TRITA-LWR DEGREE PRO JE CT 14:01
A
RSENIC
R
EMOVAL
F
ROM
F
LUE
G
AS
C
ONDENSATE
W
ITH
F
ERRIHYDRITE
P
RECIPITATION
Louise Waldenström
© Louise Waldenström 2014 Masters level Degree Project
Department of Land and Water Resources Engineering Royal Institute of Technology (KTH)
SE-100 44 STOCKHOLM, Sweden
S
UMMARYCombustion of waste wood requires extra measures to be taken to prevent emissions of toxic substances. Waste wood is often contaminated with heavy metals; some substances can originate from chromated copper arsenate (CCA), which would increase the As concentrations in the flue-gas condensate. At the Idbäcken combined heat and power (CHP) plant waste wood is combusted and in the resulting flue gases toxic substances will end up. The flue gases are then condensed and the toxic substances are transferred to the condensate water. Arsenic (As), among other contaminants, needs to be removed before the treated water is sent to the recipient. There are various techniques for removal of As; precipitation with iron(III) chloride (FeCl3) is one method, and it is used today at the Idbäcken CHP plant. To ensure that the emission limit for As may be met also in the future, this method needs to be optimized.
As occurs mainly in the two valance states +5 (arsenate, As(V)) or +3 (arsenite, As(III)), where the latter is more difficult to remove from water. Experiments have shown that the As in the flue gas condensate at Idbäcken CHP plant contains mostly As(V), which can be removed by precipitation with FeCl3. In contact with water FeCl3 forms ferrihydrite, a substance with a very high specific area. The high specific area of ferrihydrite contributes to a high adsorptive capacity for As.
The removal efficiency of As is dependent on several factors of which pH and FeCl3 dosage are the most important. Good mass transfer is also believed to be important; however this factor is beyond the scope of the present work. Several experiments have been made in the laboratory with the aim to find the optimal conditions for As removal from condensate water using ferrihydrite precipitation. Both pH and FeCl3 dosages have been varied and different As concentrations have been used. Efficient adsorption of As(V) to ferrihydrite requires a condensate pH of 3-8, and is found to be most efficient in the lower range of this pH interval. According to literature studies, the maximum As(V)-removal occurs at pH = 4, which is consistent with the results in this work An increase in FeCl3 dosage increases the As(V)-removal up to a certain point (about 48 mol Fe3+/mol As), whereas further addition does not contribute to a more complete removal. Further investigations are needed to identify the minimum Fe3+ dosage and to verify the lab results by plants measurements.
A full-scale experiment was carried out at the Idbäcken CHP plant; unfortunately no conclusions regarding As removal efficiency could be drawn from this experiment. However, measurements showed that pH was lowered considerably when FeCl3 was added (also seen in lab experiment). Hence, there is a high risk of ending up at pH values lower than 3, which is lower than recommended. Therefore, pH measurement and adjustment of pH is highly recommended after FeCl3 addition. Additionally, the full-scale test showed that the high FeCl3 dosage necessary at higher As concentrations, could clog the filters. This may be prevented by introducing an additional filter.
appears to be a better alternative than MP 7 and the condensate pH for efficient metal removal should be above 5, based on the recommendations of the manufacturer. Also, the lab-scale experiments have shown that pH should not be too low (<4) for an efficient metal removal. At pH equal to 6 the removal of metals with TMT 15 works, but the optimal pH would be around 8.
S
UMMARYIn S
WEDISHEldning av returträ kräver extra åtgärder för att förhindra utsläpp av toxiska ämnen. Returträ är oftast kontaminerat med tungmetaller; vissa kan ursprungligen komma från kromaterad koppararsenat (CCA), vilken också kan höja arsenikkoncentrationen i rökgaskondensatet. På Idbäckens kraftvärmeverk eldas returträ, vilket leder till att giftiga ämnen hamnar i rökgasen. Gaserna kondenseras och de giftiga ämnena överförs till kondensatvattnet. Liksom andra giftiga ämnen måste även arsenik (As) avskiljas från kondensatvattnet innan det behandlade vattnet kan släppas ut till recipient. Olika tekniker för att avskilja As finns tillgängliga; utfällning med järn(III)klorid (FeCl3) är en metod, vilken
används idag på Idbäckens kraftvärmeverk. Metoden behöver dock optimeras för att försäkra att utsläppsgränsen för As inte överskrids i framtiden.
As återfinns vanligtvis som arsenat (As(V)) eller arsenit (As(III)), där den senare är svårare att avskilja från vatten. Experiment har visat att kondensatvattnet på Idbäcken främst innehåller As(V), vilket kan avskiljas med hjälp av utfällning med FeCl3. I kontakt med vatten bildar
FeCl3 ferrihydrit, ett ämne med hög specifik area. Den höga specifika
arean hos ferrihydrit bidrar till en hög adsorptionskapacitet för As. As-avskiljningen är beroende av många faktorer; pH och järnkloriddos är de viktigaste. Bra materietransport tros också påverka avskiljningen, men innefattas inte i det här arbetet. Flertalet experiment har gjorts på laboratoriet med målet att finna optimala förhållanden för en effektiv As-avskiljning från kondensatvattnet med hjälp av ferrihydritutfällning. Både pH och FeCl3-dos har varierats och olika As-koncentrationer har
använts. Effektiv avskiljning av As kräver ett pH mellan 3-8, där pH-värden i den lägre delen av intervallet har visat sig ge effektivast avskiljning. Enligt litteraturstudier sker maximal As(V)-avskiljning vid pH = 4, vilket är enligt resultaten i det här arbetet. En höjning av FeCl3
-dosen medför en bättre As-avskiljning upp till en viss gräns (ungefär 48 mol Fe3+/mol As); högre doser medför alltså inte någon effektivare
avskiljning av As. För att hitta minsta FeCl3-dos och för att bekräfta
labbresultaten på fullskalenivå på Idbäckens kraftvärmeverk krävs fortsatta efterforskningar och försök.
Ett fullskaleförsök har gjorts på Idbäckens kraftvärmeverk; tyvärr kunde inte några slutsatser dras från experimentet gällande As-avskiljningens effektivitet. Dock visar mätningar att pH sjönk avsevärt under försöket då FeCl3 tillsattes (vilket också har setts under labbförsöken). Alltså finns
en stor risk att pH sjunker under 3, vilket är lägre än rekommenderat. pH-mätningar och justering av pH efter tillsatsen av FeCl3
rekommenderas därför starkt. Vidare visade även fullskaleförsöket att den höga FeCl3-dosen som behövs vid höga As-koncentrationer kan
orsaka igensättningar i filtrena. Det skulle kunna förhindras genom att sätta in ett extra filtersteg.
Avskiljning av metaller med hjälp av komplexbindare som tillsatts till kondensatvattnet har också studerats. De två kemikalierna TMT 15 och MP 7 har jämförts för att se vilken av dessa som är mest effektiv för metallavskiljning och för att se om någon av dessa påverkas av eller påverkar FeCl3-tillsatsen. Dessa kemikalier används för avskiljning av Cd,
avskiljning. Vid pH = 6 fungerar avskiljningen av metaller bra, men optimal avskiljning fås vid ett pH kring 8.
Rekommendationer efter den här studien är att mäta pH efter FeCl3
-tillsatsen och att ha ett pH-höjande steg innan sandfiltret. pH på kondensatvattnet bör ligga runt 4 (eftersom optimalt pH = 4 enligt litteraturstudier och pH enligt labbexperiment ska ligga mellan 3-8). Vad gäller metallavskiljningen bör pH ligga över 5, helst ännu högre. Ett extra filtersteg är att rekommendera utöver de befintliga filter som finns på Idbäcken, detta för att inte igensättning ska ske vid höga FeCl3-doser
A
CKNOWLEDGEMENTT
ABLE OFC
ONTENTSummary ii
Summary In Swedish iv
Acknowledgement vi
Table of Content viii
1. Introduction 1
1.1. Aims and objectives 2
1.2. The Idbäcken combined heat and power plant 2 1.3. Waste wood and pressure treated wood 4
1.4. Arsenic 4
1.5. Removal of arsenic from water 5
1.6. Iron 5
1.7. Precipitation and adsorption with ferrihydrite 6 1.8. Parameters that effect the As removal precipitation process 6
1.8.1. Competitive ions/COD 7
1.8.2. pH 8
1.8.3. Dosage of FeCl3 8
1.8.4. Species with beneficial effect 9
1.8.5. Other parameters 9
1.9. TMT 15 and MP 7 9
1.10. Literature studies, experiments, and flue gas condensate 10
2. Methodology 11 2.1. Field sampling 11 2.2. Analysis 11 2.3. Chemicals 12 2.4. Lab-scale experiment 12 2.4.1. Arsenic-speciation 12
2.4.2. Variation of pH and different [Fe3+] 13
2.4.3. Variation of [As] at initial pH 7.5 13
2.4.4. Addition of phosphate 13
2.4.5. Addition of sulphate 15
2.4.6. Addition of TMT 15 or MP 7 15
3. Results and discussion 16
3.1. Lab-scale results and discussion 17
3.1.1. Arsenic-speciation 17
3.1.2. Variation of pH and [Fe3+] 17
3.1.3. Variation of [As] at initial pH 7.5 19
3.1.4. Addition of phosphate 21
3.1.5. Addition of sulphate 21
3.1.6. Addition of TMT 15 and MP 7 21
3.2. Full-scale results and discussion 27
3.3. Plant measurements 27 4. Conclusion 28 5. Recommendations 28 6. Further studies 30 References 32 Other references 32 Appendix I, Calculations I
A
BSTRACTAt the Idbäcken combined heat and power (CHP) plant waste wood is combusted. The flue gas from the combustion is condensed and heavy metals and other toxic species ends up in the condensate water. Since the condensate water in this way contains many toxic substances it needs to be treated before it is sent to the recipient Kilaån (the Baltic Sea). Arsenic (As) is one substance that needs to be removed, and this thesis aims to find the optimal conditions for As removal by using precipitation with iron(III) chloride (FeCl3).
When FeCl3 comes in contact with water it forms ferrihydrite, an efficient sorbent due
to its high specific area. The adsorption of As to ferrihydrite is dependent on different variables; pH, Fe(III) dosage, competitive ions et cetera. Lab-scale experiments have shown that a pH value between 3 and 8 is required for efficient As removal. Concerning the Fe(III) dosage further experiments are needed in order to tell what dosage that is the optimal in this case.
Further, the removal of metals has also been studied. A comparison between two chemicals (TMT 15 and MP 7) used for metal removal has been made, which showed that TMT 15 is to prefer for metal removal. Also, it was found that efficient removal of metals require pH > 5, preferably in the slightly alkaline pH range.
Key words: Waste wood; Pressure-treated wood; Flue gas condensate; Arsenic removal by precipitation; Ferrihydrite, TMT 15
1. Introduction
There is a potential economic benefit for co-firing waste, such as waste wood, in a bio-fuel power plant. However, waste wood contains more contaminants compared to virgin bio-fuel. Some contaminants can originate from chromated copper arsenate (CCA), which is used as a wood preservative to protect the wood from fungi and insects. When waste wood is incinerated the flue gas and/or flue gas condensate are most likely to be contaminated with heavy metals and other toxic substances from the waste wood material and a co-firing combined heat and power (CHP) plant using As-contaminated waste wood products with a flue gas condenser would likely need an efficient flue gas condensate treatment in order to remove As from the condensate. At the Idbäcken CHP plant waste wood is fired and even though the arsenic (As) concentrations in the outgoing condensate water is below existing limits (150 µg As/L), observed As concentrations have often been close to the limits. Therefore an improved solution for As removal is needed in order to secure the use of 100 % waste wood as a fuel mix in the long term. Equipping plants with textile filters combined with activated carbon is an effective flue gas cleaning method but it is very costly and therefore hard to defend for older plants. It would be desirable to find a simple method that can be implemented without any major redesign of the current condensate treatment unit in order to save costs. One such method could be precipitation of As from the flue gas condensate by adding an ordinary industrial chemical, such as FeCl3. This
method is currently used at the Idbäcken CHP plant, however, the method is not optimized and there is little knowledge about its removal efficiency for As and the impact of other constituents.
This thesis aims to find optimal conditions for As removal by precipitation with FeCl3. The process of finding the optimal conditions
recommendations were given concerning desirable conditions to reach a highly efficient As removal. The Idbäcken CHP plant has been chosen for this diploma work, since waste wood is fired there and there is a need for optimizing the condensate water treatment with respect to As removal. Also, an idea with this thesis was that other similar plants firing waste wood can use the same results and optimize their treatment as well.
1.1.
Aims and objectives
• To find optimal conditions for arsenic removal by precipitation (using FeCl3 as precipitation chemical)
• To investigate factors affecting the arsenic removal
• To find out whether the parameters affecting the arsenic removal also would affect the removal of other heavy metals (this is important to keep in mind when optimizing the arsenic removal process in order to avoid release of other heavy metals)
• To suggest a suitable method for the removal of arsenic from flue gas condensate at Idbäcken CHP
• To provide an overview of arsenic removal techniques in general An overview of As removal techniques was acquired through a literature review. Further, optimal conditions and factors affecting the removal of As were investigated by literature studies and lab-scale experiments. A comparison between metal removal agents and how they affect the As removal was also carried out. A suitable method for As removal at the Idbäcken CHP plant is suggested as a result of analysis of lab-scale and plant-scale experiments.
The research described in this thesis was carried out in collaboration with KTH (Royal Institute of Technology), Department of Land and Water Resources Engineering, on behalf of Vattenfall.
1.2. The Idbäcken combined heat and power plant
At the Idbäcken CHP plant almost 100 % waste wood is fired in boiler 3 and the flue gas formed has to be cleaned from particles, heavy metals, and other water-soluble substances such as HCl, NH3 and SO2. The first
step in the cleaning is an electro filter, which removes particles from the flue gas. After that the gas reaches a condenser so that it condensates. In the condenser the flue gas is washed with pH regulated water (pH 7-9) in order to eliminate as much acidic flue gas components as possible (HCl, HF, and SO2). Particles, heavy metals and ammonium are separated from
the flue gas and will end up as contaminants in the flue gas condensate. The process scheme for condensate water treatment can be seen in figure 1.
A membrane technique is used in order to remove ammonium, but the condensate must be free from particles before it passes by the membrane and therefore the condensate water is first passed through a sand filter (continuous), ultra filter and finally a softening filter. The condensate will also go through a cation-exchanger, where cations will be removed. However, anions such as As(V) and Cr(VI) will not be removed there. A TMT solution (tri-sodium salt solution) is added before the sand filter in order to reduce the concentrations of some heavy metal cations, mainly mercury and cadmium, by forming complexes that can be removed by filtration in the sand filter. FeCl3 is added both before the sand filter and
the sand filter. However, the addition of FeCl3 in the lamella separator is
used to increase the flocculation, not for As removal, as before the sand filter. The sludge that is formed is separated by a lamella separator and the cleaned condensate water is desalinated by reversed osmosis (2 m3/h) or sent to the recipient Kilaån.
Since 2007-2008 the dissolved As concentrations have been occasionally elevated (about once a year) in the flue gas condensate at the Idbäcken CHP plant and, during the spring of 2010, pressure-treated wood (about 400 ton) was test-fired. It resulted in high As concentrations in the condensate. Since the As concentrations also had been occasionally elevated before the test-firing of pressure-treated wood it was decided to take daily samples during the autumn of 2010 in order to see how the As concentration varied in the condensate. It could be established, after about one hundred daily samples were taken, that the margins were unsatisfactory small even when only waste wood was fired. Since November 2010 FeCl3 is added to the condensate and the result shows
that the A limits are not exceeded. However, the conditions for an optimal As removal with FeCl3 are not known at the Idbäcken CHP
plant, and therefore these conditions need to be studied, which is also the aim for this diploma work.
Table 1. Emission limit values (Naturvårdsverket, 2002).
Since the flue gas condensate contains toxic species that affects the environment there are regulatory standards set for the outgoing treated condensate by the Swedish Environmental Protection Agency (Naturvårdsverket) and by the Counties Agency (Länsstyrelsen). The limit values can be seen in Table 1.
1.3. Waste wood and pressure treated wood
At the Idbäcken CHP plant biofuel, mostly waste wood and wood chips, is fired. Less than 1 % fuel oil is burned of the total amount of fuel combusted. The fuel oil is only used when the boiler is started or during boiler malfunctions. All biofuel that is transported and fired is quality tested; however, the analysis takes time and it takes between 3 to 7 weeks to get the results. Therefore it is not possible to sort out As-containing fuel before it is fired. The fuel arriving to the Idbäcken CHP plant is inspected by staff at the plant as time permits. The driver of the fuel delivery truck takes a fuel sample from the load. The sample is taken by means of a 5-liter pail. Half amount of the sample is for gravimetric moisture content determination. Further, one quarter of the sample is for a month composite sample for each supplier and the last quarter of the sample goes to a total composite sample for the current month. Waste wood from demolitions is not allowed to be put on landfills; therefore the only way to take care of the waste wood is to combust it. The Idbäcken CHP plant buys all of the waste wood that is locally available; however, it is not enough to fill their demand for waste wood. Therefore waste wood is also bought from other parts in Sweden, as well as from other countries such as Norway and Great Britain.
Chromated copper arsenate (CCA) treated wood is considered hazardous waste and contains diarsenic pentaoxide (As2O5), chromium trioxide
(CrO3), and copper(II) oxide (CuO). Since As and Cr(VI) are
carcinogenic, the sum of the As and Cr(VI) content in wood must be less than 0.1 % (1000 mg/kg) in order to classify the wood as non-hazardous; for CCA treated wood the As content is typically around 0.2 % (2000 mg/kg) (Sundqvist et al., 2009). A plant that burns CCA treated wood must have a more complex flue gas treatment than other plants burning other types of non-hazardous waste. This is mostly because of the CCA treated wood contains As, which is difficult to remove from the flue gas since it is gaseous at fairly low temperatures. Although CCA-treated wood is generally not fired at the Idbäcken CHP plant, it may happen that CCA-treated wood arrives mixed in with the usual waste wood delivery and therefore is fired anyway. Thereby, it is important with an effective As removal at the Idbäcken CHP plant.
1.4. Arsenic
Arsenic occurs in five different valence states; +5, +3, +1, 0, and -3, which of the +5 state (arsenate or As(V)) and +3 state (arsenite or As(III)) are the most important valence states for dissolved As in water (Fetter, 1999). Dissolved As can be adsorbed by ferric hydroxides - As(V) more strongly than As(III) (Fetter, 1999).
Dissociation constants for As(V) are; pK1=2.22 (for Eq. 1), pK2=6.96
(for Eq. 2), and pK3=11.5 (for Eq. 3), which means that, for example, at
pK2, when pH is equal to 6.69, 50 % of the As(V) exists as a monovalent
(1) 𝐻3𝐴𝑠𝑂4↔ 𝐻2𝐴𝑠𝑂4−+ 𝐻+
(2) 𝐻2𝐴𝑠𝑂4−↔ 𝐻𝐴𝑠𝑂42−+ 𝐻+
(3) 𝐻𝐴𝑠𝑂42−↔ 𝐴𝑠𝑂43−+ 𝐻+
As(V) and As(III) can form surface complexes with several different oxides; Fe, Al, Mn, and Tl oxides for example (Gustafsson & Bhattacharya, 2007). Further, it is widely recognized that As(V) and As(III) form strong inner-sphere complexes on Fe and Al oxides (Gustafsson & Bhattacharya, 2007).
An important aspect for this thesis was to find out if the condensate contains As(V), As(III), or both, since the removal of As(V) differs from the removal of As(III).
1.5. Removal of arsenic from water
There are different ways of eliminating As from water. Most literature discusses the removal of As from ground water, which differs from the flue gas condensate studied in this work in respect to concentration and composition. Therefore the parameters affecting precipitation with metal salts can be somewhat different in the case of flue gas condensate than those mentioned in the literature review.
Different techniques can be used in order to remove As compounds from water. The most common is to use precipitation with metal salts. In this work FeCl3 was studied as the precipitation chemical. Paragraph 1.7
discusses As removal by the use of ferrihydrite.
Another method that can be used is activated alumina (AA), which can be suitable for small facilities with acidic to neutral raw waters and As(V) (Driehaus, 200-). Ion exchange is another method, however, this method takes a lot of costly maintenance work for regeneration and is not selective to As (Driehaus, 200-). A third method worth mentioning is a technique with iron oxide based adsorbents, which probably will play an important role in As removal in the future (Driehaus, 200-). This technique does not require regenerations as frequently as the other two methods just mentioned, and the breakthrough of As in the treated water is slow. The operational costs are equal to those of a coagulation plant; for small facilities this method could be the most economical solution (Driehaus, 200-).
1.6. Iron
Rapid hydrolysis of Fe(III) solutions results in the initial precipitate called ferrihydrite (Schwertmann & Cornell, 2000). Ferrihydrite plays an important role as an active sorbent in natural environments, due to its very high surface area (Schwertmann & Cornell, 2000).
Ferrihydrite can be described as an amphoteric ion exchange medium; as pH conditions change it has the capacity to offer hydrogen or hydroxyl ions (H+ or OH-) for cation and anion respectively. The adsorption
capacity is roughly about 0.5 mmol ionic material/g adsorbent (Vance, 2008). The ionic material includes As, Se, Cr, heavy metal cations, and organic anions or ligands.
1.7. Precipitation and adsorption with ferrihydrite
A method that can decrease the As concentration is the conventional coagulation-flocculation method (Han et al., 1994; Gregor, 2001; Wickramasinghe et al., 2004 cited in Laky & Licskó, 2011). There are three main steps in this process of As removal and these are (Han et al., 1994; Floch & Hideg, 2004; Wickramasinghe et al., 2004 cited in Laky & Licskó, 2011):
1. Oxidation of As(III) → As(V) (this step is only needed if As(III) is present)
2. Addition of a coagulant chemical (for example Fe(III) or Al(III) salt) in order to convert soluble As(V) to insoluble
3. Sedimentation, rapid sand filtration or micro filtration so that the solid particles can be removed
The coagulation method, followed by filtration, is one of the most conventional methods, suitable for large plants and coagulants used are based on ferric or ferrous salts; typical Fe dosages are 1-10 mg/L (Driehaus, 200-).
Pierce and Moore (1982) found that the oxidation state is important for the adsorption of As; As(V) adsorbs to a greater extent than As(III). Ferrihydrite is believed to be one of the most effective adsorbents of minor elements because of its reactivity and large specific surface area (Davis et al., 1993). Under aqueous conditions colloidal ferrihydrite particles can form large floccules or aggregates of particles (Bottero et al., 1991; Tchoubar et al., 1991 cited in Davis et al., 1993). As(V) adsorbs on the ferrihydrite surfaces by forming an inner sphere bidentate (bridging) complex, sharing apical oxygens of two adjacent edge-sharing Fe-oxyhydroxyl octahedral (Davis et al., 1993). Monodentate complexes have also been observed (Davis et al., 1993). Moreover, ferric hydroxides are stable over a wide Eh-pH range and therefore the mobility of As is limited (Fetter, 1999; Le Guern et al., 2007).
The rates of adsorption of As(V) on ferrihydrite have been established by Pierce and Moore (1982); the adsorption was fast with 90 % completion after 1 hour of stirring at pH 4, 8 and 9.9. Further, the rates were faster at an As concentration of 13.3 µmol/L than 0.667 µmol/L. 99 % of maximum adsorption had taken place after 4 hours of stirring for both As(III) and As(V). However, for equilibrium to occur 24 hours was adequate time. Usually, adsorption due to electrostatic processes is very rapid (in the order of seconds), but the adsorption of As is more slow, in the order of hours, indicating that a specific adsorption or formation of a chemical bond between the As and the adsorption site takes place (Pierce & Moore, 1982). Experimental work done by Pierce and Moore (1982) shows that the adsorption of As(III) increases to a maximum around pH 7, however, above this pH value the adsorption of As(III) decreases. Further, the pH dependence is greater at higher As(V) concentrations (Pierce & Moore, 1982).
In experiments done by Pierce and Moore (1982) more than 50 % of the As was adsorbed on ferrihydrite before a sample could be taken, which implies that a significant amount of As would be adsorbed as soon as the As species come in contact with the surface of the oxide.
1.8. Parameters that effect the As removal precipitation process
Many different parameters affect the As removal by precipitation from water. Below the most important parameters are discussed; competitive ions, pH, and dosage of FeCl3. Species with beneficial effect can alsopreviously been investigated for As removal from flue gas condensate. However, one can assume that these parameters will affect the removal of As from condensate water in the same way as from water, but to another extent.
1.8.1. Competitive ions/COD
There are several species that can affect the removal of As – phosphate, sulphate, silica, organic ligands, calcium, and magnesium (Le Guern et al., 2007). These ions will adsorb on the sites of metal oxide surfaces and in this way interact with As (Stachowicz et al., 2007).
Phosphate
Phosphate has a strong affinity for goethite, ferrihydrite and other iron oxides (Stachowicz et al., 2007). Further, phosphate is highly adsorptive on iron oxide surfaces, close to that of As(V) (Pierce, 1981 cited in Pierce & Moore, 1982). The effect of phosphate was evident at all pH values (7.5 – 8.0) and coagulant dosages (0.84–3 mg/L Fe) tested by Laky and Licskó (2011); phosphate has a greater effect the lower dosage coagulant used and the higher the pH value is. However, experiments made by Stachowicz et al. (2007) showed that As(V) is very sensitive to the presence of elevated concentrations of phosphate and that this effect is most distinct at lower pH, but still clearly visible throughout the whole pH range. The phosphate ions are relatively important at the surface of iron oxides even at very low concentrations and is therefore a competitor to As, and the adsorption ratio P/As will therefore be high (Stachowicz et al., 2007). Both the structure and the sorption behavior of phosphate and As(V) are similar (Dzombak & Morel, 1990 cited in Laky & Licskó, 2011), which results in a competition between them for the adsorption sites on ferric hydroxide (Holm, 2002; Manning & Goldberg, 1996 cited in Laky & Licskó, 2011). Laky and Licskó (2011) found that if phosphate phosphorous (PO4-P) is present in a system at the concentration 0.3
mg/L less As(V) is adsorbed compared to a phosphate free system. However, Pierce and Moore (1982) found that the adsorption efficiency of As was very little affected when phosphate and sulfate were added after the adsorption of As(III)/As(V) had taken place. The negative effect on the adsorption of As at low concentration was significant after phosphate or sulfate had already been adsorbed (Pierce & Moore, 1982). Silicate
When silicate is present in a solution it inhibits the removal of As with ferric hydroxides (Liu et al., 2007; Meng et al., 2000; Swedlund & Webster, 1999 cited in Laky & Licskó, 2011). Liu et al. (2007 cited in Laky & Licskó, 2011) observed that the ζ potential decreases when silicate adsorbs and form surface complexes on Fe(III) hydroxides at elevated pH values. The decreased ζ potential inhibits the particle agglomeration of ferric hydroxide because of increased repulsive forces between the Fe(III) hydroxide precipitates. Since the particle agglomeration fails the small Fe(III) hydroxide colloids can pass through a 0.2 µm pore-sized membrane (that is they cannot be separated) and furthermore the As(V) is not able to associate with the small flocks since the association of silicate with Fe(III) hydroxides results in a reduction in available adsorption sites for As(V) (Liu et al. 2007 cited in Laky & Licskó, 2011).
α-FeOOH (goethite) decrease in size and above concentrations of 0.5 and 1.0 mol-% for silicate and phosphate, respectively, only very fine particles are formed.
Carbonates
Metal hydroxide formation and pH are two parameters affected by the presence of inorganic carbon (H2CO3, HCO3-, and CO32-). Carbonate
and bicarbonate are important for the coagulation process since hydroxides are formed from iron coagulants due to hydrolysis if the buffering capacity of the water is high enough (Stumm & Morgan, 1962 cited in Laky & Licskó, 2011). A high buffering capacity completes the transformation of the metal ions to metal hydroxides (Licskó, 1990 cited in Laky & Licskó, 2011). However, the precipitation of carbonates leads to an increase of the pH, which may lead to desorption of As (Le Guern et al., 2007). Laky and Licskó (2011) also show in their work that high inorganic carbon concentration has an indirect negative effect by affecting the final pH, which leads to a higher remaining As concentration.
An adverse effect of inorganic carbon on As removal due to the competition between carbonate, bicarbonate and As for the adsorption sites on the metal hydroxides has been demonstrated by Holm (2002 cited in Laky & Licskó, 2011) when the concentration of inorganic carbon is 12 mg/L and 120 mg/L (1 and 10 mmol/L).
Natural organic matter
A considerable negative effect on the As removal process by natural organic matter has been observed by Kelemen (1991 cited in Laky & Licskó, 2011) and Laky and Licskó (2007 cited in Laky & Licskó, 2011).
1.8.2. pH
In an oxidizing environment with pH above 4.09 colloidal ferrihydrite will be found and As will be adsorbed, that is little As will be found in solution (Fetter, 1999). In environments that are mildly reducing and lack hydrogen sulfide mostly mobile As would be found, since iron would be in the soluble ferrous state and As would be in the As(III) form (Hounslow, 1980 cited in Fetter, 1999). When removing As better removal efficiency can be achieved with a lower pH value; the effect of silicate is more significant at high pH values for example (Laky & Licskó, 2011). Pierce and Moore (1982); Fuller et al., (1993 cited in Le Guern et al., 2007); Lumsdon et al., (2001 cited in Le Guern et al., 2007) state that the adsorption of As(V) is more efficient in a lower pH-range and that the uptake of As(V) decreases with increasing pH when adsorbed to iron hydroxide.
The surface of the adsorbent has a primarily net negative charge due to the adsorption of anions, so when pH increases there should be a decrease in the adsorption of anions (Pierce & Moore, 1982). Also, at low pH values the As(V) adsorption to ferrihydrite is independent of the pH, and the certain point where adsorption starts to decrease shifts towards lower pH values as the initial As concentration is increased (Pierce & Moore, 1982). Further, Pierce and Moore (1982) found that pH=7 for As(III) and pH=4 for As(V) are the pH values at which maximum As removal occur for concentrations normally found in nature.
1.8.3. Dosage of FeCl3
results in a greater number of adsorption sites for As (Laky & Licskó, 2011).
The dosage of FeCl3 is an important parameter when it comes to As
removal; better removal efficiency can be achieved with a higher coagulant dose (Laky & Licskó, 2011). Driehause (200-) estimates the minimum dosage of Fe(III) to be 10-20 times the As concentration at pH values between 6-7 and between 40-50 times the As concentration for pH values above 8. Further, an increased dose of 80 % Fe(III) was needed to achieve the target As concentration when the concentration of PO4-P was increased from 0.17 mg/L to 0.3-0.4 mg/L (Laky & Licskó,
2011). Laky and Licskó (2011) also found that the dose of iron needed to reach the target As concentration (10 µg/L) in a system with a silicate concentration of 14.23 mg/L was 2.5-3.5 times higher compared to a silicate free system.
Also, since particulates are formed when FeCl3 is added to water, more
particulates will be formed with a higher Fe(III) dosage, which can create clogging problems in the ultra-filters.
1.8.4. Species with beneficial effect
The adsorption of negatively charged ions like As(V) may be promoted by positively charged ions like calcium and magnesium (Stachowicz et al., 2007). Bivalent cations, like calcium, enhance the adsorption of As to ferrihydrite in the range of pH where cations show two positive charges (Le Guern et al., 2007). The bivalent cations seem to act as an electrostatic bridge between the solid surface of ferrihydrite and the anions of As (Le Guern et al., 2007).
1.8.5. Other parameters
The filtration process, the retention time, and the mass transfer are other parameters that could affect the As removal by precipitation The filtration process is important since particles containing As need to be removed. If the filters are not efficient, then there is a risk that As adsorbed to the particles end up in the recipient. Further, if, for example, the sand filter is not efficient, then it might lead to clogging in the ultra-filters. The retention time might not be as important as the filtration process, since the adsorption of As to ferrihydrite is fast, but if pH is too low a longer retention time might lead to the desorption of As. However, how these parameters affect the As removal is not investigated in this thesis.
1.9. TMT 15 and MP 7
TMT 15 is a 15 % aqueous solution of the salt trimercapto-s-triazine. It is used to precipitate metal ions – Cd, Cu, Pb, Hg, and Ni – present in process water. The solution is effective over a wide pH range, however, a neutral to slightly alkaline pH value is to prefer. Metal ions form together with TMT 15 stable, almost insoluble compounds and the precipitate can easily be flocculated and filtered in order to remove the metal ions from
the process water. The active substance in TMT 15 is announced as Best Available Technique (BAT) for waste combustion by the European Commission (Peldszus, 2011). TMT 15 is an almost odorless liquid with a pH ≈ 12.3 (Evonik Industries AG, 2012b). The structural formula is shown in figure 2. An average dosage of 50 mL TMT 15 /m3 (0.05
mL/L) is appropriate according to Evonik Industries AG (Evonik Industries AG, 2010). The removal of Hg is most efficient at a pH value above 5 (see Fig. 3) (Evonik Industries AG, 2011).
Amersep MP 7 is also used to precipitate metal ions and is a 38 % liquid water-based polythiocarbonate with an active CS3 molecule (ICIS, 1991).
30-45 minutes are needed for a complete reaction and the separation of metal ions is dependent on pH, reaction time, the concentration of metals in solution, what kind of flocculent chemical that is used et cetera (ICIS, 1991).
1.10. Literature studies, experiments, and flue gas condensate
As mentioned before, most of the literature available considers ground water, not flue gas condensate. Hence, this has to be taken into account while applying this information for the case of the treatment of flue gas condensate. The composition between ground water and condensate water differ, also the pH before FeCl3 addition can be assumed to bedifferent. Further, FeCl3 is used in the experiments in this study, while
literature sometimes considers other types of As-adsorbing iron materials. However, the tendencies in adsorption behavior and pH dependence et cetera are expected to be more or less the same between ground water and condensate water. Another parameter worth mentioning is the addition of As(V) when the lab scale experiments were carried out. In the flue gas condensate other As-species than As(V) can be present, and it is possible that these species adsorb weaker to ferrihydrite than As(V). If this is the case, then a higher Fe(III) dosage would probably be needed for full-scale experiments and operation than what was needed for the lab-scale experiments, since a higher Fe(III) dosage would provide a higher amount of adsorption sites for the As-species and thereby increase the As removal.
2. Methodology
The first thing that was done before the lab experiments was the collection of flue gas condensate samples from the Idbäcken CHP plant. After that lab experiments were carried out; FeCl3 was added to samples
of the flue gas condensate in order to remove As from the condensate. Different parameters, e.g. pH and FeCl3 dosage, were varied to see how
they affected the As removal. Calculations made are shown in Appendix I. After the lab-scale experiments were done and the results analyzed a full-scale experiment was carried out at the Idbäcken CHP plant.
2.1. Field sampling
Flue gas condensate water has been taken from the Idbäcken CHP plant. The samples used for the lab-scale experiments were taken in the flush tank when the boiler was operated with a power of 82 MW. This condensate was used in all the lab-scale experiments described under paragraph 2.4. For the condensate composition, see below in Table 2. The samples were stored in sealed plastic bottles in the refrigerator. The dosage of FeCl3 added at the Idbäcken CHP plant was measured at
the time for sampling and the Fe(III):As molar ratio was calculated to 255:1, which theoretically could remove between 480-1440 µg As/L (the maximum value of As that has ever been measured at the Idbäcken CHP plant before treatment was approximately 1000 µg As/L). Hence, if only waste wood is fired (that is, no CCA treated wood) and the As concentration remains low, the Fe(III) dosage could probably be lowered. The molar ratio 255:1 is equal to an addition of about 100 mL FeCl3/m3 flue gas condensate (when the As concentration is 113 µg/L
and the flue gas condensate flux is 5.5 m3/h).
Samples were also taken during the full-scale experiment discussed in paragraph 2.5, when pressure treated wood was mixed with the ordinary waste wood.
2.2. Analysis
atomic emission spectroscopy), AFS (atomic fluorescence spectrometry), and ICP-SFMS (inductively coupled plasma sector field mass spectrometry). It can also be mentioned that all results from analysis have a margin of error, which can be an explanation to, for example, why the S concentration is higher in the filtered sample than in the unfiltered sample in Table 2. The full analysis results are shown in Appendix II.
2.3. Chemicals
Chemicals used in the lab-scale experiments are as follows: • Fe(III) chloride (FeCl3)
• hydrated sodium arsenate (Na2HAsO4·7H2O) • sodium hydroxide (NaOH)
• hydrochloric acid (HCl) • sodium sulfate (Na2SO4)
• sodium phosphate (Na2HPO4·2H2O) • nitric acid (HNO3)
• trimercapto-s-triazine (TMT 15)
• polythiocarbonate solution (Amersep MP 7)
2.4. Lab-scale experiment
The experiments were in general carried out the same way, except for the As-speciation (see paragraph 2.4.1.). The condensate water used for all experiments was taken at the Idbäcken CHP plant at one single sample-taking described in paragraph 2.1. In all experiments, 30 mL of condensate water was added in centrifugal bottles. pH was measured with a pH-electrode. In some experiments pH was adjusted by adding small amounts of HCl or NaOH. However, even in cases when pH was adjusted to be constant throughout a set of experiments it could happen that pH differed marginally from sample to sample since it was difficult to adjust to exact values. The initial pH is the value before any addition of FeCl3 and final pH the value after FeCl3 addition. Since pH is lowered
when FeCl3 is added and the pH is important for an effective As
removal, it is the final pH that is important for the removal process. Therefore pH was adjusted after the addition of FeCl3 for most lab-scale
experiments. However, in some of the first experiments pH was only measured initially (before FeCl3 addition), and final pH was
approximated afterwards by preparing new samples according to the same procedure and measuring pH in those. This might have resulted in final pH values somewhat different from the actual values. Table 5 in paragraph 6 shows all measured pH values.
The same chemicals were used in all experiments. All samples were shaken for 24 hours and then centrifuged at 3000 rpm for 15 minutes. After that all samples were filtered with a 0.2 µm filter. The samples were then conserved by adding HNO3 so that the HNO3 concentration in the
samples was 0.1 M. Analyzes were made by ALS, see paragraph 2.2. The molar ratios of Fe(III):As used (between approximately 20:1 and 60:1) are based on literature studies.
2.4.1. Arsenic-speciation
µm filter), that is, As in solution was analyzed, but not As in particulates. The condensate water then ran through the cartridge. The samples were conserved with HNO3 so that the concentration in the samples was 0.1 M. Analysis was made at the University of Stockholm with an ICP-OES (inductively coupled plasma optical emission spectrometry).
In this way it could be determined that the flue gas condensate mostly consisted of As(V). It is important to know if the condensate water contains As(V) or As (III) since As(V) is adsorbed more easily than As(III) and at a different pH. Also, if As(III) is present it has to be oxidized to As(V) before the removal process.
2.4.2. Variation of pH and different [Fe3+]
In this experiment a total of nine samples were prepared and three different initial pH values were used; 7, 8 and 9. Further, three different Fe(III):As molar ratios were used; 24:1, 48:1, and 71:1. pH was measured before the addition of FeCl3. As concentration was 113 µg As/L (that is the initial As concentration in the condensate water). The experiments were carried out in order to see how the variation of pH and [Fe3+] affected the As removal.
Also, six experiments were made at fixed pH and with the addition of 10000 µg As(V)/L. Initial pH was adjusted to 9 after As(V) addition and thereafter FeCl3 was added to the samples with a molar ratio Fe(III):As equivalent to 20:1 and 40:1. After the FeCl3 addition pH was adjusted to 3, 5, and 7, respectively. These experiments were made to show the effect of pH after the addition of FeCl3. Table 3 shows a summary of the experiment parameters.
2.4.3. Variation of [As] at initial pH 7.5
A total of six samples were prepared and analyzed. An As(V) solution was added so that the concentrations in the samples were 1000, 5000 respectively 10000 µg As(V)/L higher than the initial concentration of 113 µg As/L. All samples were adjusted to an initial pH of 7.5. Thereafter FeCl3 was added to the samples in two different molar ratios; Fe(III):As equivalent to 54:1 and 81:1. The addition of FeCl3 lowered the pH and the final pH values can be seen in the summary of the experiment parameters shown in Table 4. These experiments were carried out so that the removal of As(V) could be studied at different As concentrations.
2.4.4. Addition of phosphate
In this set of experiments six samples were prepared. As(V) was added to a concentration corresponding to 1000 µg As(V)/L above the initial 113 µg As/L in the samples (total concentration = 1113 µg As/L). The initial P concentration was 541 µg/L and phosphate was then added to the samples with a molar ratio PO43−
Fe3+ equal to 0.2 and 0.4, respectively – that
Table 4. A summary of the experiment parameters described under paragraph 2.4.3. As addition, the molar ratio between Fe(III) and As, initial and final pH, and the remaining As concentration are shown. Remaining As concentration is the concentration after precipitation and subsequent filtration.
2.4.5. Addition of sulphate
Totally six samples were prepared in this set of experiments. Sulphate was added to the samples with a molar ratio 𝑆𝑂42−
𝐹𝑒3+ = 0.2 and 0.4
respectively, equivalent to 31 and 61.7 mg sulphate/L respectively. The added sulphate concentrations equals 0.4 respectively 0.75 times the initial S concentration of 82 mg/L. Also, As(V) was added so that the concentration in all samples was equivalent to 1000 µg As(V)/L above the initial 113 µg As/L. pH was adjusted to 7, 8 and 9 (initial pH values). FeCl3 was added in the molar ratio Fe(III):As equivalent to 60:1. These
experiments were carried out to see how sulphate affects the removal of As(V) by competing for the adsorption sites at the FeCl3. Table 6 shows
a summary of the experiment parameters.
2.4.6. Addition of TMT 15 or MP 7
In these experiments four to six samples were prepared for each set of experiments. A summary of the experiment parameters are shown in Table 7. The experiments were carried out in order to see which of the two precipitations agents that are the most effective for metal removal, and also to see if the addition of FeCl3 affects the metal removal in
general or if the As precipitation is affected by TMT 15 and MP 7. TMT 15 and MP 7 were added to flue gas condensate of elevated As concentrations (5113 and 10 113 µg As(V)/L) with varied pH. In the first two sets of experiments presented in Table 7, 5000 µg As(V)/L was added and the initial As concentration was 113 µg/L, resulting in a total As concentration of 5113 µg/L. Also, 0.03 mL/L TMT 15 or MP 7 was added to the samples. pH before the FeCl3 addition was adjusted
Table 5. A summary of the experiment parameters described under paragraph 2.4.4. As addition, the molar ratio between Fe(III) and As, initial and final pH, the addition of phosphate, and the remaining As concentration are shown. Remaining As concentration is the concentration after precipitation and subsequent filtration.
to 7, 8 or 9. FeCl3 was added in the molar ratio Fe(III):As equivalent to
40:1 respectively 60:1. Final pH after FeCl3 addition was measured for
the experiments with an addition of TMT 15 (see Table 7). The final pH values were measured after the experiments were done, that is, new samples with the same composition as the first ones were made and pH was measured for these samples. However, final pH for the MP 7 experiments was only measured for a few samples.
Further, the other two sets consisting of four experiments each were made with an addition of TMT 15 or MP 7 at fixed pH values. 10000 µg As(V)/L and 0.03 mL TMT 15/L was added to all samples and pH was set to 9 before the addition of FeCl3. The molar ratio Fe(III):As
was 40:1. The pH was adjusted again after the FeCl3 addition to 2, 4, 6,
and 8. The experiments made were able to show whether the metal removal was affected by the different final pH values, in a wider pH range than the experiments when only initial pH was controlled.
3. Results and discussion
displayed in figures. The results are tabulated in Table 3-7. Appendix II, Samples, gives a more detailed version of the results from the lab scale experiments (i.e. results for more elements than discussed in this report).
3.1. Lab-scale results and discussion
Below a discussion and presentation of the results from the lab-scale experiments follow. It is important to take into account that the final pH values might not be exact, since final pH was measured for different samples than the samples that were analyzed by ALS. However, the samples used for measuring the final pH values were prepared the same way as the analyzed samples. For the experiments with fixed pH, the measuring of final pH was made and thereafter the samples were sent for analysis. It can also be mentioned that since the concentrations of MP 7 (38 % aqueous solution) and TMT 15 (15 % aqueous solution) are different from each other, and it is two different substances, a comparison between these chemicals might not be accurate.
3.1.1. Arsenic-speciation
The results from the As-speciation showed that the condensate water mainly consisted of As(V) and not As(III). Hence, the removal of As with FeCl3 from the flue gas condensate should be fairly straight forward
and there is no need for oxidation from As(III) to As(V). An analysis of the unfiltered condensate water showed a concentration of 113 µg As/L, while after As-speciation the As concentration (As(III)) was analyzed to 7.45 µg/L, that is, most of the As in the condensate water is present as As(V) or As bound to particles (113 µg As/L – 7.45 µg As(III)/L = 105.55 µg As(V) (or As bound to particles)/L). An important result from the As-speciation is that a small amount of As(III) is present in the flue gas condensate. As adsorbed to particles will be removed by filtration.
3.1.2. Variation of pH and [Fe3+]
Variation of pH and [Fe3+] with no adjustment of pH after FeCl3- addition
From these experiments it can be seen that the As concentration decreases after the addition of FeCl3 up to a certain point where it is
leveling out, see figure 4. That is, the adsorption of As will not be improved by increasing the FeCl3 dosage above 48 mol Fe(III)/mol As.
The experiments show that the As removal is clearly affected by pH. figure 4 shows a less effective As removal at a high initial pH (9) before FeCl3 addition (final pH = 8.9). This is consistent with the literature,
which recommends pH 4-8 for As removal, expecting better removal in the lower end of this range. Initial pH 7 and 8 resulted in final pH (i.e. pH after FeCl3 addition) equal to 6.9 and 7.5 and resulted in a
As species as well. However, further studies are needed in order to find the optimal dosage at the Idbäcken CHP plant.
Variation of pH and [Fe3+] with adjustment of pH after FeCl3- addition
In the experiments with fixed pH (pH was adjusted after the addition of FeCl3) the remaining As concentration in the samples was <5 µg/L (i.e.
below detection limit). It can be seen in figure 5 that these experiments, where pH was adjusted, resulted in a better removal of As compared to the experiments where pH was not adjusted (for remaining As concentrations for the experiments see Table 3). The difference between these experiments, except from the adjustment of pH, is the addition of As(V). For these experiments with adjusted pH, 10000 µg As(V)/L was added to the samples, while no As was added to the other samples. It is unclear why the remaining As concentrations are lower for these examples compared to those with no added As(V). One explanation could be that the As present in the condensate water is not only As(V) and therefore the removal is not as efficient as it should be if only As(V) was present. However, when the FeCl3 dosage is increased (to keep the
same Fe:As ratio as in the other experiment with a lower As(V) concentration) more particulates are formed that can adsorb As and thereby the As present in the condensate will be removed even though it is not As(V). This could, as said, be one explanation, but there can be other explanations as well. A comparison (As removal efficiency (%)) between the experiments with or without adjusted pH can be seen in figure 5. Another explanation could be that at higher As(V) concentrations the As(V) ions are in excess over competitive ions (P for example) and thereby a higher percentage of As can be adsorbed. 0 20 40 60 80 100 120 0 20 40 60 80 Re m ai ni ng [ As ] ( µg/L )
Molar ratio Fe3+/As
Variation of [Fe
3+] at different pH values
pH 7 pH 8 pH 9
Figure 4. Variation of different [Fe3+]; 0.002, 0.004, and 0.006 g Fe3+/L,
equivalent to molar ratios (Fe(III):As) 24:1, 48:1, and 71:1 respectively. Initial As concentration is 113 µg/L and initial pH values are 7 (6.9), 8 (7.5)
A conclusion from these experiments can be drawn; an effective As removal is reached above pH 3 for As(V) additions up to 10000 µg/L and Fe(III):As molar ratios above 20:1 (this was the lowest FeCl3 dosage
tested, however, lower dosages might be enough for effective removal as well).
As earlier mentioned, not only As concentrations were analyzed for, but also other metals. These remaining metal concentrations are still quite low for all pH values (final pH of 3, 5, and 7); no limit values were exceeded. Since the final pH values are relatively high the metals will not be in solution to a great extent, thereby the metal concentrations will be low. When comparing to the condensate for which no FeCl3 had been
added, it was obvious that the addition of FeCl3 decreased the
concentrations of metals in solution, probably because the metals form metal hydroxides with the formed ferrihydrite.
3.1.3. Variation of [As] at initial pH 7.5
To investigate the degree of As removal at elevated concentrations (1113 µg As/L, 5113 µg As/L, 10113 µg As/L), As(V) was added to the condensate collected at Idbäcken CHP plant. pH was set to 7.5 before FeCl3 was added, since this is a typical pH for the condensate at
Idbäcken CHP plant. The results are presented in figure 6. The FeCl3
dosage was based on As concentrations and two different dosages were used; 54:1 and 81:1 (Fe(III):As molar ratios). Hence, the resulting FeCl3
concentration was higher for higher As concentrations. Since the addition of FeCl3 lowers pH, the final condensate pH decreased with
increasing As concentrations. Final pH for the samples with an addition of 1000 µg As(V)/L was 5.5 and 3.9 for the two molar ratios 54:1 and 81:1 respectively. According to literature this is within the pH range for good As(V) adsorption, which is confirmed by the experimental results shown in figure 6 (high removal of As is shown as a low remaining As Figure 5. A comparison between the experiments with a variation of pH and [Fe3+] and with or without an As(V) addition (i.e. As concentrations of 113 and 10113 µg As(V)/L)). Fe(III):As molar ratios is between 20:1 and 48:1. It can be seen that those experiments with adjusted pH have resulted in a more effective As removal (near 100 %), while the other experiments without adjusted pH or As(V) addition have an As removal efficiency as low as 60 %. Final pH 6.9 and 7.5 are less effective than adjusted pH 7, which could be explained by the addition of As(V) to the experiments with adjusted pH (and also a higher Fe(III) dosage) and that the initial 113 µg As/L was not only As(V), but also other As-species. 0 20 40 60 80 100 120 0 20 40 60 As r em ova l
Fe(III):As molar ratio
concentration). However, when the As(V) concentration is increased from 1000 to 5000 µg As(V)/L, the dosage of FeCl3 is also increased,
resulting in final pH of 2.7 and 2.6 for the molar ratios 54:1 and 81:1 respectively. At these pH values, the As-adsorption is expected to be less efficient, especially at pH 2.6, which is clearly shown in figure 6. Even though higher FeCl3 addition is normally beneficial for As-adsorption,
this effect is, in this case, outweighed by the decrease in pH. At even higher As concentration, 10000 µg added As(V)/L, pH after FeCl3
addition is even lower, 2.5 and 2.4 for the molar ratios 54:1 and 81:1 respectively. However, it should be emphasized that the adsorption of As still was good and below limit values for all experiments in figure 6. Figure 6. Variation of As(V) concentration at initial pH = 7.5. Final pH was measured to 5.5 and 3.9 (1113 µg As/L), 2.7 and 2.6 (5113 µg As/L), and 2.5 and 2.4 (10113 µg As/L) for the Fe:As molar ratios 54:1 and 81:1, respectively. The figure shows that a pH <2.6 results in a less effective As removal, even though the Fe(III) dosage is increased. However, when pH is 3.9 or higher the removal of As is more effective when the Fe(III) dosage is higher.
0 5 10 15 20 2,84 2,86 2,88 2,9 2,92 2,94 As co nce nt ra tio n r em ai ni ng ( µg/ L) Final pH
Phosphate addition
15,6 mg added phosphate/l 0 5 10 15 20 25 30 35 0 20 40 60 80 100 Re m ai ni ng [ As ] ( µg/L )Molar ratio Fe3+/As
Variation of [As] at pH 7.5
[As]=100 0 µg/l [As]=500 0 µg/l
Figure 7. Phosphate addition. Initial P concentration in samples was 541 µg/L and added phosphate was 15.6 respectively 31.2 mg/L. Initial As concentration was 113 µg/L and added As(V) was 1000 µg/L (i.e. total As concentration = 1113 µg/L). pH values was initially 7 respectively 8 resulting
in final pH values of 2.8 respectively 2.9. Added [Fe3+]=0.045 g/L, equal to a
3.1.4. Addition of phosphate
The effect of phosphate on As removal was investigated by adding phosphate to the flue gas condensate prior to the FeCl3. It can be seen in
figure 7 that the As removal decreases with a higher phosphate addition, however, the removal of As is still good. Since the added amount of phosphate in the samples was 30 respectively 60 times higher than the P concentration in the flue gas condensate used in the experiments, the conclusion would be that phosphate has an effect on As removal, but that it takes a much higher concentration of phosphate in the condensate in order to affect the removal to a significant extent. It can also be mentioned that it seems like less As is removed at higher pH values, however, the difference in remaining As concentration is quite small.
3.1.5. Addition of sulphate
The effect of sulphate on As removal was tested during these experiments. The results can be seen in figure 8. Since the S concentration in the flue gas condensate was as high as 82 mg/L initially, the added sulphate did not result in a significantly higher concentration in the final samples. Sulphate was only added in amounts equal to 0.4 respectively 0.75 times the initial S concentration. As in the case for phosphate addition it seems like the removal of As is less affected by the addition of sulphate at low pH values, but still these differences are small so it is difficult to say whether the effect of pH is significant or not, especially for a wider pH range. According to the experiments, the removal of As is good at all added sulphate concentrations. However, since the added sulphate did not result in a significantly higher S concentration in the samples, it cannot be excluded that sulphate could affect the As removal if added in higher concentrations.
3.1.6. Addition of TMT 15 and MP 7
Experiments with no adjustment of pH after FeCl3 addition
The results from the experiments with an addition of TMT 15 show that the As removal was effective for all pH values (initial pH values 7, 8, and 9 resulting in final pH between 2.3 and 5.61) and FeCl3 dosages (mol
Fe(III)/mol As equal to 40 and 60) tested, the remaining As 0 0,5 1 1,5 2 2,5 3 3,5 4 4,5 2,64 2,66 2,68 2,7 2,72 2,74 Re m ai ni ng [ As ] ( µg/L ) Final pH
Sulphate addition
31 mg added sulphate/ lFigure 8. Sulphate addition. Initial S concentration in samples was 82 mg/l and added sulphate was 31 respectively 61.7 mg/l. Initial As concentration was 113 µg/l and added As(V) was 1000 µg/l (i.e. total As concentration = 1113 µg/l). pH values was initially 7 respectively 8 resulting in final pH values of
2.6 respectively 2.7. Added [Fe3+]=0.045 g/l, equal to a Fe(III):As molar ratio
concentrations are shown in Table 7. These experiments showed a similar efficiency in As removal compared to the experiments with different [As] at pH 7.5 (see Table 4). However, it is difficult to compare these experiments since the final pH and FeCl3 dosages are not exactly
the same, but for example, the TMT 15 experiment with a final pH = 2.68 resulted in a remaining As concentration of 6.23 µg/L, while for the other experiment a final pH = 2.7 resulted in a remaining As concentration <7 µg/L (i.e. under detection limit) for Fe(III):As molar ratios 60:1 and 54:1 respectively. The difference in remaining As concentration is very small between the experiments and a conclusion that TMT 15 does not affect the As removal negatively can be drawn. Since final pH was around 2.7 for both sets of experiments and the As removal was efficient it can also be concluded that the As removal is effective at this pH value.
The experiments with the addition of MP 7 resulted in quite high remaining As concentrations; the highest value was about 98 µg As/L. The value was measured in a sample where final pH was not measured, however, if compared to the TMT 15 experiment with the same initial pH (7) and added Fe(III):As molar ratio (60:1) the final pH should be around 2.7, see Table 7. A comparison between the experiments where MP 7 and TMT 15 were used shows that the addition of MP 7 gives a higher remaining As concentration in all samples (for remaining concentrations see Table 7). It seems like the addition of MP 7 results in a less effective As removal, both compared to the experiments with TMT 15 and the experiments with varied [As] at initial pH 7.5 (MP 7 resulted in an As removal efficiency of 98 %, while the other experiments resulted in an efficiency >99.8 %). The remaining As concentration (98 µg/L) for the MP 7 experiment was below the emission limit value for As (150 µg/L), but since TMT 15 seems like a better alternative MP 7 is not to prefer for metal removal at the Idbäcken CHP plant.
Furthermore, the experiments for which both complex binders and FeCl3 were added were carried out in order to compare the efficiency in
metal removal between MP 7 and TMT 15. The results from these experiments show high concentrations of most metals present in the Table 8. A comparison between two samples with added TMT 15,
one sample without any added FeCl3, the other with an addition
of FeCl3 equal to a Fe(III):As molar ratio of 60:1. As
concentration in both samples was 5113 µg/l.. pH for the sample without TMT 15 was measured to about 9 and the sample with
added FeCl3 had a final pH of 3. It can be seen that the low pH
because of FeCl3 addition results in a higher remaining metal
concentration (except for Hg) compared to the other sample
samples. A comparison between a sample prepared with TMT 15, but without any FeCl3 addition, and the samples prepared with both TMT 15
and FeCl3 shows that the metal concentrations (Cd, Co, Cr, Cu, Ni, Pb,
and Zn) are much higher in the samples with FeCl3 (Table 8). The only
exception is the Hg concentration that was lower when FeCl3 was added.
An explanation to this can be the low final pH (2.3-5.61) that is the result of the FeCl3 addition. The low pH dissolves metal hydroxides, that is,
releasing metal ions into solution, which results in a higher concentration of free metal ions in the samples (Fig. 9). For a better metal removal a higher dosage of TMT 15 or MP 7 would be required if pH is not raised (Fig. 3).
Recommended pH value for an effective metal removal is, due to the manufactures of TMT 15, above 5 for Hg removal and neutral to slightly alkaline for removal of other metals. No optimal pH for MP 7 has been found, but since the two chemicals work in a similar way, it is reasonable to expect that MP 7 is less effective at low pH values (below 5) as well. Experiments with adjustment of pH after FeCl3 addition
The experiments with an addition of TMT 15 and MP 7 at fixed pH, i.e. pH is adjusted after FeCl3 addition, (tested pH values = 2, 4, 6, and 8)
show that at low pH values the metal removal is not as good as at higher pH values, see figure 10-13. The experiments with TMT 15 at pH 2 resulted in high concentrations of zinc and lead after precipitation and filtration, which can be seen in figure 10 (for Pb 1660 µg/L and for Zn 2210 µg/L). These concentrations are above the emission limit values (200 and 1500 µg/L for Pb and Zn respectively), which means that a pH value of 2 is too low for metal removal. The remaining Zn concentration is high even at pH 4 (1950 µg/L), but low at pH 6 (424 µg/L). Since no experiments have been carried out at pH 5 it is difficult to say whether it is good enough for the removal of Zn, so in order to secure a remaining Zn concentration below the emission limit value a higher pH than 5 could be used, for example pH 6 up to pH 8. Other metals – Cu, Cr, Cd, Co, and Ni – were present in the condensate, but the concentrations were below their emission limit values, since the concentrations in the flue gas condensate from start were below limit values (for limit values see Table 1). The concentrations of these metals increased as pH decreased (Fig. 11). Therefore, it is important to adjust the pH so that it is not too low in order to get a satisfying metal removal from the
condensate water. The remaining concentrations of Zn and Pb can be seen in figure 12 and those of Cu, Cr, Cd, Co, and Ni in figure 13. The addition of TMT 15 kept the Hg concentration low for all pH values tested (Fig 14).
It can also be mentioned that the precipitation of metal hydroxide to metal ions in solution is dependent on pH; in the lower pH range metal ions are present in solution, while at the higher pH range metals are precipitated as metal hydroxides (see Fig. 9). Hence, the experiments made at pH 2 and 4, and for some metals even at pH 6, show a higher metal concentration in solution than at pH 8. The higher concentration of metal ions at these low pH values (pH 2-6), can partly be explained by the dissolution of metal hydroxides, but also by the fact that TMT 15 removes metal ions less efficiently at pH below 5. The As removal was almost complete; however, as expected less As was removed at pH 2, where 17.6 µg As/L remained in solution (see Table 7). At pH 4, 6, and 8 the remaining As concentration was below detection limit.
For the experiments with an addition of MP 7 at fixed pH the removal of As was almost complete, except at pH 2, where 13 µg As/L remained in solution (this is still a satisfying result, but less effective compared to the other pH values), see Table 7. The most significant result from the MP 7 experiments can be seen in figure 14, where the Hg concentrations exceeded the emission limit (10 µg Hg/L) for pH 4 and 6; concentrations of 31 and 52 µg Hg/L were found.
Concerning As removal, both TMT 15 and MP 7 give good results for the tested pH values 4, 6, and 8. At pH 2 the removal of As is less effective; this is probably not because of TMT 15 or MP 7, but because of the low pH affecting the ferrihydrite formation and adsorption of As. Figure 10. Remaining concentrations of Zn and Pb after an addition of 0.03 mL TMT 15/L. Initial concentrations of Zn = 2590 µg/L (unfiltered) and Pb = 1520 µg/L (unfiltered). Total concentration of As = 10113 µg/L and final pH values was adjusted to 2, 4, 6, and 8. The molar ratio Fe(III):As was 40:1. Higher remaining concentrations are obtained in the lower pH-range.
0 500 1000 1500 2000 2500 0 2 4 6 8 10 Re m ai ni ng co nce nt ra tio n ( µg/L ) Final pH
Metal concentration after TMT 15 addition
0 100 200 300 400 500 0 2 4 6 8 10 Re m ai ni ng co nce nt ra tio n ( µg/L ) Final pH
Metal concentration after TMT 15 addition
Ni Cu Cr Co Cd
Figure 11. Remaining concentrations of Ni, Cu, Cr, Co, and Cd after the addition of 0.03 mL TMT 15/L. Initial concentrations (unfiltered) of Ni = 100 µg/L, Cu = 397 µg/L, Cr = 206 µg/L, and Cd = 11.5 µg/L. Total concentration of As = 10113 µg/L and final pH values was adjusted to 2, 4, 6, and 8. The molar ratio Fe(III):As was 40:1.
0 500 1000 1500 2000 2500 0 2 4 6 8 10 Re m ai ni ng co nce nt ra tio n ( µg/L ) Final pH
Metal concentration after MP 7 addition
Zn Pb
0 50 100 150 200 250 300 350 400 450 0 2 4 6 8 10 Re m ai ni ng co nce nt ra tio n ( µg/L ) Final pH
Metal concentration after MP 7 addition
Cr Co Cd Ni Cu
Figure 13. Remaining concentrations of Ni, Cu, Cr, Co, and Cd after the addition of 0.03 mL MP 7/L. Initial concentrations (unfiltered) of Ni = 100 µg/L, Cu = 397 µg/L, Cr = 206 µg/L, and Cd = 11.5 µg/L. Total concentration of As = 10113 µg/L and final pH values was adjusted to 2, 4, 6, and 8. 0 10 20 30 40 50 60 0 2 4 6 8 10 Re m ai ni ng Hg co nce nt ra tio n ( µg/ L) Final pH
Hg concentration after TMT 15/MP 7 addition
Hg (TMT 15) Hg (MP 7)
