Självständigt arbete Nr 102
Mineral Barrier Systems for the Treatment of Metal-polluted Water from an Alum Shale Deposit
Mineral Barrier Systems for the Treatment of Metal-polluted Water from an Alum Shale Deposit
Isabell Gärtner
Isabell Gärtner
Uppsala universitet, Institutionen för geovetenskaper Kandidatexamen i Geovetenskap, 180 hp
Självständigt arbete i geovetenskap, 15 hp Tryckt hos Institutionen för geovetenskaper Geotryckeriet, Uppsala universitet, Uppsala, 2014.
Oil and gas were recovered from alum shale (black shale) at Kvarntorp, Närke, during a period of 24 years. One of the remnants of this industry is a 100 m high deposit with high contents of uranium, arsenic, molybdenum, vanadium and other elements. Presently the leakage of metals from the deposit into nearby streams is rather modest but will most likely accelerate in the near future. One way to prevent an uncontrolled leakage of these elements from the deposit into the environment would be to install a filter with an effective adsorbent that after saturation could easily be regenerated and reused.
The filter could not only be used to reduce the impact of metal leakage to the environment but also to reclaim valuable elements like nickel, vanadium, uranium and molybdenum. Unfortunately such a filter does not exist today, but there are a wide range of minerals that have high adsorbing capacities and could serve as filter components in a mineral barrier system. The adsorbing properties of some natural minerals with respect to selected metals that are abundant in the Kvarntorp deposit are studied in this project. These minerals are bentonite clay, Leca© (burnt clay), unburnt shale (stybb), burnt shale (rödfyr), apatite, peat and bark. In the experiments each sorbent was blended with artificial groundwater and a certain amount of metal stock solution.
Samples were taken at five different pH.
The results show that the two shale-products stand out from the others and have the best adsorbing qualities for nickel, copper and zinc. At a pH over 7 between 96 and 99% of the metal ions were removed from the solution.
Unburned shale especially shows remarkably good results for vanadium throughout the tested pH range of 3,4-7,7 between 98 and 99% of the vanadium ions were removed from the solution. Unburned shale performs best of the tested adsorbents and could probably be used as an adsorbent in a geological barrier but more research is needed.
Handledare: Bert Allard, Stefan Karlsson och Viktor Sjöberg (externa, ÖU); Roger Herbert (UU)
Självständigt arbete Nr 102
Mineral Barrier Systems for the Treatment of Metal-polluted Water from an Alum Shale Deposit
Isabell Gärtner
Abstract
Oil and gas were recovered from alum shale (black shale) at Kvarntorp, Närke, during a period of 24 years. One of the remnants of this industry is a 100 m high deposit with high contents of uranium, arsenic, molybdenum, vanadium and other elements. Presently the leakage of metals from the deposit into nearby streams is rather modest but will most likely accelerate in the near future. One way to prevent an uncontrolled leakage of these elements from the deposit into the environment would be to install a filter with an effective adsorbent that after saturation could easily be regenerated and reused. The filter could not only be used to reduce the impact of metal leakage to the environment but also to reclaim valuable elements like nickel, vanadium, uranium and molybdenum. Unfortunately such a filter does not exist
today, but there are a wide range of minerals that have high adsorbing capacities and could serve as filter components in a mineral barrier system. The adsorbing
properties of some natural minerals with respect to selected metals that are abundant in the Kvarntorp deposit are studied in this project. These minerals are bentonite clay, (burnt clay), unburnt shale (stybb), burnt shale (rödfyr), apatite, peat and bark. In the experiments each sorbent was blended with artificial groundwater and a certain amount of metal stock solution. Samples were taken at five different pH.
The results show that the two shale-products stand out from the others and have the best adsorbing qualities for nickel, copper and zinc. At a pH over 7 between 96 and 99% of the metal ions were removed from the solution.
Unburned shale especially shows remarkably good results for vanadium throughout the tested pH range of 3,4-7,7 between 98 and 99% of the vanadium ions were removed from the solution. Unburned shale performs best of the tested
adsorbents and could probably be used as an adsorbent in a geological barrier but more research is needed.
Sammanfattning
Under en period av 24 år utvanns olja och gas ur alunskiffer i Kvarntorp. Ett av de tydligaste spåren efter skifferoljeutvinning är den 100 m höga Kvarntorpshögen som enbart består av restprodukter från denna industri, innehållande höga halter av uran, arsenik, molybden, vanadin och andra grundämnen. För tillfället sker utlakning av metaller i begränsad omfattning men den kommer förmodligen att öka betydligt i framtiden. Ett filter med en effektiv adsorbent skulle kunna rena lakvattnet och på så sätt förhindra en okontrollerad spridning av metaller från deponin. När filtret är mättat kan värdefulla metaller som nickel, vanadin, uran eller molybden utvinnas ur
filtermaterialet. Tyvärr finns inte något sådant filter för närvarande. Däremot finns en rad lovande adsorbenter som skulle kunna användas som filtermaterial.
I detta projekt studeras några mineralers förmåga att adsorbera utvalda metaller vilka är vanligt förekommande i laklösningarna från skifferdeponin. Dessa mineral är bentonitlera, (bränt lera), obränt skiffer (stybb), bränt skiffer (rödfyr),
apatit, torv och bark. I försöken blandades varje adsorbent med konstgjort grundvatten och en fast mängd metallstamlösning. Prover togs vid fem olika pH.
Resultaten visar att i synnerhet de båda skifferprodukterna stybb och rödfyr utmärker sig med sina goda adsorptionsresultat för nickel, koppar och zink. I pH området över 7 adsorberades mellan 96 och 99 av joner från lösningen. Stybb visar särskild god adsorption för vanadin under alla uppmätta pH värdena som låg mellan 3,4 och 7,7. Mellan 98 och 99 % av vanadin jonerna avlägsnades från vattenlösningen. Av de testade mineralerna skulle stybb lämpa sig bäst som adsorbent i en geologisk barriär men mer forskning behövs.
Table of contents
1. Introduction ... 1
1.1 Aim ... 1
2. Background ... 1
2.1 Historical background and the industrial process ... 1
2.2 The Kvarntorp deposit ... 3
2.3 Geology ... 4
2.4 Adsorption process ... 5
2.5 Treatment methods for metal contaminated leakages ... 6
3. Experimental methods ... 6
3.1 Preparing the artificial groundwater ... 7
3.2. Preparing the sorbents ... 8
3.3 Experimental procedure ... 9
3.4 Data evaluation ... 9
5. Discussion ... 11
6. Conclusions ... 13
7. Acknowledgement ... 13
8. References ... 13
9.1 Appendix I ... 16
9.2 Appendix II ... 18
9.4 Appendix IV ... 34
1
1. Introduction
The Kvarntorp shale deposit is a residual product of the oil shale industry which was active in the area during 1942-1966. Scandinavian black shale is naturally rich in metals and metalloids and the process of oil shale extraction increased these concentrations even more. Since deposition the spoil heap is exposed to rain and snowfall. Natural weathering is leading to an uncontrolled leakage of these metals and metalloids to the nearby streams. Although the extent of problem is limited right now it will increase in the future. At the same time, some of those metals are a valuable resource which should be reclaimed. A mineral sorbent could act as a geological barrier or filter where metal ions get temporarily adsorbed.
1.1 Aim
The aim of this study is to investigate the capacity of selected minerals to adsorb a number of metal ions under controlled laboratory conditions representative of the conditions at the field (Kvarntorp deposit). A number of mineral sorbents were chosen which are known to have good sorption qualities, such as clays and organic material (peat and bark) but also black shale and incinerated shale (stybb and rödfyr respectively) which have not been used for this purpose before (Socialstyrelsen, 2006). These potential adsorbents were also chosen because they are readily available and relatively cheap.
This study is part of a larger project where the ultimate goal is to find adsorbents with both high capacity and high selectivity. These adsorbents could then be used to make a full scale, and ideally reversible, filter which could be tested in the Kvarntorp area, as well as other similar sites.
2. Background
In this chapter a brief introduction to the overall geology of the Kvarntorp area, the industrial process of shale oil extraction and the remains of that process are given. A short overview of the most important adsorption mechanism and a summary of a number of studies on how mineral absorbents were used to treat metal polluted water will give a better understanding on which mechanism are important.
2.1 Historical background and the industrial process
During the second World War an import stop caused an oil crisis in Sweden and forced the authorities to look for alternative sources of oil within the country. One suitable alternative was the kerogen-containing black shale found in several places in Sweden including them the Kvarntorp area, Närke, about 10 km south of Örebro (see Appendix I) and mount Billingen, Västergötland (Länsstyrelsen, 2005). The content of organic matter is higher in the Närke shale (20% versus 13%) whereas the
uranium content is higher in the Billinge shale (300 g/t versus 140 g/t). The oil shale plant was built in Kvarntorp in 1941 and a factory to mine and extract uranium was
2 built in 1960 in the Billingen area (Ranstad) (SGU 1985). The oil shale at Kvarntorp was mined in open pits. The mining product was transported to the nearby factory (see Figure 1 and 2) where the shale was roasted in an oven without oxygenation until the oil started to vaporize. The vapor was then cooled down and the liquid oil and gas collected (Miljödepartementet, 2001). ). This process, pyrolysis, is also known as dry distillation (Nationalencyklopedin 2014).The ash, sometimes still hot, was then transported to the Kvarntorp waste deposit along with the fine-grained unprocessed shale called stybb (Länsstyrelsen, 2005).
The annual production was about 82 700 t of oil and 34 900 t of gasoline were produced per year as well as about 11 100 t of gas (Sveriges geologiska
undersökning, 2003).
Figure 1: Shale oil company at Kvarntorp 1945 (Source: Örebro stadsarkiv, photographer unknown).
3 Figure 2: Kvarntorpshögen and the oil shale company in the 1970s (Source: Örebro
stadsarkiv, photographer unknown).
2.2 The Kvarntorp deposit
The Kvarntorp waste pile (see Figure 3) is approximately 100 m high and consists of approximately 28 million ton material; basically all of it is either incinerated shale (rödfyr) or discarded shale (stybb). The deposit covers an area of about 0,44 (Länsstyrelsen 2005). In total about 50 million ton of shale were mined but there are still about one billion ton of shale left in the ground in the eastern part of Närke (SGU 1985). After the war the oil extraction plant proved to be economically unsuccessful and even a widening of the product range and improvements in the production where not enough and the plant closed in 1966 after 24 active years (Länsstyrelsen 2005).
The overall content of metals and metalloids is very high in the Kvarntorp black shale (see Table 1) some of those are uranium, vanadium, nickel and
molybdenum which are economically valuable and could be extracted (Andersson 1985, Länsstyrelsen 2005). Both the mineral composition and the concentration of the different metals are not homogenous throughout the shale layer. Some metals can be abundant in one horizon and almost missing in another; the same applies to the organic content (Westergård 1941, SGU 1971, SGU 1972, Andersson 1985).
Figure 3: Kvarntorpshögen in the 1970s (Source: Örebro stadsarkiv, photographer unknown).
Table 1: Content of some metals given in ppm in different kind of shales (Karlsson 2013;
Andersson 1985, U.S. Geological Survey 2014).
Närke (Sweden) Västergötland (Sweden)
Millard (Utah, USA)
V 580-690 450-750 61-88
Ni 95-110 90-200 35-60
Cu 120 110-190 3,1-36
Zn 27-32 130-150 59-110
Mo 240 70-340 2-3,2
U 140 10-300 -
4
2.3 Geology
The sedimentary rocks of eastern Närke were deposited on Precambrian bedrock during the Middle Cambrian to the Early Ordovician era in a shallow sea (see Figure 4), during a time period of approximately 30 million years. The border between bedrock and sedimentary layers is marked by an unconformity in the form of a conglomerate layer, followed by Cambrian sandstone, clay, alum shale (also called black shale) and limestone on top (see Figure 5). The sedimentary layers are slightly tilted and dipping shallow to the south-west. Average thickness of the black shale layer is 19 m in the north of the area and about 17 m in the south (Westergård 1941, SGU 1971, SGU 1972).
Figure 4: Paleogeografic map of Europe including Sweden in the Ordovician 450 Ma (credit to Ron Blakey, Colorado Plateau Geosystems).
5 Figure 5: Geological map of the Kvarntorp area (Westergård 1941, C442). Ortocerkalksten = Ordovician limestone; Alunskiffer = alum shale; Skifferlera = clay; Sandsten = sanstone;
Urberg = crystalline bedrock.
The alum shale contains few fossils and those that can be found are usually concentrated in the stinkstones (Westergård 1941, SGU 1971, SGU 1972,
Andersson 1985). Stinkstones are lenses of limestone with a high content of organic matter and a distinct smell (Nationalencyklopedin 2014). Lenses of kolm (coal like bedrock) are quite common and those lenses have a high percentage of organic material even higher than the black shale and are exceptionally rich in uranium, about 5000 ppm or 0,5% (Westergård 1941, SGU 1971, SGU 1972, Andersson 1985, Nationalencyklopedin 2014). The alum shale is not only rich in organic material but also in pyrite (see Table 2) which occurs in nodules throughout the shale layer.
Table 2: Average mineral composition of the Närke alum shale (SGU 1972).
Mineralogical composition Content (%)
Muscovite and other mica minerals 34-35
Quartz 27-28
Organic material 21
Pyrite 13
Kaolinite 4
2.4 Adsorption process
There are several different adsorption processes. Physical adsorption, chemisorption and electrostatic sorption are the three main types. Some are more rapid than others and some are reversible others not. These differences are of importance because
6 they determine on how fast the sorbent is binding the metal ions, if the process is too slow the sorbent might not be usable as a geological barrier. A non reversible
process would make a reusable filter impossible.
Physical adsorption is the effect of weak van der Waals forces. Atoms or molecules need to be quite close to another for the forces to be active. No new chemical bonds are formed and this is a fast process but usually easily reversible.
Chemisorption means that some bonds break and new ones are formed. This process is slower than physical adsorption and often irreversible (Atkins and Jones 2010, Majer 1980, Nationalencyklopedin 2014). In electrostatic adsorption the charge of the atom is of major importance but also the pH. This process is also called an ion exchange process. The atoms don´t need to be as close together as in physical adsorption (Atkins and Jones 2010, Majer 1980).
A clear distinction must also be made between primary and secondary adsorption. In the first one the element is embedded in the sorbent´s mineral
structure whereas in the second one the element is only bound to the surface of the sorbent ( Majer 1980).
2.5 Treatment methods for metal contaminated leakages
The removal of metal ions from a water solution by the use of a mineral absorbent has been investigated in a great number of studies. Some of those studies have been chosen to give a short overview of which adsorbents and metal ions were used and which results were achieved.
Laboratory experiments showed that apatite has a very good ability to remove lead from a water solution but can also adsorb cadmium and zinc, but to a lesser degree especially at low pH (<4). Up to 41 mg zinc per gram apatite can be bound, probably through surface complexation and ion exchange but also due to the formation of amorphous solids (Chen et al 1997). A study at a mine site in Ontario (Blowes et al, 2000) showed that a geological barrier containing mainly biological waste products like wood chips and compost can reduce the nickel content of the mine water significantly, probably due to bacterial sulfate reduction and precipitation of metal sulfides.However an oxygenation of the barrier would release high
concentration of nickel and other adsorbed metals (Blowes et al, 2000). Many readily available products like different algaes, ground coffee, lemon peel and bark exhibited medium to high adsorption capacities and are efficient in the removal of mercury (Carro et al, 2009).
3. Experimental methods
The uptake of metals from an aqueous solution by some selected solid minerals expected to have high adsorption capacity was measured. The adsorption capacity is the ability of an adsorbent to bind metal ions at the surface and inside the mineral grain. The sorbent, artificial groundwater and metal stock solution were mixed together. To give metal ions time to adsorb to the sorbent both were placed in a shaker for approximately 18 hours and then the samples were collected. The batch
7 experiments were done at constant temperature but with varying pH, to determine if the sorption is pH depended or not. The liquid-to-solid ratio (L/S-ratio) was constant (50 L/kg).
3.1 Preparing the artificial groundwater
An artificial groundwater was prepared (see Table 3), to resemble a natural
groundwater but with fewer trace elements like iron and without humic acids which could have an effect on the sorption process as heavy metals often react with iron or magnesium to form solid complexes. Organic compounds like humic and fulvic acids can together with metal ions form a chelate and then precipitate from the solution (Drever 1997). This process removes metal ions from the solution and does prevent a correct interpretation if the mineral adsorbent actually adsorbs ions and is therefore not wanted in these experiments.
Table 3: Preparation of artificial groundwater.
Salt M Amount, mg/l Content, mg/l Content, mmol/l x
O
246,51 24,65 Mg: 2,43 : 9,61 Mg: 0,10 : 0,10 KCl 74,55 7,46 K: 3,91 Cl: 3,54 K: 0,10 Cl: 0,10 x
O
147,02 29,40 Ca: 8,02 Cl: 14,18 Ca: 0,20 Cl: 0,40 Na 84,02 84,02 Na: 22,99 :
61,02
Na: 1,00 : 1,00
To prepare two liters of artificial groundwater, the four salts were weighed separately and then 50 ml of distilled water was added to each of them. The four solutions were then mixed and 1800 ml of distilled water was added.
To this artificial groundwater a metal solution was added with a number of metals that are common both in the Kvarntorp spoil heap and in the stream water which are recipients of leakage from the spoil heap. The solution consists of five readymade metal solutions with a concentration of 1g/l. The solutions where weighed on a scale to give a concentration of 2x mol/l in the finished solution (see Table 4). The uranium solution was prepared in the lab. 210 mg of U (N x 6 O was dissolved in 100 ml distilled water. 10,158 g of this solution and the five other metal solutions (amounts given in Table 4) where then diluted to make 100 ml metal stock solution.
Table 4: Metal stock solution.
Element Amount (g) in stock solution Concentration I final solution(μmol/l)
Vanadium 2,217 20,9
Nickel 2,497 19,6
Zinc 2,792 19,5
Copper 2,666 20,7
Molybdenum 4,098 20,6
8
3.2. Preparing the sorbents
The sorbents can be grouped into clays (bentonite and ), shale products (stybb and rödfyr), a mineral (apatite) and biological products (peat and bark). Clays are known for their ion exchange capacities and their negative surface charges (Morel and Hering 1993) should attract positively charged ions, but will probably repel negatively charged ones. Biological materials have shown good sorption qualities in previous studies (Carro et al, 2009) and have therefore been included. The shale products have not been tested in previous studies as an adsorbent but as there are large amounts of them in the spoil heap and they are therefore both easily available and cheap they are included. Another reason is that they are naturally rich in a number of elements which might indicate that they have very good adsorption capacities.
The sorbents were crushed in a mortar and then sieved (see Table 5).
Table 5: Size fraction and available amounts of sorbents.
Sorbent Sieve (micron)
Bentonite 250
1 60
Stybb (Kvarntorp) 60
Rödfyr (Kvarntorp) 20
Apatite [Ca5(PO4)3(F,Cl,OH)] 20
Peat 60
Bark2 20
To deplete the two shales of any metals and peat and bark of organic acids those four sorbents (see Table 5) were first leached in 100 ml 0.01 M HCl. The leaching of shales was necessary to prevent a leakage of metals from the sorbents which could falsify the results of the experiments. After 72 hours the acid was poured off and the sorbents washed with distilled water. Then 100 ml 0.1 M NaOH was added to each sorbent and leached for 72 hours. For peat and bark the leaching process with NaOH was repeated because the leaching solution was still strongly discolored. The second leaching process was stopped after 24 hours and the sample again washed in
distilled water.
All of the sorbents were then washed with 100 ml distilled water. After 10 minutes of sedimentation the water was poured off to get rid of colloid particles. This process was repeated once and then the sorbents were mixed with 50 ml artificial groundwater (without metal stock solution). The samples were then placed in a shaker at 30 RPM . After 18 hours the samples were again washed in distilled water and centrifuged at 5000 RPM for one minute. For the bentonite samples this was repeated once more. The excess water was then poured off and the samples placed in an oven at 105°C for 72 hours to dry out.
1 is a porous ceramic material consisting mainly of burnt clay.
2 Dried pinebark, sold under the name Zugol©.
9
3.3 Experimental procedure
The experiments were divided in two phases. In the first one the metal ion content for all sorbents were analyzed at three different pH values. In the second one the
sorbents with the highest sorption capacity were selected and new samples were taken at one pH level but with different metal ion contents to confirm the results from phase one. For all sorbents, samples at two additional pH were taken.
In the first phase for each sorbent three test tubes where prepared. Into each tube 0,80 g sorbent, 35 ml artificial groundwater and one milliliter metal stock solution was blended. The pH was measured and adjusted to one low, one neutral and one high value for each sorbent by adding 0,1 M sodium hydroxide and 0,1 M sulfuric acid. The samples were then weighed and artificial groundwater was added so that every sample weighed 40 g. The samples were placed in a shaker at 30 RPM. After 18 hours the samples were centrifuged and the pH was measured again.
One milliliter of each sample was taken, unfiltered, and diluted 1:100 and then analyzed with the ICP-MS, Agilent 7500cx.
In the second phase the pH of the samples with the low and the high pH were adjusted again to a value in between the low and the neutral pH and in between the neutral and the high pH. At the same time new samples were prepared for the most promising four sorbents as above in phase one but with a diluted metal stock solution 1:5 and 1:20. In the samples of , unburned shale, burnt shale and apatite at neutral pH from phase one, additional metal stock solution was added to make a concentration of 6x mol/l .
3.4 Data evaluation
The raw analyzing results are listed in Appendix II. To get a better understanding of how much of the metals ions were removed from the solution, what amount was probably bound to the sorbent and how the distribution between the metals ions in solution and the metal ions bound to the sorbent are the raw data was put in the mathematical equations listed below and the results are given in Appendix II.
V= volume (L) of the samples= 40x m= mass (kg) of the samples = 0,8x
= Metal ion concentration start (mol/L) = 2x
C = Metal ion concentration at the end of the experiment (mol/L)
= adsorbed amount (mol/kg) Distribution coefficent = (L/kg)
S = metal ions removed from solution (%)
=
= [ ] x =
= [ ] x 100
10
4. Results
The batch experiments were performed to show how many metal ions from an aqueous solution different mineral sorbents can adsorb. All sorbents showed some adsorption but some performed better than others.
Bentonite showed good adsorption for nickel, copper, zinc and uranium at high pH and molybdenum at low pH, but little adsorption for vanadium (see Table 6). As a side note, it can be mentioned that amongst others strontium (120-2004 ppm) was found in the analyzed samples, this ion was not part of the experiment.
shows very good adsorption for nickel, copper especially at higher pH and very modest adsorption for molybdenum. Uranium adsorption was also very good at least at both the highest and lowest pH (see Table 6). The good adsorption results for nickel, copper and zinc were confirmed in the second phase of the experiments with tripled concentration of metal stock solution (see Table 7).
The experiments show that unburned shale has high adsorption for vanadium and copper throughout the tested pH range. Unburned shale has also very good adsorption for nickel and zinc at a high pH and for molybdenum and uranium at a more acidic pH (see Table 6). The very good absorption results for vanadium and copper were confirmed in the second phase of the experiments. The results of the adsorption for the other metals in the second phase were mixed (see Table 7).
The burnt shale has very good adsorption of copper and zinc at all pH levels. This shale has also very good adsorption capacity for vanadium at a low pH and nickel at high pH. The results of uranium adsorption were mixed and with
somewhat surprisingly high concentration of uranium at a neutral to slightly basic pH but very low at low and high pH (see Table 6 and 7). The shale did also release molybdenum and arsenic (31-147 ppm) into the samples.
Apatite seems to have very good sorption capacity for nickel, copper and zinc at higher pH and for uranium at all pH levels with exception for a very high
concentration at pH around 8 (see Table 6 and 7).
Peat and bark both have very good absorption qualities for nickel, zinc, copper, vanadium and uranium. All analyzing tubes had a brownish precipitation at the bottom (see Table 6).
Table 6: Metal ions removed from solution (%). Starting concentration of metal stock solution 2x mol/l.
Bentonite Stybb Rödfyr Apatite Peat Bark
pH range 1,86–
9,76
5,86–
10,29
3,41–
7,70
5,07–
10,00
5,18–
10,24
6,02–
7,98
6,31–
8,03
V 8,8–39 19-60 98-99 122-95 10-20 13-78 12-71
Ni 16-85 35-98 8,2–97 45-99 2,8–98 89-91 93-95
Cu 48-84 77-97 67-99 93-98 69-97 84-87 87-91
Zn 9,8–82 58-97 12-99 67-99 16-98 86-90 92-96
Mo 8,7–96 2,6–25 2,2–99 -19-42 4,6–31 17-51 29-49
U 1,8–62 32-89 62-98 27-99 37-98 82-88 76-82
11 Table 7: Metal ions removed from solution (%).Starting concentration of metal stock solution 6x mol/l.
Stybb Rödfyr Apatite
pH 9,19 6,56 7,37 6,46
V 32 96 69 22
Ni 100 76 92 34
Cu 99 99 99 93
Zn 100 94 99 73
Mo 11 20 3,9 13
U 59 82 48 96
5. Discussion
The leakage of metals from the Kvarntorp spoil heap is at the moment only modest mainly due to the fact that the heap is in most areas about 80-90° C hot in some even as much as 500-700° C. Cause of this heat development is well oxygenated shale that is slowly burning underground. This process is expected to last for at least another one to two hundred years. During the cooling period the metals leakage will most likely increase (Länsstyrelsen 2005).
The differences in adsorption for bentonite are probably due to a
negative surface charge of the clay particle. The Pourbaix diagram (see Appendix III) of nickel, copper and zinc show that in a well oxygenated system at a pH below 3 those elements predominantly exist as and uranium as
which would fit the theory. Molybdenum exists as different ions throughout the pH range but all of them are negatively charged or uncharged species. As adsorption happened at a low pH it was either due to precipitation or the surface charge of the clay was positively charged which is probably unlikely. The vanadium concentration was more or less stable throughout all pH levels but vanadium can exist in a variety of species and it is not predictable which one was predominant in the artificial
groundwater. Bentonite was not considered for the second phase of the experiments because it released strontium into the water solution. The strontium must be naturally occurring in the clay as it was not part of the metal stock solution. It must be
mentioned that the pH values of the bentonite were out of the targeted zone and the results of the sorption experiments might be misleading.
The sorbent showed very good adsorption for nickel and copper probably due to a negative surface charge. Despite these good results would not be suitable in a full scale geological barrier for uranium because even a minor change in the pH of the water could release high concentrations of uranium.
Unburned shale had the best sorption results of the tested sorbents and should be considered highly promising for a future full scale geological barrier. The somewhat mixed results of the second phase maybe due to the fact that the pH was close to neutral (6,56) and if the pH had been either lower or higher the adsorption would have probably been better for either nickel and zinc or molybdenum and uranium.
12 The results for burnt shale, especially for uranium, should be checked and confirmed if they truly are representative or show erroneous results. It is also surprising that burnt shale and unburned shale which are the same kind of shale but one heated and the other not show quite different results in the adsorption
experiments.
The results for apatite are similar to those for burnt shale and should also be checked.
All batches of peat and bark were strongly colored which speaks for a release of humic acid even after one acid and two basic washings. As there where brownish precipitation in the sample tubes it is impossible to tell if the lower
concentration of metal ions in the solution was due to an adsorption process or if the ions bound to the humic acids and precipitates. As the main goal of the experiments was to show the adsorbing qualities of the sorbents neither peat nor bark where considered for the second phase of the experiments even if their capacities might be good.
It seems that none if the sorbents reached saturation even when the amount of metal stock solution was tripled the adsorption was same. Overall it seems that the anion (molybdenum) showed better adsorption at the lower end of the ph range and the cations (nickel, zinc and copper) at the higher end. For vanadium and uranium the results were more mixed, which in the case of vanadium is plausible as it can exist both as cations and anion in a solution (see Appendix III).
The plot of the distribution coefficient ( ) versus the pH (see Appendix IV) illustrates that nickel, copper and zinc show a similar adsorption pattern for all four sorbents from phase two. The is lower at a low pH and increases as the pH gets more basic, this means that the concentration in the solution of these metal ions is higher at low pH. For molybdenum there seems to be an opposite trend high at a low pH and lower at a higher pH. Vanadium, and uranium does show the same clear pattern wich could be due the different charges of the metal ions at different pH (see Appendix III) or some other property of the sorbent like ion exchange.
The pH adjustments were rather difficult as some of the sorbents
reacted and started to leach acids (see Table 8 and 10) which lowered the pH. Some pH values changed quite dramatically for unknown reasons which could also have lead to misleading results.
The experiments show that there are a number of promising adsorbents for most of the metals ions, except for molybdenum, but many of them have their best sorption results in quite a narrow pH range. This suggest that in a future full- scale filter not one but several different sorbents should be used to maximize the effect and efficiency of the geological barrier.
13
6. Conclusions
Based on the results of the study, the following conclusions can be drawn:
- There were not enough ph values to determine the maximum capacity of each sorbent.
- Despite acid and basic leakage of the sorbents peat and bark leached organic acids and the burnt shale arsenic and molybdenum.
- Peat and bark might have good potential for the removal of nickel, copper, zinc and uranium but should be leached more thoroughly to remove the organic acids.
- Vanadium is most effectively adsorbed by unburned shale.
- Nickel, zinc and copper are adsorbed equally well by all the sorbents.
- None of the sorbents is effective for the removal of molybdenum.
- The adsorption for uranium was mixed but unburned shale performs best.
-The minimum loading capacity is between 10 and 14 mmol/kgfor the four sorbents from phase two.
Suggestions for further experiments/studies are:
- A similar batch experiment like in this study with the four sorbents ( , unburned shale, burned shale and apatite), plus maybe peat and bark and some other
elements that are abundant in the Kvarntorp spoil heap, like arsenic and cadmium.
- Experiments with both groundwater and surface water from the Kvarntorp area, to determine how effective the sorbents are.
- A long-term study with a full-scale geological barrier with unburned shale in combination with one or several other adsorbents in the Kvarntorp area.
7. Acknowledgement
I´d like to thank the librarians at Geobiblioteket and SGU´s library for their good advice and help in finding maps and articles. Many thank to my supervisors for their input and help in analyzing my samples. Special thanks to Bert Allard without his knowledge and experience this project would not have been possible.
8. References
8.1 Literature
Andersson, Astrid (red.) (1985). The Scandinavian alum shales. Uppsala: SGU Atkins, P. W. & Jones, Loretta (2010). Chemical principles: the quest for insight. 5.
ed. New York: W.H. Freeman, 792 pages.
, Petr & Majer, Vladimír (1980). Trace chemistry of aqueous solutions: general chemistry and radiochemistry. Amsterdam, 252 pages.
Blowes, D.W., Ptacek, C.J., Benner S.G., McRae Ch.W.T., Bennett T.A., Puls R.W.
(2000). Treatment of inorganic contaminents using permeable reactive barriers. Journal of contaminant hydrology, Volume 45, Issues 1-2, September 200, pages 123-137.
14 Carro L., Herrero R., Barriada J.L., Sastre de Vicente M.E. (2009). Mercury removal:
a physicochemical study of metal interaction with natural materials.
Journal of Chemical Technology and Biotechnology, volume 84, issue 11, pages 1688-1696.
Chen X., Wright J.V., Conca J.L., Peurrung L.M. (1997). Evaluation of heavy metal remedation using mineral apatite. Water,Air & Soil Pollution, Volume 98, issue 1-2, pages 57-78.
Drever, James I. (1997). The geochemistry of natural waters: surface and
groundwater environments. 3. ed. Upper Saddle River, N.J.: Prentice Hall, 436 pages.
Karlsson Lovisa E. (2013). Release of metals from unprocessed- and processed shale from Kvarntorp – As a function of solution pH. Örebro University, School of Science and Technology.
Länsstyrelsen Örebro län (2005). Kvarntorpsområdet-Studie av kvarntorpshögen, (Dr. 5771-08677/2005).
Miljödepartementet (2001). Kartläggning av vissa förorenade områden m.m. (Dnr 08- 748/2001).
Morel, François M. M. & Hering, Janet G. (1993). Principles and applications of aquatic chemistry. New York: Wiley, 588 pages.
Socialstyrelsen (2006). Dricksvattenrening med avseende på uran. (Artikelnr. 2006- 123-11).
Sveriges geologiska undersökning (1971). Berggrundsgeologiska kartbladet Örebro SV: med 4 geologiska profiler. Beskrivning till berggrundskartbladet Örebro SV. Stockholm: SGU.
Sveriges geologiska undersökning (1972). Geologiska kartbladet Örebro SV.
Beskrivning till geologiska kartbladet Örebro SV = Description of the geological map Örebro SV. Stockholm: SGU.
Sveriges geologiska undersökning (1985). The Scandinavian Alum Shales. (ISSN 0348-1352).
Sveriges geologiska undersökning (2003). Mineralmarknaden -Tema:Uran. (ISSN 0283-2038).
Westergård, A. H. (1941). Skifferborrningarna i Yxhultstrakten i Närke 1940.
Stockholm.
8.2 Internet resources
Hydra/Medusa freeware, http://www.kth.se/en/che/medusa/downloads-1.386254, retrieved 2014-04-01.
Leca lättklinker,
www.weber.se/fileadmin/user_upload/pdf/leca/broschyrer/leca_lattklinker _egenskr.pdf, retrieved 14-11-04.
Nationalencyklopedin (2014). Adsorpion.
www.ne.se/lang/adsorption, retrieved 2014-04-23.
Nationalencyklopedin (2014). Kolm.
15 www.ne.se/lang/kolm, retrieved 2014-05-07.
Nationalencyklopedin (2014). Orsten.
www.ne.se/lang/orsten, retrieved 2014-06-07.
Nationalencyklopedin (2014). Pyrolys.
www.ne.se/lang/pyrolys, retrieved 2014-04-22.
U.S.Geological Survey (2014). tin.er.usgs.gov/ngdb/rock/find-rock.php, retrieved 14- 04-17.
Zugol AB, zugol.com/Filer/pdf/sakerhetsdatablad.pdf, retrieved 14-04-11.
8.3 Pictures and maps
Lantmäteriet, www.kso2.lantmateriet.se/#, retrieved 14-03-14 Ron Blakey, Colorado Plateau Geosystems,
cpgeosystems.com/450_Ord_EurMap_sm.jpg, retrieved 14-05-16.
Örebro stadsarkiv,
mediaarkiv.orebro.se:8080/Cumulus_ark/ShowRecord.jsp?server=localh ost&catalogName=Bildarkivet&recordID=32003&encoding=UTF-
8&useGuestLogin=true&recordView=SearchResult_RecordInfo, 14-05- 07
Örebro stadsarkiv,
http://mediaarkiv.orebro.se:8080/Cumulus_ark/ShowRecord.jsp?server=l ocalhost&catalogName=Bildarkivet&recordID=13831&encoding=UTF- 8&useGuestLogin=true&recordView=SearchResult_RecordInfo, 14-05- 07
Örebro stadsarkiv,
http://mediaarkiv.orebro.se:8080/Cumulus_ark/ShowRecord.jsp?server=l ocalhost&catalogName=Bildarkivet&recordID=13832&encoding=UTF- 8&useGuestLogin=true&recordView=SearchResult_RecordInfo, 14-05- 07
16
9. Appendix
9.1 Appendix I
Figure 17: Map over the Kvarntorp area (Lantmäteriet 2014).
17 Figure 18: Map over Örebro, Kumla and Kvarntorp (Lantmäteriet 2014).
18
9.2 Appendix II
Table 8: pH values at the start of the experiment (first measurement) and after 18 hours before the samples for analysis where taken (second measurement).
Sorbent pH (first measurement) pH (second measurement)
Bentonite 1,76 1,86
8,72 8,60
10,17 9,76
Leca 5,31 5,86
9,32 8,83
10,50 10,29
Stybb (black shale) 5,38 3,41
8,72 6,23
10,40 7,70
Rödfyr (incinerated shale) 5,67 6,76
9,32 8,65
10,52 10,00
Apatite 5,67 5,91
8,22 8,14
10,51 10,24
Peat 5,02 6,02
8,60 6,42
10,40 7,70
Bark 5,47 6,45
8,88 6,70
10,37 7,43
Artificial groundwater with sorbent (blank)
5,70 5,53
7,79 6,27
10,34 9,88
Water solution (without metal stock solution)
6,78
Table 9: Analysis result phase one, concentrations in ppm. Samples with 1 ml metal stock solution.
Bentonite Leca Stybb Rödfyr Apatit Peat Bark Blank Element pH
V low 842 427 15 137 953 285 518 1072
neutral 748 842 15 572 932 338 792 1060
high 971 864 18 760 852 927 942 1061
Ni low 965 751 1057 327 1119 104 68 1160
neutral 168 42 306 49 264 99 69 1159
high 248 33 45 22 23 99 66 1144
Cu low 648 100 410 26 176 166 116 1253
neutral 200 47 20 33 37 157 141 1240
high 278 44 16 39 33 173 168 1228
Zn low 1203 560 1174 28 896 174 78 1367
neutral 242 43 82 11 32 151 69 1318
high 375 38 11 13 37 134 51 1308
19
Mo low 115 1853 64 2341 1849 1131 1132 1969
neutral 1812 1907 1301 2363 1838 1528 1167 2057
high 1700 1934 1942 2312 1895 1651 1417 1923
U low 4812 541 1851 1136 94 659 950 4924
neutral 3866 3327 173 3577 3075 643 1184 4908
high 1886 401 476 93 1440 764 937 4859
Table 10: pH values after second adjustment (third measurement) and after 18 hours before samples were taken for analysis (forth measurement).
Sorbent pH (third measurement) pH (forth measurement)
Bentonite 2,38 2,36
9,69 8,89
Leca 7,31 7,77
10,08 6,28
Stybb (black shale) 6,36 4,67
9,21 7,23
Rödfyr (incinerated shale) 6,54 5,07
10,01 9,80
Apatite 7,70 8,30
9,91 5,18
Peat 6,18 6,20
9,12 7,98
Bark 6,66 6,31
9,24 8,03
Table 11: Analysis result phase two, concentrations in ppm. Samples with 1 ml metal stock solution.
Bentonite Leca Stybb Rödfyr Apatite Torv Bark Element pH
V lower 648 797 10 57 947 230 381
higher 840 546 23 941 958 443 309
Ni lower 875 227 917 637 1073 124 85
higher 188 26 38 17 64 107 59
Cu lower 491 280 59 92 383 193 141
higher 237 44 18 59 38 203 142
Zn lower 1177 305 1060 442 1118 238 114
higher 518 37 22 23 40 189 92
Mo lower 71 1485 28 1153 1381 972 1139
higher 1468 1485 1504 1922 1395 1288 1024
U lower 3956 2760 117 210 80 605 924
higher 2806 2402 1414 59 605 889 863
Table 12: Analysis result phase two, concentrations in ppm. Samples with 3 ml metal stock solution.
Leca Stybb Rödfyr Apatite
Element pH 9,19 6,56 7,37 6,46
V 2168 118 989 2495
Ni 16 846 287 2277
Cu 30 48 49 263
20
Zn 20 255 30 1063
Mo 5302 4785 5724 5212
U 6075 2629 7673 625
Table 13: Analysis result phase two, concentrations in ppm. Samples with 0,2 ml metal stock solution.
Leca Stybb Rödfyr Apatite
Element pH 7,21 5,66 8,87 6,18
V 148 10 252 276
Ni 444 85 16 295
Cu 386 16 35 53
Zn 270 34 19 167
Mo 283 293 916 595
U 889 30 909 185
Table 14: Analysis result phase two, concentrations in ppm. Samples with 0,05 ml metal stock solution.
Leca Stybb Rödfyr Apatite
Element pH 8,81 6,41 8,61 7,49
V 66 8 107 76
Ni 7 13 16 59
Cu 12 11 28 19
Zn 11 16 16 20
Mo 259 339 622 162
U 175 48 345 247
Table 15: Concentration of metal ions in solution (C) and concentration of metal ions bound to the sorbent bentonite ( ), their distribution coefficient ( ) and how many percent of the metal ions were removed from the solution (S). Samples with 1 ml metal stock solution.
C μmol/l
logC C0 μmol/l
Csolid μmol/kg
logCsolid kd l/kg
logkd S (%) Element pH
V 1,86 16.5 1,22
20,9
219 2,34 13,3 1,12 20,9
2,36 12,7 1,10 409 2,61 32,1 1,51 39,1
8,60 14,8 1,17 303 2,48 20,4 1,31 29
8,89 16,5 1,22 221 2,34 13.4 1,13 21,1
9,76 19,1 1,28 92 1,96 4,8 0,68 8,8
Ni 1,86 16,4 1,22
19,6
158 2,20 9,61 0,98 16,1
2,36 14,9 1,17 235 2,37 15,8 1,20 24
8,60 2,86 0,46 837 2,92 293 2,47 85,4
8,89 3,20 0,51 820 2,91 256 2,41 83,7
9,76 4,22 0,63 769 2,89 182 2,26 78,5
Cu 1,86 10,2 1,01
19,5
465 2,67 45,6 1,66 47,7
2,36 7,73 0,89 589 2,77 76,2 1,88 60,4
8,60 3,15 0,50 818 2,91 260 2,41 83,8
21
8,89 3,73 0,57 789 2,90 212 2,33 80,8
9,76 4,38 0,64 756 2,88 173 2,24 77,5
Zn 1,86 18,4 1,26
20,4
100 2 5,43 0,73 9,8
2,36 18 1,26 120 2,08 6,66 0,82 11,7
8,60 3,70 0,57 835 2,92 226 2,35 81,9
8,89 7,92 0,90 624 2,80 78,8 1,90 61,2
9,76 5,74 0,76 733 2,87 128 2,11 71,9
Mo 1,86 1,20 0,80
20,7
975 2,99 813 2,91 94,2
2,36 0,74 -0,13 998 3,00 1349 3,13 96,4
8,60 18,9 1,28 91 1,96 4,82 0,68 8,7
8,89 15,3 1,18 279 2,45 18,2 1,26 26,1
9,76 17,7 1,25 149 2,17 8,41 0,92 14,4
U 1,86 20,2 1,31
20,6
19 1,28 0,94 - 0,03 1,8
2,36 16,6 1,22 199 2,30 12 1,08 19,3
8,60 16,2 1,21 218 2,34 13,4 1,13 21,2
8,89 11,8 1,07 441 2,64 37,4 1,57 42,8
9,76 7,92 0,90 634 2,80 80 1,90 61,6
Table 16: Concentration of metal ions in solution (C) and concentration of metal ions bound to the sorbent Leca ( ), their distribution coefficient ( ) and how many percent of the metal ions were removed from the solution (S). Samples with 1 ml metal stock solution.
C μmol/l
logC C0
μmol/l
Csolid μmol/kg
logCsolid kd l/kg
logkd S (%) Element pH
V 5,86 8,38 0,92
20,9
626 2,80 74,7 1,87 59,9
6,28 15,7 1,19 263 2,41 16,8 1,23 25,1
7,77 10.7 1,03 509 2,71 47,5 1,68 48,7
8,83 16,5 1,22 219 2,34 13,3 1,12 20,9
10,29 17 1,23 197 2,29 11,6 1,07 18,9
Ni 5,86 12,8 1,11
19,6
341 2,53 26,7 1,43 34,7
6,28 3,87 0,59 787 2,90 203 2,31 80,3
7,77 0,44 -0,35 958 2,98 2177 3,34 97,8
8,83 0,80 -0,10 940 2,97 1175 3,07 95,9
10,29 0,56 -0,25 952 2,98 1700 3,23 97,1
Cu 5,86 1,57 0,20
19,5
897 2,95 571 2,76 91,9
6,28 4,41 0,64 755 2,88 171 2,23 77,4
7,77 0,69 -0,16 941 2,97 1364 3,13 96,5
8,83 0,74 -0,13 938 2,97 1268 3,10 96,2
10,29 0,69 -0,16 941 2,97 1364 3,13 96,5
Zn 5,86 8,57 0,93
20,4
592 2,77 69,1 1,84 58
6,28 4,67 0,67 787 2,90 169 2,23 77,1
7,77 0,57 -0,24 992 3,00 1740 3,24 97,2
8,83 0,66 -0,18 987 2,99 1495 3,17 96,8
10,29 0,58 -0,24 991 3,00 1709 3,23 97,1
