm column and using a nitric acid solution at pH 4 (Fällman, 1997a).
Saturation indices (SI) for the leachates were calculated with the LeachXS expert system version 1.0.4.1 including the MINTEQA2 database (USEPA, 1999) and DOC interaction according to the NICA-Donnan model (ECN et al., 2008; van Zomeren et al., 2007). Minerals were accepted as near saturation when SI was between -0.15 and +0.15, and the choice between minerals within this group was based on agreement with measured trends of saturated concentrations vs liquid:solid ratio.
Elements below the detection limit were included at half detection limit value.
2.5 Accelerated ageing
The ageing experiments used the road centre sample as starting material, since this was most similar to materials from large-scale storage (see section 3). The identified primary ageing process was oxidation or carbonation with moisture contact less than 0.1 L/kg (see section 3). The parameters tested in the ageing experiments were gas type, moisture content, temperature and time (Table 1).
The following procedure was used:
The sample was sieved to particle size < 4 mm and divided into subsamples. < 4 mm is the particle size used for leaching tests for by-products and wastes (see for example EN 12457, CEN/TS 14405 and CEN/TS 14997). The particle size was larger than in the pH measurement in order to have more materials for the tests. 8 weight percent of the material had a particle size of < 2 mm, 15 weight percent was < 4 mm (
1. ).
2. 16 g (undried) subsample was placed in a 50 ml plastic flask.
3. 0.6, 1.25 or 2.5 g deionized water was dripped onto the sample (named small, medium and large moisture addition respectively). This resulted in 8, 12 and 20 % (wet weight) moisture content respectively (Table 2).
4. Nitrogen, oxygen or carbon dioxide was flowed through a washing flask with deionized water into the flask with the slag sample. The flask was sealed and placed in double plastic bags filled with the same gas. Nitrogen and oxygen gas were of chemical quality, the carbon dioxide of instrument quality, containing max 2 ppm hydrocarbons and 15 ppm oxygen.
5. Flasks in bags were placed in an end-over-end rotator, either inside a 40 ºC heating cabinet or at room temperature, for one or two weeks.
6. The samples were taken into room temperature and 51 g deionized water was added. This corresponded to the pH measurement of the road profile.
7. The next day, pH and Eh were measured and the flasks were refrigerated.
All accelerated ageing tests were duplicated.
Four of the samples, the two with the lowest pH (CO2, 40 ºC, 7 d) and the two corresponding blanks (N2, 40 ºC, 7 d), were further analyzed:
8. They were placed in the end-over-end rotator for 24 hours.
9. The water was filtered (0.45 µm) and analyzed for metals, macro cations, Cl, F, SO4, DOC and DIC (dissolved inorganic carbon).
Table 1: Experimental matrix for accelerated ageing. Shaded samples were also analyzed for metals, macrocations, Cl, F, SO4, DOC and DIC.
Moisture: Small Medium Medium Medium Large
Temp and time 40 ºC, 7d 25 ºC, 7d 40 ºC, 7d 40 ºC, 14d 40 ºC, 7d
CO2 2 2 2 2 2
O2 2 2 2 2 2
N2 2 2 2 2 2
Table 2: Moisture content for the accelerated ageing (% wet weight).
Moisture: Small Medium Medium Medium Large
Temp and time: 40 ºC, 7d 25 ºC, 7d 40 ºC, 7d 40 ºC, 14d 40 ºC, 7d
CO2 7.6 11.9 12.0 11.7 19.8
CO2 7.8 11.7 11.9 11.7 19.8
O2 7.7 11.8 11.9 11.8 19.7
O2 7.8 11.6 11.7 11.7 20.0
N2 7.7 11.7 11.7 11.9 19.6
N2 7.8 11.7 11.7 11.7 19.5
3 Results
3.1 Characteristics of the original slag
The slag placed in the road was produced 1995, but no data for the fresh slag at that time was available. However, samples of fresh slag produced in 1992 was analysed by (Fällman, 1997a), and fresh slag produced 2006 was analysed in the present study. The total content of elements in those slag samples together with the samples from the road and the 7-months pile is shown in Table 3. The slag consisted mainly of Fe, Ca, Si, Mg and Mn. The leaching behaviour in percolation tests of the slag produced 1992 and 2006 were very similar (see supporting information).
Geotechnical properties such as proctor compaction results, resilient modulus and accumulated permanent compressive strain in cyclic load triaxial test etc, both for the road samples and the material produced 2006, were also investigated and data is reported in (Arm et al., 2008; Arm et al., 2009).
However, the differences in properties between fresh material and road aged material were too small to allow for measurement on accelerated aged samples, and therefore the geotechnical data is not discussed here.
Table 3: Total content of elements after digestion, in mg/kg. (Data for 1992 from (Fällman, 1997a)) Road
centre
Pavement edge
Produced 1992
Produced 2006
7-months pile
Al 17600 17400 21850 25700 25900
As < 3 2,92 5,26 7,12 4,29
Ba 721 710 728 648 635
Be < 0,7 < 0,7 < 0,7 <0,6
Ca 190000 189000 221500 197000 192000
Cd 0,22 0,16 0,45 0,16 0,197
Co 4,91 3,73 5,78 8,55 4,95
Cr 8530 8450 7765 12200 12300
Cu 158 158 166 233 195
Fe 341000 341000 242500 306000 396000
Hg < 0,02 < 0,02 < 0,010 <0,01 K < 900 < 900 455 < 900 < 900
La < 6 7,48 14,1 15,7
LOI < 2,3 < 2,5 <0,1 < 3,0 -2,2
Mg 46800 45600 45150 49800 48700
Mn 39000 41100 39300 43000 44300
Mo 30 27,7 20,6 47,1 43,9
Na 838 684 370 < 500 <500
Nb 95,6 107 186 188
Ni 25,9 21 45 73,5 52,6
P 3110 3220 2320 2350
Pb 15 7,86 21,5 8,42 6,36
S 1150 1040 1260 901 1070
Sb 1,14 1,11 1,97 1,09
Sc < 1 < 1 < 1 <1
Se 1,51 0,9 0,57 0,796
Si 54700 52800 57650 51900 53300
Sn < 20 < 20 < 20 <20
Sr 145 145 219 220
Ti 2190 2190 2720 2580
TS 98,9 99,1 97,6 96
V 1260 1310 1210 1710 1590
W 233 195 244 304
Y 8,1 8,19 7,15 4,3
Zn 214 165 244 321 364
Zr 89,9 91,6 134 126
TS (%) 98,9 99,1 97,6 96
LOI1000°C (%) < 2,3 < 2,5 <0,1 < 3,0 -2,2
3.2 Key ageing processes in the road
The SEM analyses indicated an open structure in the road slag particles. The particle surfaces had a high frequency of micro cracks along mineral grain boundaries and, in lower frequency, larger cracks cutting across the mineral structure of the grains, giving easy access for water (Fig 2). Precipitates of gypsum and the iron sulphate mineral melanterite were observed on the particle surfaces (Fig 3 and Fig 4). Crystallization was observed on the surfaces of the slag particles during sample drying when sampling in field, and the gypsum crystals may have been precipitated during sampling. Also calcium silicates were observed on the particle surfaces (Fig 5 and Fig 2: Surface of a steel slag particle from the road centre. Dark lines are cracks along or across mineral grains. SEM image with instrumental
magnification 200 times
Fig 3: Diffractogram from XRD analysis at the surface of a steel slag particle. Gypsum and melanterite have been identified.
Fig 4: Precipitates of gypsum crystals on the surface of a slag particle
Fig 5: Calcium silicates on the particle surfaces. See also Fel! Ogiltig självreferens i bokmärke..
Table 5: Analyses of calcium-silicates observed on the particle surfaces. Results from EDXA analyses in weight percent of the elements. See also the BSE image in Fig 5.
Element 1 2 3 4 5 6
C 3 10 5 1 - -
Na - - - - tr tr
Mg - 1 - - - -
Al 1 1 1 1 2 2
Si 14 8 13 12 16 15
S tr tr tr tr tr tr
Ca 39 26 35 44 41 47
Cr - tr - - tr tr
Mn - tr - tr - -
Fe tr 3 tr 3 2 2
Polished cross sections of slag particles from the pavement edge and the road centre were analyzed using SEM/EDS. The particles from the road centre were dark and showed little variation between individual particles. The particles from the pavement edge were slightly lighter in colour with some colour variation between the particles.
No change from particle edge to the central part was observed in the BSE images (Fig 6). Nor could any chemical reaction fronts be identified in SEM images of particle cross sections (Fig 7 and Table 6). The EDXA analyses also showed little or no difference in chemical composition between particle surfaces and particle centra. A profile from the particle edge obtained from a dark phase sampled in the road centre is given in Table 7 and a profile from a particle edge in light mineral and a dark mineral sampled in the pavement edge are given in Table 8 and Table 9 respectively. The chemical profiles of the light phase showed a scatter in the results that corresponds to the precision of the EDS method. The dark phase showed a variation in composition not related to the distance from the edge of the particle. Chemical changes closer than approximately five microns are difficult to identify using EDS or WDS. The reason for this is that the interaction volume from which the analyzed signal is obtained has a volume of a few cubic microns. The edge of the polished particle may have a slight slope resulting in a change of the geometry increasing the uncertainty of the analysis in the outermost few microns. Within these limitations the analyses show no change in mineral composition close to the particle surface.
Fig 6: BSE images showing no change from particle edge to the central part of the particle.
Fig 7: Cross section through a typical steel slag particle at the pavement edge. SEM image with instrumental magnification 350 times. Chemical composition in positions A, B, C and D is reported in Table 6.
). More calcium sulphates and silicates were observed in slag from the pavement edge, and in part of the slag in the road centre that had been exposed to extended constructional traffic. Possibly the traffic crushed the slag and increased the rate of ageing reactions and the resulting ageing induced calcium silicates on the particle surfaces.
Fig 2: Surface of a steel slag particle from the road centre. Dark lines are cracks along or across mineral grains. SEM image with instrumental magnification 200 times
Fig 3: Diffractogram from XRD analysis at the surface of a steel slag particle. Gypsum and melanterite have been identified.
Fig 4: Precipitates of gypsum crystals on the surface of a slag particle
Fig 5: Calcium silicates on the particle surfaces. See also Fel! Ogiltig självreferens i bokmärke..
Table 4: Analyses of calcium-silicates observed on the particle surfaces. Results from EDXA analyses in weight percent of the elements. See also the BSE image in Fig 5.
Element 1 2 3 4 5 6
C 3 10 5 1 - -
Na - - - - tr tr
Mg - 1 - - - -
Al 1 1 1 1 2 2
Si 14 8 13 12 16 15
S tr tr tr tr tr tr
Ca 39 26 35 44 41 47
Cr - tr - - tr tr
Mn - tr - tr - -
Fe tr 3 tr 3 2 2
Polished cross sections of slag particles from the pavement edge and the road centre were analyzed using SEM/EDS. The particles from the road centre were dark and showed little variation between individual particles. The particles from the pavement edge were slightly lighter in colour with some colour variation between the particles.
No change from particle edge to the central part was observed in the BSE images (Fig 6). Nor could any chemical reaction fronts be identified in SEM images of particle cross sections (Fig 7 and Table 5). The EDXA analyses also showed little or no difference in chemical composition between particle surfaces and particle centra. A profile from the particle edge obtained from a dark phase sampled in the road centre is given in Table 6 and a profile from a particle edge in light mineral and a dark mineral sampled in the pavement edge are given in Table 7 and Table 8 respectively. The chemical profiles of the light phase showed a scatter in the results that corresponds to the precision of the EDS method. The dark phase showed a variation in composition not related to the distance from the edge of the particle. Chemical changes closer than approximately five microns are difficult to identify using EDS or WDS. The reason for this is that the interaction volume from which the analyzed signal is obtained has a volume of a few cubic microns. The edge of the polished particle may have a slight slope resulting in a change of the geometry increasing the uncertainty of the analysis in the outermost few microns. Within these limitations the analyses show no change in mineral composition close to the particle surface.
Fig 6: BSE images showing no change from particle edge to the central part of the particle.
Fig 7: Cross section through a typical steel slag particle at the pavement edge. SEM image with instrumental magnification 350 times. Chemical composition in positions A, B, C and D is reported in Table 5.
Table 5: Chemical composition in the pavement edge slag. A, B, C and D are marked in Fig 7. A is at the particle surface. Results from EDXA analyses in weight percent of the elements.
Element A B C D
Mg 10 11 0.8 0.4
Al 0.5 0.8 8 0.8
Si 0.8 1.1 1.5 15
P – – – 0.8
Ca 1.8 2 31 41
Ti – – 0.5 –
V – – 0.4 0.3
Cr 1.5 2 2 –
Mn 8 8 1.3 0.7
Fe 51 48 24 4
Table 6: Chemical composition of a dark mineral in the road centre slag, from particle surface inward to 40 µm into the particle. Results from EDXA analyses in weight percent of the elements.
Element
Particle surface
40 µm depth
Mg 0.6 0.5 0.6 0.6 0.8 0.5 0.5 1 0.6
Al 0.9 1 0.9 0.8 0.9 0.8 0.7 - 1.1
Si 16 14 14 14 13 13 14 12 14
P 0.9 1 0.9 1 1 1.2 0.9 0.7 0.7
Ca 38 38 38 38 37 35 38 35 38
Ti 0.4 0.5 0.3 0.3 0.4 0.3 0.4 0.6 0.3
Mn 0.9 0.9 0.9 1.1 1.3 0.9 1 1.3 1
Fe 6 7 7 7 9 7 7 13 8
Table 7: Chemical composition of a light mineral in the pavement edge slag, from particle surface inward to 220 µm into the particle. Results from EDXA analyses in weight percent of the elements
Particle surface
220 µm depth
Mg 10 4 9 10 6 8 9 8 7 9 3 14
Al 1 1 1 2 2 2 1 3 2 2 3 1
Si 1 2 1 1 1 1 1 2 1 1 2 1
Ca 4 6 3 3 4 4 3 6 4 3 4 3
Cr 2 2 2 1 2 1 1 2 2 2 2 1
Mn 8 8 8 8 9 9 9 7 9 9 6 8
Fe 47 50 49 46 50 49 47 43 48 47 54 43
Table 8: Chemical composition of a dark mineral in the pavement edge slag, from particle surface inward to 220 µm into the particle. Results from EDXA analyses in weight percent of the elements.
The concentrations of Na and K in the leachates decreased with L/S (Fig 8), indicating that their leaching was restricted by the availability in the solid phase, as expected. The pavement edge slag had the lowest concentrations of soluble salts (Fig 8), which was not surprising. It has been shown that precipitation on and road run-off from an impermeable paved road lead to high water flow through the embankment at the pavement edge (Apul et al., 2007; Flyhammar and Bendz, 2006). Thus, more soluble salts would have been removed from the pavement edge. The difference in remaining soluble salts between road centre and pavement edge corresponded to a liquid:solid ratio less than 0.1 L/kg.
Leaching of alkalinity is the major cause of loss of buffering capacity and pH for bottom ash (Astrup et al., 2006; Yan et al., 1998a; Yan et al., 1998b). Though leaching was observable in the present study, the inferred water amount was low and is unlikely to have caused a significant pH decrease by leaching of alkalinity.
Fig 8: Detail of Na concentrations during leaching in percolation tests.
0 5 10 15 20
0 0,2 0,4 0,6 L/S (L/kg)0,8 1
Na (mg/L)
Road centre Pavement edge Produced 1992 Produced 2006 7 months pile
The pH in the road profile ranged from 10.5 to 12.4 (Table 9). The base material of crushed gneiss had pH 10.5–11.5, and the till in the subgrade had pH 11.7. Alkaline pH is not unusual in crushed rock (Silva et al., 2005), but pH above 10 is unusual in Swedish till and may be due to leaching from the slag layer. The slag layer had pH values between 11.6 and 12.4 with a major decrease close to the pavement edge and a minor decrease in the upper 20 cm under the base. The remainder of the slag had pH 12.2 (Table 9).
Table 9: pH distribution in the road.
Distance from road centre (m)
Depth (m) 0.5 0.85 1.2 1.7 2.15 2.7 3.2 3.5
0.12a 11.3 11.4 11.4 11.5 11.5 11.4 10.7 10.5
0.20 12.0 12.0 12.0 11.9 12.0 12.1 11.8 11.6
0.35 12.3 12.3 12.2 12.2 12.2 12.3 12.1 11.8
0.52 12.3 12.4 12.3 12.3 12.4 12.3 12.0 11.9
0.62b 11.7 11.7
a road base
b subgrade
The pH in the fresh slag – both produced 1992 and produced 2006 – was 12.2, measured in the first
value (Fig 9). The pH decrease at the pavement edge may be due to oxidation of sulfidic compounds in the slag or carbonation of oxides and hydroxides. Carbon dioxide could be available either from the atmosphere or from microbial degradation of organic compounds.
Huijgen et al observed that carbonation of fresh basic oxygen furnace steel slag involved Ca(OH)2 and Ca from C-S-H, which diffused through the particles and out into the water phase, to form CaCO3 on the outer rim (Huijgen et al., 2005). Carbonation mainly involves Ca from calcium silicates, which forms calcium carbonate, leaving the silicates with a lower Ca content (Huijgen and Comans, 2006;
Huijgen et al., 2005; Taylor, 1997). A similar process may be in progress in this study’s electric arc furnace slag. However, Huijgen et al also observed CaCO3 in a rim on the particle surfaces (Huijgen et al., 2005). No such rim was present in the road slag samples.
Fig 9: pH in the percolation tests.
10 11 12 13
0 0,2 0,4 0,6 L/S (L/kg)0,8 1
pH
Road centre Pavement edge Produced 1992 Produced 2006 7 months pile
3.3 Accelerated ageing
The analyses of the road samples lead to the hypothesis that ageing in the road had been controlled by oxidation or carbonation, and not by significant leaching. Thus, moisturised slag was exposed to O2 to test oxidation, CO2 to test carbonation, and N2 as a blank, at different temperatures and moisture levels.
3.3.1 pH
Exposure of the road centre sample to N2 or O2 gave no change in pH (Table 10), while carbon dioxide lowered pH in all samples. Carbonation as the major ageing influence for EAF-slag, more important than moisture or heat, has also been confirmed by (Diener et al., 2008). The lowest pH, 11.3, was achieved in the samples exposed to CO2 and medium moisture in 40 ºC for 7 days. These samples were chosen for subsequent leaching tests, on the hypothesis that pH was the most important parameter to describe leaching of contaminants.
Table 10: pH in the accelerated aged slag.
Moisture: Small Medium Medium Medium Large
While moisture is essential for the carbonation reaction, high moisture content hinders the diffusion of CO2. The lowest pH was achieved with 12 % (wet weight) initial slag moisture, while 8 % and 20 % lead to higher pH under otherwise identical conditions (Table 10: Table 10). Similar results have been found for other materials: 15 % initial moisture resulted in higher carbonation than 5 % when
accelerating carbonation of cement kiln dust (Gardner et al., 2006), and 13–25 % moisture gave the best leaching results for carbonation of municipal solid waste incinerator bottom ash (Van Gerven et al., 2005).
Carbonation at room temperature led to higher pH than at 40 °C (both exposed for 7 days), although the rate of carbonation of lime and hydrated lime decreases at temperatures over 20 ºC due to the lower solubility of calcium hydroxide and CO2 (Shih et al., 1999). However, the sample aged for 14 days at 40 °C had also a higher pH. This can be due to slower base buffering reactions after the
“initial” pH decrease. Consequently, it cannot be excluded that carbonation at room temperature for e.g. ½ week would give results similar to 40 ºC for 7 days.
The moisture content varied up to 0.3 % (wet weight) between duplicates, due to variation in the original sample weighing of up to 0.8 g slag (wet weight). The pH results varied at most 0.2 units (Table 10).
3.3.2 Leaching
The duplicate samples with the lowest pH (CO2, medium water, 7 days at 40 ºC) and corresponding N2
blanks were leached in a batch test. The CO2 aged samples showed higher V, Si, S, Al and Se leaching and lower Ba, Fe, Mn, Pb, Ca and Cu leaching than the N2 samples (Table 11 right column and Fig 10, other elements shown in the supporting information). Comparison of pavement edge (most aged) and road centre (largely unaffected) showed the same effects of ageing as above, except for S and Se.
Thus, the laboratory ageing with carbon dioxide reproduced the road ageing with respect to macro elements, but S and Se concentrations in the leachate increased by CO2 treatment while they were lower for the pavement edge than the road centre. No depletion of S or Se could be observed for the pavement edge sample (Table 3 and supplementary material), which could have explained lower leaching.
Na and K leaching were not affected by the accelerated carbonation. This conclusion was drawn since the N2 and the CO2 treated slag leached similar amounts. Differences between accelerated aged slag samples and the original samples may easily be explained with difference in liquid/solid ratio (Fig 8 and Fig 10).
Fig 10: Concentrations of selected macroelements in the leachate. Accelerated aged samples are duplicates at L/S 3.4 L/kg, Road centre and pavement edge values are at 0.1 and 2 L/kg.
0 300 600 900
Ca
(mg/L)
0 5 10 15
Si DOC S DIC Na Ba
(mg/L)
CO2 CO2 N2 N2
Edge 0.1 Edge 2 Centre 0.1 Centre 2
0 0.5 1 1.5 2
Al
(mg/L)
Fig 11: Concentrations of selected trace elements in the leachate. Accelerated aged samples are duplicates at L/S 3.4 L/kg, Road centre and pavement edge values are at 0.1 and 2 L/kg.
0 30 60
Cr Mo Cu Pb
(µg/L)
CO2 CO2 N2 N2
Edge 0.1 Edge 2 Centre 0.1 Centre 2
0 100 200 300
V
(µg/L)
Table 11: Changes in element leaching in road aged and accelerated aged steel slag.
Pavement edge compared to road centre
CO2 compared to N2
pH -- --
Al ++ +
Ba -- ---
Ca -- --
Cl 0
Cr -- 0
Cu - --
DOC -- 0
Fe ---
K 0 -
Mn ---
Mo - 0
Na 0 -
Pb ---
S - ++
Sb 0
Se - +
Si +++ ++
V +++ ++
Other elements not evaluated due to detection limit CO2/N2 or Pavement
edge/road centre +++ CO2/N2 > 10
++ CO/N 2-10
Vanadium concentrations in the leachate were increased by both road ageing and accelerated
carbonation (Fig 11 and Table 11), at the same time as pH decreased. This is contrary to the results of Apul et al, Fällman et al and Huijgen and Comans, who found that V leaching of aged or carbonated slag decreased with decreasing pH between pH 13 and 9 (Apul et al., 2005; Fällman et al., 1999;
Huijgen and Comans, 2006). Apul et al and Fällman et al concluded that V leaching could be
controlled by sorption processes in this pH range. Our results are more reminiscent of the behaviour of fresh, non carbonated slag: Huijgen and Comans found that fresh blast oxygen furnace slag leached less V than carbonated slag, and that leaching of V from fresh slag increased with decreasing pH (Huijgen and Comans, 2006). The leachates from the present experiments were undersaturated for all V containing minerals in the MINTEQA2 database.
Chromium, another metal with high total contents in the slag, showed decreased leaching with road ageing and outdoors storage but was largely unaffected by carbonation (Fig 11 and Table 11).
Ba(SCr)O4 minerals were the only relevant minerals for Cr leaching in the database, with solubility indices between -0.15 and +0.08 for the non-aged slag samples (produced 2006 and 1992, road centre, N2-aged). Ba(SCr)O4 solubility control is in agreement with results for fresh slag (Fällman, 1997b), where Ba(SCr)O4 was formed within 6 h, and with results from aged slag (Apul et al., 2005). Thus Ba may be expected to influence the Cr leaching. In the aged slag samples (pavement edge, CO2-aged, 7- months pile), no minerals were identified as near saturation, neither for barium nor for chromium.
Fällman et al found Cr in crystalline iron phases during sequential extraction of the slag weathered for five years in an open air lysimeter (Fällman et al., 1999). Transformation of iron and inclusion of Cr in iron phases might explain the decreased importance of Ba(SCr)O4 for concentrations of Cr in the leachate from the pavement edge.
The laboratory aged samples (both N2 and CO2) had higher Cr concentrations in the leachates than the samples of fresh and road aged material (Fig 10). This is most likely a result of the test method, since laboratory aged samples were leached with a batch test and the other samples with a percolation test.
4 Conclusions
The electric arc furnace slag sampled from the centre of the asphalt paved road was nearly identical to fresh material. No reaction fronts were observed in the particles from the road centre, and pH and leaching of controlling minerals were very similar to that of fresh material produced in 1992 and 2006.
The major ageing effects were noticed under the pavement edge. Observed ageing consisted of a pH decrease from 12.2 to 11.6 and a small leaching effect. pH decrease was reproduced by carbonation and not by oxidation. The lowest pH values were achieved with exposure to carbon dioxide for seven days at 40 °C. The simple batch carbonation reproduced the effect of road ageing for leaching of macro elements except S. Most trace elements, for example Cu, V, and Pb, but not Cr, were also reproduced by the laboratory carbonation.
5 Acknowledgements
Several colleagues at SGI and VTI have taken part in this research. The authors want to express their appreciation for the cooperation of Cecilia Toomväli, Lennart Larsson, Martin Lyth, Håkan Arvidsson, Gunilla Franzén, Karl-Johan Lorentz and Jimmy Carlsson. The work was funded by the Swedish Road Administration. Additional funding was provided by the Swedish Research Council Formas, the Thermal Engineering Research Institute, the Swedish Association of Graduate Engineers and the Swedish Geotechnical Institute (SGI).
6 References
Apul DS, Gardner KH, Eighmy TT. Modeling hydrology and reactive transport in roads: The effect of cracks, the edge, and contaminant properties. Waste Management 2007; 27: 1465-1475.
Apul DS, Gardner KH, Eighmy TT, Fallman AM, Comans RNJ. Simultaneous application of dissolution/precipitation and surface complexation/surface precipitation modeling to contaminant leaching. Environmental Science & Technology 2005; 39: 5736-5741.
Arm M, Suer P, Arvidsson H, Lindqvist J-E, Frogner Kockum PC, Larsson L, et al. Förutsägelse av långtidsegenskaper hos restprodukter - teknik och miljö i vägar (Prediction of long-term properties of by-products - Technical and environmental properties in roads). 1083. Värmeforsk, Stockholm, 2008.
Arm M, Suer P, Arvidsson H, Lindqvist J-E, Frogner Kockum PC, Larsson L, et al. Förutsägelse av långtidsegenskaper hos restprodukter - teknik och miljö i vägar. VTI rapport. 641. VTI, 2009.
Astrup T, Jakobsen R, Christensen TH, Hansen JB, Hjelmar O. Assessment of long-term pH
developments in leachate from waste incineration residues. Waste Management & Research 2006; 24:
491-502.
Cappai G, Dettori C, Muntoni A, Mostbauer P. Artificial weathering of Waelz slags. In: Fentiman CH, Hills CD, Mangabhai RJ, editors. First International Conference on Accelerated Carbonation of Environmental and Materials Engineering, London, 2006.
Carter R, Gardner K, Hjelmar O, Lopez S, Kosson D. Mineralogical and leaching characteristics of weathered industrial ash. In: Ortiz de Urbina G, Goumans H, editors. WASCON 2003: fifth international conference on the environmental and technical implications of construction with alternative materials, San Sebastián, Spain, 2003, pp. 691-695.
COM. Towards a thematic strategy on the urban environment. Communication of the European Commission 2004(60); 11 Feb 2004.
Diener S, Andreas L, Herrmann I, Ecke H, Lagerkvist A. Accelerated carbonation of ashes and steel slags in a landfill cover construction. ACEME08, 2nd International Conference on Accelerated Carbonation for Environmental and Materials Engineering, Rome, Italy, 2008.
ECN, Vanderbilt University, DHI. Leaching eXpert System - A new environmental impact assessment tool. 2008. http://www.leaching.org/LeachXSFlyer.pdf; accessed 2008-07-03 2008.
Flyhammar P, Bendz D. Vatten- och masstransporter i en dränerad vägkropp med asfaltsbeläggning.
Lund University, 2004.
Flyhammar P, Bendz D. Leaching of different elements from subbase layers of alternative aggregates in pavement constructions. Journal of Hazardous Materials 2006; 137: 603-611.
Fällman A-M. Characterisation of residues - Release of contaminants from slags and ashes. Dept of Physics and Measurement Technology. Linköping University, Linköping, Sweden, 1997a.
Fällman A-M. Leaching of chromium from steel slag in laboratory and field tests - a solubility controlled process? In: Goumans JJJM, Senden GJ, Sloot HAvd, editors. WASCON ´97 Waste materials in construction - putting theory into practice. Elsevier, 1997b, pp. 531-539.
Fällman A-M, Eighmy TT, Salaneck WR. Aging reactions in residues. AFR, 1999.
Fällman A-M, Hartlén J. Leaching of slags and ashes - controlling factors in field experiments versus in laboratory tests. In: Goumans JJJM, van der Sloot HA, Aalbers TG, editors. Environmental Aspects of Construction with Waste Materials. Elsevier, 1994, pp. 39-53.
Fällman A-M, Hartlén J. Utilisation of electric arc furnace slag in road construction. In: Kamon M, editor. Environmental Geotechnics. Balkema, Rotterdam, 1996, pp. 703-708.
Gardner KH, Fisher J, Melton J, Eighmy TT. Carbon sequestration using recycled materials under ambient atmospheric conditions. In: Fentiman CH, Hills CD, Mangabhai RJ, editors. First
International Conference on Accelerated Carbonation of Environmental and Materials Engineering, London, 2006.
Huijgen WJJ, Comans RNJ. Carbonation of steel slag for CO2 sequestration: Leaching of products and reaction mechanisms. Environmental Science & Technology 2006; 40: 2790-2796.
Polettini A, Pomi R. The leaching behavior of incinerator bottom ash as affected by accelerated ageing. Journal of Hazardous Materials 2004; 113: 211-217.
SGU. Aggregates - production and resources 2004. Swedish Geological Survey, 2005.
Shih SM, Ho CS, Song YS, Lin JP. Kinetics of the reaction of Ca(OH)(2) with CO2 at low temperature. Industrial & Engineering Chemistry Research 1999; 38: 1316-1322.
Silva MTB, Hermo BS, Garcia-Rodeja E, Freire NV. Reutilization of granite powder as an amendment and fertilizer for acid soils. Chemosphere 2005; 61: 993-1002.
Taylor HFW. Cement chemistry. London: Thomas Telford, 1997.
USEPA. MINTEQA2/PRODEFA2, a geochemical assessment model for environmental systems: user manual supplement for version 4.0. Athens, GA, USA: National exposure research laboratory,
ecosystems research division, 1999.
Van Gerven T, Van Keer E, Arickx S, Jaspers M, Wauters G, Vandecasteele C. Carbonation of MSWI-bottom ash to decrease heavy metal leaching, in view of recycling. Waste Management 2005;
25: 291-300.
van Zomeren A, van den Berg P, Bleijerveld R, van der Sloot HA. Identification of in-situ processes controlling emissions of a stabilised waste landfill by field measurements and geochemical modelling.
Sardinia 2007, 11th Int. Waste Managem. and Landfill Symposium. CISA, Cagliari, Italy, 2007.
Yan J, Baverman C, Moreno L, Neretnieks I. Evaluation of the time-dependent neutralising
behaviours of MSWI bottom ash and steel slag. Science of the Total Environment 1998a; 216: 41-54.
Yan JY, Moreno L, Neretnieks I. Neutralizing processes in leaching of solid waste: Modeling of interactions between solid waste and strong acid. Journal of Environmental Science and Health Part a- Toxic/Hazardous Substances & Environmental Engineering 1998b; 33: 923-950.
Ye G, Burström E, Fällman A-M. Utilisation and stabilisation of steelmaking slags. AFR, 1995.
