当前位置:首页 >> 电力/水利 >>

Influence of dry and humid gaseous atmosphere on the thermal decomposition of calcium chloride and i_图文

Chemical Engineering and Processing 48 (2009) 380–388

Contents lists available at ScienceDirect

Chemical Engineering and Processing: Process Intensi?cation
journal homepage: www.elsevier.com/locate/cep

In?uence of dry and humid gaseous atmosphere on the thermal decomposition of calcium chloride and its impact on the remove of heavy metals by chlorination
¨ G. Fraissler a , M. Joller a,b,? , T. Brunner a,b,c , I. Obernberger a,b,c
a

Austrian Bioenergy Centre GmbH, Inffeldgasse 21b, 8010 Graz, Austria Graz University of Technology, Institute for Process Engineering, Inffeldgasse 21b, 8010 Graz, Austria c BIOS BIOENERGIESYSTEME GmbH, Inffeldgasse 21b, 8010 Graz, Austria
b

a r t i c l e

i n f o

a b s t r a c t
In the last years chlorination has become an important industrial technique for the removal of heavy metal impurities (especially cadmium, copper, lead and zinc) in raw materials. In order to estimate the chlorination effect of different chlorine donors on heavy metals in sewage sludge ashes and the in?uence of gas humidity, the thermal decomposition behaviour of the chlorine donors was investigated. Therefore, thermal analysis of calcium chloride between 20 ? C and 1400 ? C in synthetic dry and humid ?ue gas atmosphere simulating ?ring methane with air was conducted in this work. In the ?rst case the gas composition contained 12 v% oxygen, 5 v% carbon dioxide and 83 v% nitrogen whereas in the second case water vapour was added to the gas mixture in an amount to equal 10 v% of the total gas. In both cases (dry and humid ?ue gas), the ?rst mass losses of calcium chloride up to 260 ? C are based on the decomposition of its hydrates. Chlorine release occurs above the melting temperature of calcium chloride (782 ? C) only: in the ?rst case chlorine gas (Cl2 ) by reaction of gaseous calcium chloride with oxygen and in the second case hydrogen chloride (HCl) by reaction with water vapour are formed. These different reaction products may in?uence the chlorination effect on heavy metals and further their removal in industrial processes. The solid reaction product is in both cases calcium oxide. ? 2008 Elsevier B.V. All rights reserved.

Article history: Received 31 May 2007 Received in revised form 9 May 2008 Accepted 12 May 2008 Available online 27 May 2008 Keywords: Calcium chloride Thermal decomposition Chlorination effect Heavy metals Volatilisation

1. Introduction Ashes from sewage sludge combustion offer a high content of the element phosphorus and represent therefore a valuable fertilising resource. The main reasons why phosphorus is not exploited from sewage sludge ashes are heavy metal impurities in the ashes (especially cadmium, copper, lead and zinc) and also the lack of economic processing methods to remove these heavy metals [1–3]. Therefore, a new method for heavy metal removal is investigated. In metallurgical processes, chlorination has developed an important method for the removal of many heavy metals in the last years [2]. It is based on the high chemical af?nity of the chlorinating agents towards heavy metals which can be achieved at relatively moderate temperatures compared to other methods (oxidation or reduction processes) [3]. Some of the relevant chlorinating agents are gaseous chlorine, hydrochloric acid, carbon tetrachloride and metal chlorides, espe-

? Corresponding author at: Graz University of Technology, Institute for Process Engineering, Inffeldgasse 21b, 8010 Graz, Austria. Tel.: +43 316 481300 22; fax: +43 316 481300 4. ¨ E-mail address: markus.joeller@tugraz.at (M. Joller). 0255-2701/$ – see front matter ? 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.cep.2008.05.003

cially alkali or alkaline earth chlorides (natrium chloride, potassium chloride, magnesium chloride, calcium chloride) [2,3]. Metal chlorides offer the advantages of low cost, easy availability, low toxicity and unproblematic handling of the chlorination process. The most commonly researched metal chloride is calcium chloride because of its high chlorination ef?ciency [4]. Nevertheless, in literature little is known about the in?uence of the gaseous atmosphere on the thermal behaviour of calcium chloride. The composition of the gaseous atmosphere in the reactor depends, at least partly, on the heating system. Generally there are two types of industrial reactors, namely directly and indirectly heated systems (with the heat source inside or outside the reactor, respectively). Whereas directly heated systems are preferably ?red with fossil fuels like methane or heating oil, indirectly heated systems can also use electricity or other heat media. In the case of the directly heated reactor, ?ue gas resulting from combustion of the fuel creates the atmosphere, in the second case the atmosphere can be adjusted freely (ambient air, inert gas—no reaction with the feedstock, reactive gas—reaction with the feedstock). In spite the ?exibility of the atmosphere of the indirectly heated reactor its protection from the aggressive chlorine atmosphere requires the use of a reactor lining which is corrosion resistant and shows a good heat conduction.

G. Fraissler et al. / Chemical Engineering and Processing 48 (2009) 380–388

381

In the case of the directly heated system the chlorine rich gas released from the heated material gets diluted and therefore a ceramic lining (for example aluminium oxide of low cost) can be utilised. Due to this technically relevant advantage for an industrial application, the directly heated reactor is investigated within this work. The incoming atmosphere in the directly heated reactor consists of oxygen and nitrogen (from the combustion air), carbon dioxide (from the combustion of the fuel) as well as water vapour. According to literature, water has a great in?uence on the removal of heavy metals [5]. This is the reason why this work aims to investigate the in?uence of water vapour in the ?ue gas: ? ?rstly on the thermal decomposition of calcium chloride and ? secondly on its chlorination effect on heavy metals, by simulating the combustion of methane in the reactor. Regarding the occurrence of calcium chloride, information can be gained from literature [6,7]. Calcium chloride is highly hygroscopic and forms several hydrates. In literature four solid hydrates of calcium chloride (hexa-, tetra-, di- and monohydrate) have been identi?ed [6]. The decomposition of the hydrates may occur under the following conditions, which are de?ned by a calcium chloridewater system [6,7]. CaCl2 ·6H2 O(s) → CaCl2 ·4H2 O(s) + 2H2 O(s, l) CaCl2 ·4H2 O(s) → CaCl2 ·2H2 O(s) + 2H2 O(l) CaCl2 ·2H2 O(s) → CaCl2 ·H2 O(s) + H2 O(l, g) CaCl2 ·H2 O(s) → CaCl2 (s) + H2 O(g) ?55–30 ? C 30–45 ? C 45–176 C
? ?

method was used in the work intended in order to achieve an improved characterisation of calcium chloride and its chlorination effect under consideration of the chlorine release in a methane ?red reactor. In addition, the in?uence of water on the chlorine release was investigated by synthetically produced dry and humid ?ue gas conditions. 2. Materials and methods 2.1. Materials The calcium chloride used was not pure but applicative for industrial processes. Its chemical composition is listed in Table 1. Concerning the content of calcium, chlorine and water (on a wet basis), it is a mixed hydrate, assuming that the whole water (21 wt%) is bound to the calcium chloride. The material used contains a small amount of impurities particularly in the form of sodium and potassium compounds. Before starting the thermal analysis the calcium chloride in ?ake form had to be crushed in a swing mill (<63 m) in order to improve the thermal contact with the sample holder and thus to optimise the heat transfer. Because of its hygroscopicity the calcium chloride had to be stored in an airtight sample box [12]. 2.2. Methods Two methods of thermal analysis were used, namely thermogravimetric analysis (TGA) and differential thermal analysis (DTA). The thermal analyser used (Netzsch STA 409 PC LUX) offered the possibility to perform both analyses simultaneously. During the TGA the mass change (in wt%) of calcium chloride due to release of its constituents was measured as a function of temperature (in ? C) while the sample was subjected to a controlled temperature program, whereas during the DTA the temperature differences between calcium chloride and the reference were measured. These temperature differences indicated enthalpy changes (in mW mg?1 ) due to chemical and physical processes. Maxima of the DTA curve indicate that at these temperatures the highest rate of enthalpy change occurred. Additionally, for a better evaluation of the mass change the TGA curve can be differentiated (in wt% min?1 ) to render a differential thermogravimetric curve (DTGA). Maxima of the DTGA curve indicate that at these temperatures the highest rate of mass change during the thermal process occurred.
Table 1 Elemental composition of the calcium chloride used within the investigations performed Element Cd Cr Cu Ni Pb Zn Ca Cl K Mg Na P S Al Fe Si H2 O Values are given on wet basis. Content (mg kg?1 ) <0.33 <3.3 <3.3 <0.66 <3.3 <0.66 278,000 498,000 1,600 10.6 8,670 <3.3 222 <33 51.7 <330 210,000

(1) (2) (3) (4)

176–260 C

anhydrous calcium chloAt least at temperatures above ride should be formed. Depending on the presence of water vapour, carbon dioxide and oxygen in the atmosphere there may be several possible pathways of chlorine release from anhydrous calcium chloride including the intermediate reaction product calcium hydroxychloride which is formed with water [6–11]. CaCl2 (s) + H2 O(g) → Ca(OH)Cl(s) + HCl(g) Ca(OH)Cl(s) + CO2 (g) → CaCO3 (s) + HCl(g) Ca(OH)Cl(s) → CaO(s) + HCl(g) CaCl2 (s) + H2 O(g) + CO2 (g) → CaCO3 (s) + 2HCl(g) < 640 ? C (8) CaCl2 (s) + 0.5O2 (g) → CaO(s) + Cl2 (g) 600–800 ? C (9) 640–740 C
?

260 ? C

410–740 ? C 410–640 C
?

(5) (6) (7)

Reactions (5)–(9) show that water vapour in the atmosphere might cause chlorine release (in the form of hydrogen chloride) in a wider temperature range than without water vapour (in the form of chlorine gas). But the reactions and temperature ranges illustrated above are not related to ?ue gas conditions as caused by the combustion of methane. According to literature only little variations of the gas composition, especially variations of the water and carbon dioxide concentration in the atmosphere, may in?uence the chlorine release and the chlorination of heavy metals [6–11]. This circumstance motivates further research on the decomposition of calcium chloride under ?ue gas conditions as caused by the combustion of methane, which has been performed by means of thermal analysis. Thermal analysis is a physical method which studies the chemical and physical behaviour of materials in dependence of the temperature under controlled atmospheric conditions. This

382

G. Fraissler et al. / Chemical Engineering and Processing 48 (2009) 380–388

Table 2 Composition (volume percent) of dry and humid ?ue gas from combustion of methane with air Concentration (v%) Dry gas O2 CO2 N2 H2 O 12.0 5.0 83.0 0.0 Humid gas 10.8 4.5 74.7 10.0

Excess air ratio of 2.2, corresponding to an adiabatic ?ame temperature of 1008 ? C.

the gas feed pipe before the entrance of the thermal analyser could not be heated due to the design of the analyser. A scheme of the experimental setup is illustrated in Fig. 1. In order to check the consistency and repeatability of the results, all measurements under dry and humid ?ue gas conditions were repeated four times. Furthermore, with the aim to understand the thermochemical behaviour accompanying thermodynamic analyses were performed using the software HSC Chemistry Version 3.0. These thermodynamic considerations provide information about possible reaction pathways and about the stability of the reaction products. 3. Results Fig. 2 illustrates the result of a representative thermal analysis for dry gas conditions. The TGA curve indicates mass losses of 21 wt% up to 260 ? C. From 260 ? C to approximately 800 ? C the mass remains constant at 79 wt% of the weighed input sample. From 800 ? C to 1400 ? C the mass decreases by 32 wt% during the heating phase and further 30 wt% during the cooling phase, resulting in a remaining mass of 17 wt% of the input sample. Regarding the DTA curve, three distinct endothermic peaks can be identi?ed in the whole temperature range. The ?rst peak is, in fact, an overlap of several peaks with maxima between 220 ? C and 250 ? C, the second peak has its maximum at 490 ? C and the third at 760 ? C. In contrast to dry gas conditions, the measurements in humid gas indicate that the TGA curves differ distinctly only above 800 ? C (Fig. 3). During the heating phase in the temperature range from 800 ? C to 1400 ? C the mass losses are higher at the beginning although, compared to dry gas conditions, the total mass losses up to 1400 ? C (35 wt%) are similar. In the course of the cooling phase another 9 wt% are lost resulting in a remaining mass of 35 wt% of the input sample. This decrease of mass loss indicates that the thermal behaviour (type and/or rate of reaction) of calcium chloride in humid ?ue gas differs from its behaviour in dry ?ue gas in the considered temperature range. Changing to the DTA, the curve for humid gas conditions is similar to dry gas and indicates

Before starting the measurements, the crushed and weighed (100 mg ± 5%) calcium chloride was put into the sample holder and compressed with a te?on rod. The material of the sample holder was corundum. The furnace was heated at a rate of 20 K min?1 from 20 ? C (ambient temperature) to 1400 ? C and, after having reached the maximum temperature, it was cooled at a rate of 10 K min?1 . The decreased cooling rate depends on the cooling limitation of the analyser used. The purge gas used in the furnace was in the ?rst case synthetically produced dry ?ue gas and in the second case humid ?ue gas simulating the combustion of methane (at an excess air–fuel ratio of 2.2, corresponding with an adiabatic ?ame temperature of 1008 ? C). The ?ue gas composition for both cases is given in Table 2. The gas ?ow of the dry gas in the furnace was adjusted to 100 ml min?1 . Considering the water concentration, the adaptation of the gas feed pipe with two thermostats connected in series offered the opportunity of controlled humidi?cation of the dry combustion gas. Whereas the ?rst thermostat was heated up to approximately 70 ? C and provided a ?rst, more than desired humidi?cation, the second thermostat was set at exactly 45 ? C. This temperature allowed the condensation of excess water vapour in order to adjust the remaining humidity to approximately 10 v%. In order to avoid condensation, the gas feed pipe between the ?rst and second thermostat as well as between the second thermostat and the inlet to the thermal analyser were heated at 70 ? C. Only the last 10 cm of

Fig. 1. Scheme of the experimental setup for humid ?ue gas conditions. For dry conditions the purge gas used is fed into the thermal analyser directly.

G. Fraissler et al. / Chemical Engineering and Processing 48 (2009) 380–388

383

Fig. 2. Thermal analysis of calcium chloride (cH2 O = 21 wt%) under dry gas conditions containing O2 (12 v%), CO2 (5 v%) and N2 (83 v%) by illustrating TGA (thermogravimetric analysis), DTGA (differential thermogravimetric analysis) and DTA (differential thermal analysis) in the temperature range from 20 ? C to 1400 ? C. Red curves—heating phase (heating rate = 20 ? C min?1 ); green curves—cooling phase (cooling rate = 10 ? C min?1 ); blue curve—theoretical mass loss of calcium chloride by evaporation (based on vapour pressure data from HSC Chemistry 3.0 and purge air); magenta curve—theoretical total mass loss by formation of calcium oxide in the crucible. (For interpretation of the references to color in this ?gure legend, the reader is referred to the web version of the article.)

the same three endothermic peaks at the temperatures mentioned above. For both cases of dry and humid gas, a detailed interpretation of the curve progressions and the reaction mechanisms behind them is given in the following chapter. With regard to repeatability of the TGA curves for dry gas conditions, in the heating phase below approximately 1200 ? C no variation (<1 wt%) occurred. Above this temperature the curves diverged increasingly up to a maximum variation of 3 wt% at 1400 ? C. Using humid ?ue gas, the same phenomenon occurred already above approximately 800 ? C and caused a mass variation of up to 9 wt% at 1400 ? C. Probably material inhomogeneities between the samples used in the repetitions, caused by handling and sam-

pling, in?uenced the decomposition of calcium chloride and hence the mass losses at higher temperatures due to different material structure and different heat transfer rates. Differences between dry and humid ?ue gas conditions may be explained by different chemical reactions illustrated in the following chapter. Changing to DTA curves, consistent positions of the peaks below 800 ? C allow a qualitative identi?cation and interpretation of chemical reactions with regard to both dry and humid ?ue gas conditions. Above 800 ? C, it must be noted that the peaks are not well-de?ned and reproducible either. This means that chemical and physical processes cannot be interpreted meaningfully. Hence, in the temperature range from 800 ? C to 1400 ? C reactions can only be discussed by means of the TG (DTG) curves.

Fig. 3. Thermal analysis of calcium chloride (cH2 O = 21 wt%) under humid ?ue gas conditions containing O2 (11 v%), CO2 (5 v%), N2 (74 v%) and H2 O (10 v%). For details see caption of Fig. 2.

384

G. Fraissler et al. / Chemical Engineering and Processing 48 (2009) 380–388

4. Discussion 4.1. Heating up to 260 ? C TGA curves produced under both (dry and humid) gaseous atmospheres show that the mass losses up to 260 ? C correspond well with the respective decomposition reactions (3) and (4) of calcium chloride hydrates. Under consideration of the water content of the input sample (approximately 21 wt%) as well as of the maxima of the peaks of the differentiated curves (180–190 ? C and 230 ? C) and their positions compared to the temperature range of the reactions as given in literature (45–176 ? C, 176–260 ? C), the calcium chloride used is a mixed hydrate of mono- and dihydrate. When temperature increases up to 260 ? C it decomposes into water and anhydrous calcium chloride which implies that no chlorine is released in this temperature range. The fact that the maxima of the peaks of the ?rst mass losses (180–190 ? C) are not exactly located in the forecast temperature range is most likely due to kinetic effects, which have been caused most probably by the time dependent heating process, thereby possibly shifting the peaks to higher temperatures [13]. 4.2. Heating between 260 ? C and 782 ? C (melting point of calcium chloride) Under dry gas conditions, in the temperature range from 260 ? C to the melting point of calcium chloride (782 ? C) no relevant mass losses were detected and therefore no chlorine release occurred although according to reaction (9) calcium chloride may react with oxygen above 600 ? C forming calcium oxide and chlorine gas. The range between start and end temperature of the chlorine release is very important for the evaluation of chlorinating agents because it may considerably in?uence the activation of the heavy metals selected. When using humid ?ue gas between 260 ? C and 782 ? C, no further mass losses occurred, too. Reactions (5)–(7) illustrate that calcium chloride may react with water vapour under formation of ?rstly calcium hydroxychloride above 410 ? C and secondly calcium oxide by its decomposition above 640 ? C or forming calcium carbonate by reaction with carbon dioxide above 410 ? C. In every step chlorine is released in the form of hydrogen chloride. But literature data illustrate that calcium hydroxychloride is formed solely under atmospheric conditions, which require more than 25 v% water vapour and less than 1 v% carbon dioxide [10]. In dependence on these data in the experiments performed the water vapour concentration was too low (10 v%) as well as the carbon dioxide concentration was too high (5 v%) for calcium hydroxychlorideformation. That is the reason why calcium hydroxychloride was most probably not formed and no related chlorine release occurred. The release of hydrogen chloride by a direct reaction of calcium chloride with water vapour and carbon dioxide forming calcium carbonate in this temperature range probably did not occur either (reaction (8)). The reason will be given within the discussion of the heating period from 782 ? C to 1400 ? C. Summarised for both cases of dry and humid ?ue gas, no signi?cant mass losses of calcium chloride were detected between 260 ? C and its melting point (782 ? C) by means of TGA within the heating time of approximately 26 min. Hence, for the activation of heavy metals in this temperature range no released chlorine is provided. 4.3. Heating between 782 ? C and 1400 ? C and subsequent cooling in dry gas Above the melting temperature of calcium chloride differences regarding the total mass loss for dry and humid ?ue gas condi-

Fig. 4. Vapour pressure (Pa) of calcium chloride based on calculations with thermodynamic software in comparison to literature data [14].

tions occur. In the case of dry ?ue gas, the total measured mass loss amounts to 62 wt% as shown in Fig. 2 (32 wt% during the heating phase and 30 wt% during the cooling phase). A calculation of the theoretical mass decrease based on evaporation of calcium chloride and a subsequent saturation of the purge gas with calcium chloride indicates a maximum mass loss of about 73 wt% as shown with the blue curve in Fig. 2. The vapour pressure data considered are the data provided by the thermodynamic software package used. The validity of these data was checked by comparison with literature data [14]. In Fig. 4 the temperature dependent vapour pressures according to both datasets are plotted. This calculated total mass loss by evaporation is higher than the total mass loss observed although during the heating phase the experiments show a higher mass loss than the calculations. The difference of mass loss cannot be explained by an uncertainty of thermodynamic data since the mass loss is similar with both of the thermodynamic datasets used for calculating the evaporation of calcium chloride. These considerations demonstrate that during the heating phase the high mass losses measured cannot simply be caused by the release of gaseous calcium chloride from the sample holder (reactions (10) and (11)) but are caused by additional chemical reactions, which have to be identi?ed. CaCl2 (s) → CaCl2 (l) CaCl2 (l) → CaCl2 (g) 782 ? C > 782 ? C (10) (11)

¨ Accompanying test runs at Netzsch Geratebau GmbH using a thermal analyser (Netzsch STA 409 CD) coupled with a mass spectrometer in order to analyse the gas formed showed that neither calcium ions nor any calcium-chlorine-compounds could be detected. Thus thermodynamic considerations were used to evaluate the probable reaction products and reaction pathways. Thermodynamic equilibrium calculations as well as phase stability diagrams indicate that calcium carbonate below approximately 700 ? C as well as calcium oxide above approximately 700 ? C are the most stable products of calcium chloride in ?ue gas. Fig. 5 shows the phase stability diagram of calcium compounds at 1000 ? C depending on the partial pressures of oxygen and chlorine gas. Calcium chloride may react with the main purge gas component oxygen by forming calcium oxide and chlorine gas or with carbon dioxide and oxygen forming calcium carbonate and chlorine gas. Both calcium products are under the conditions prevailing in solid state. Nevertheless, calcium has left the sample holder in gaseous form since the amount of remaining material after thermal treatment (17 wt%) is smaller than the theoretical mass remaining in form of calcium oxide (40 wt%). Thus the reaction pathway is of interest. According to the free enthalpies of reaction, reactions (12)

G. Fraissler et al. / Chemical Engineering and Processing 48 (2009) 380–388

385

of carbon monoxide and atomic oxygen strongly depend on temperature and reach according to equilibrium calculations for the system investigated their maximum at 1400 ? C with approximately 0.001 v% each. CaCl2 (g) + 0.5O2 (g) → CaO(s) + Cl2 (g) > 782 ? C (12)

CaCl2 (g) + CO2 (g) + 0.5O2 (g) → CaCO3 (s) + Cl2 (g) > 782 ? C (13) It is assumed, that the release of calcium and chlorine under dry gas conditions proceeds in the following way. Calcium chloride evaporates from the sample holder and mixes with the purge gas ?ow. The amount evaporated is determined by the temperature dependent vapour pressure, the purge gas ?ow and the time given due to the heating rate. Subsequently, the calcium chloride may react according to the reaction in Fig. 6 with oxygen to solid calcium oxide, which is the preferred product above approximately 900 ? C, and chlorine gas. This reaction lowers the concentration of calcium chloride in the gas leading further to enhanced evaporation. Hence, the reaction to the oxide proceeds mainly in the gas phase as indicated in Fig. 8. Therefore, a large part of the calcium oxide formed probably precipitates outside the sample holder and cannot be detected by the balance. Moreover, reactions of calcium chloride with carbon monoxide and radicals may take place in a negligible amount. Further mass losses in the range of 1–2 wt% of the sample can be related to the release of impurities contained in the input sample, particularly in the form of sodium and potassium chlorides (reactions (14)–(17)). NaCl(s) → NaCl(l) NaCl(l) → NaCl(g)
Fig. 6. Free enthalpies of reaction (kJ) of gaseous and liquid calcium chloride reacting with oxygen solely to form calcium oxide and chlorine gas and with carbon dioxide and oxygen to form calcium carbonate and chlorine gas.

Fig. 5. Phase stability diagram of the system Ca–Cl–O at 1000 ? C; dashed line indicates O2 concentration applied for the experiments.

800 ? C > 800 C 772 ? C > 772 ? C
?

(14) (15) (16) (17)

KCl(s) → KCl(l) KCl(l) → KCl(g)

and (13) only proceed below approximately 1000 ? C and if gaseous calcium chloride is present. They do not take place with solid or liquid calcium chloride under given conditions, which is depicted in Fig. 6. Thermodynamic investigations of further reactions show that reactions including carbon monoxide and the atomic oxygen forming calcium oxide and calcium carbonate should also proceed with solid and liquid calcium chloride (Fig. 7). But the amounts

4.4. Heating between 782 ? C and 1400 ? C and subsequent cooling in humid gas In the case of humid ?ue gas, the total mass losses above 260 ? C amount to 44 wt% of the sample mass (35 wt% during the heating phase and 9 wt% during the cooling phase). This result strongly differs from the theoretical mass loss by evaporation (blue curve in Fig. 3), which would lead in this case to a complete evaporation, since the gas ?ow has been enlarged with the water uptake. Compared to the dry gas experiment, where the mass decrease amounted to 62 wt%, the mass loss during heating is higher and during cooling much smaller. Obviously, the thermal behaviour of calcium chloride under humid gas conditions differs in comparison to dry gas conditions. From thermodynamic calculations it can be seen that calcium chloride may react with water vapour by the formation of calcium oxide and hydrogen chloride or with water vapour and carbon dioxide by the formation of calcium carbonate and hydrogen chloride as shown in Fig. 9 and reactions (18) and (19). Similarly to the reaction of calcium chloride with oxygen, the calculations demonstrate that calcium chloride has to be present in gaseous state to make the reaction happen (see Fig. 6). In addition, reactions of solid and liquid calcium chloride with carbon monoxide, atomic oxygen and the hydroxide radical, which have its maximum amount in the temperature region investigated of approximately 0.026 v% at 1400 ? C, take

Fig. 7. Free enthalpies of reaction (kJ) of solid calcium chloride reacting with carbon monoxide and atomic oxygen.

386

G. Fraissler et al. / Chemical Engineering and Processing 48 (2009) 380–388

Fig. 8. Scheme of the reaction mechanism of gaseous calcium chloride with oxygen forming calcium oxide and chlorine gas in the reactor of the thermal analyser by using dry ?ue gas in the temperature range between 772 ? C and 1400 ? C. ( ) Solid calcium chloride; ( ) liquid calcium chloride; ( ) gaseous calcium chloride; ( ) calcium oxide; ( ) chlorine gas.

place. The formation of relevant amounts of calcium hydroxychloride, which is also formed by a heterogeneous reaction between calcium chloride and water vapour, is not probable but the formation of small amounts cannot be ruled out completely due the fact that the local gas composition in the sample holder may differ strongly from the bulk gas composition. CaCl2 (g) + H2 O(g) → CaO(s) + 2HCl(g) > 782 ? C (18)

ditions. This assumption of a faster reaction rate under humid ?ue gas conditions is supported by the initially higher mass loss (steeper TGA curve) during the heating phase in the regarded temperature range above the melting point of calcium chloride as described in the previous chapter. Moreover, thermodynamic calculations (Figs. 6 and 9) indicate that the reaction of gaseous calcium chloride with water vapour is favoured in comparison to the reaction with oxygen. Again, reactions of calcium chloride with carbon monoxide and radicals may take place in a negligible amount. 4.5. Discussion of DTA curves

CaCl2 (g) + H2 O(g) + CO2 (g) → CaCO3 (s) + 2HCl(g) > 782 ? C (19) The difference of the mass decreased measured to the estimated mass decrease (79 wt%) by evaporation of calcium chloride only amounts to 33 wt% of the initial sample mass, whereas the difference to the calculated mass decrease by forming calcium oxide amounts only to 5 wt%. This leads to the conclusion that the gaseous calcium chloride reacts immediately after vapourisation, which means that either the evaporation of calcium chloride or the mass transport of water vapour and hydrogen chloride respectively are determining parameters for the mass loss in this case. Hence, the solid reaction products formed stay inside the sample holder and are not entrained by the gas ?ow (Fig. 10). The reason for the immediate reaction may be a faster reaction rate of reaction (18) in comparison to reaction (12) under dry gas con-

In both cases (dry and humid ?ue gas) the evaluations of the DTA curves show that the ?rst of the three endothermic peaks (maxima between 220 ? C and 150 ? C) indicates the decomposition of the hydrates. The second slight peak (maximum at 490 ? C) cannot be interpreted with certainty. The most probable hypothesis is the melting of the eutectic composition of the ternary system CaCl2 –KCl–NaCl. Furthermore, the sulphur in the sample could be present as gypsum, which undergoes a phase transformation from anhydrite III to anhydrite II. This reaction is endothermic and takes place at 500 ? C [10], which would well agree with the maximum of the related peak. The third peak (maximum at 770 ? C) can be attributed to the melting temperature of calcium chloride at 782 ? C [10]. 4.6. Relevance of experimental results for heavy metal removal In order to estimate the heavy metal activation and removal in the form of gaseous metal chlorides by the use of calcium chloride as chlorine donor, additional thermodynamic calculations were performed. Fig. 11 illustrates the free enthalpy of reaction of the investigated heavy metals cadmium, copper, lead and zinc with chlorine gas and with hydrogen chloride assuming that heavy metals contained in sewage sludge ashes occur primarily in oxide form. The results show that below approximately 600 ? C the reactions of the heavy metal oxides with hydrogen chloride are thermodynamically favoured in comparison to reactions with chlorine gas. Above this temperature the opposite is the case. The reason is that below approximately 600 ? C the conversion of HCl and O2 to Cl2 and H2 O will proceed on its own and contribute to the total free enthalpy of reaction as it can be seen from Fig. 11. Furthermore, it can be concluded from Fig. 11 that the chlorination of the heavy metals investigated should take place in the order that lead should form a chloride compound most easily fol-

Fig. 9. Free enthalpies of reaction (kJ) of gaseous and liquid calcium chloride when reacting with water vapour to form calcium oxide and hydrogen chloride and with carbon dioxide and water vapour to form calcium carbonate and hydrogen chloride.

G. Fraissler et al. / Chemical Engineering and Processing 48 (2009) 380–388

387

Fig. 10. Scheme of the reaction mechanism of gaseous calcium chloride with water vapour forming calcium oxide and hydrogen chloride in the reactor of the thermal analyser by using humid ?ue gas in the temperature range between 772 ? C and 1400 ? C. ( ) Solid calcium chloride; ( ) liquid calcium chloride; ( ) gaseous calcium chloride; ( ) calcium oxide; (?) hydrogen chloride.

Fig. 11. Free enthalpy of reaction (kJ) for the reactions of heavy metal oxides with chlorine gas or hydrogen chloride as well as for reactions of hydrogen chloride gas with oxygen (heavy metal reactions are related to 1 mol of the metal).

lowed by cadmium, zinc and copper, which shows the smallest free enthalpy of reaction for the chloride formation in a wide temperature range. Since the main chlorine release occurs, according to the experiments above the melting point of calcium chloride, the reaction with chlorine gas produced in a dry atmosphere should be the thermodynamically favoured reaction for the chlorination process. But the experiments show a slower reaction progress for the release of the chlorine gas compared to the release of hydrogen chloride (compare Fig. 2 with Fig. 3 with regard to the TGA curves in the temperature range between 800 ? C and 1200 ? C) and therefore no clear indication concerning the choice of the atmosphere can be derived from the experiments. Thus, further TGA/DTA experiments with mixtures of sewage sludge ashes with calcium chloride have to be performed in order to asses the kinetic behaviour of the chlorine release compared to the subsequent chlorination of the heavy metal oxides and to investigate possible chemical interactions of calcium chloride with sewage sludge ash at temperatures below the melting point of calcium chloride. 5. Conclusions In order to estimate the chlorination effect of calcium chloride under consideration of different gas humidities on the removal of heavy metals (cadmium, copper, lead and zinc), a comparative thermal analysis of calcium chloride between 25 ? C and 1400 ? C in dry and humid gas atmospheres (simulating a methane combustion in air) was performed. The dry gas contained 12 v% oxygen, 5 v% carbon dioxide and 83 v% nitrogen whereas the wet gas was

composed of 11 v% oxygen, 5 v% carbon dioxide, 74 v% nitrogen and 10 v% water vapour. In both cases (dry and humid gas atmosphere) the ?rst mass losses of the calcium chloride samples that occurred at temperatures up to 260 ? C are caused by the decomposition of hydrates and amount to 21 wt% of the input mass. Relevant chlorine release occurs only above the melting temperature of calcium chloride (782 ? C). It got obvious that between the two sets of experiments different mechanisms dominate the mass decrease of the sample during the thermal process. Under dry gas conditions calcium chloride evaporates from the sample according to the vapour pressure near the sample surface and reacts with the gas phase to form calcium oxide and chlorine gas. The chemical reactions proceed that slow that the mass decrease of the sample is controlled mostly by the mass transfer to the bulk gas phase. Heterogeneous reactions of calcium chloride and the gas phase seem to play no important role. Under humid gas conditions calcium chloride immediately reacts after vapourisation with water vapour forming calcium oxide and hydrogen chloride. The mass decrease of the sample is controlled either by the evaporation of the sample or by mass transport into and out of the sample holder. As in the case of dry gas heterogeneous reactions of calcium chloride and the gas phase play most probably no important role. Regarding the release behaviour of heavy metals the utilisation of humid atmosphere should lead to a faster release and thus a better availability of the chlorine in form of hydrogen chloride gas than under dry conditions, where chlorine gas is formed. But thermodynamic investigations showed that the subsequent reactions of chlorine gas with heavy metal oxides are thermodynamically

388

G. Fraissler et al. / Chemical Engineering and Processing 48 (2009) 380–388 ride volatilization of heavy metals, Kagaku Kogaku Ronbunshu 30 (2004) 778–785. Z. Youcai, S. Stucki, C. Ludwig, J. Wochele, Impact of moisture on volatility of heavy metals in municipal solid waste incinerated in a laboratory scale simulated incinerator, Waste Management 24 (6) (2004) 581– 587. The Doe Chemical Company, Calcium chloride handbook: a guide to properties, forms, storage and handling, The Doe Chemical Company Customer Information Center, Midland, USA, 2003. A. Willmes, Taschenbuch Chemische Substanzen, second ed., Verlag Harri Deutsch, Frankfurt on the Main, Germany, 2001. K.M. Allal, J.C. Dolignier, G. Martin, Determination of thermodynamical data of calcium hydroxichloride, Revue de l’IFP 52 (3) (1997) 361–368. K.M. Allal, J.C. Dolignier, G. Martin, Reaction mechanism of calcium hydroxide with gaseous hydrogen chloride, Revue de l’IFP 53 (6) (1998) 871–881. J. Partanen, P. Backman, R. Backman, M. Hupa, Absorption of HCl by limestone in hot ?ue gases. Part II: importance of calcium hydroxychloride, Fuel (2004) 1674–1684. S. Wedemeyer, Recovery of heavy metals from ?y ash of municipal solid waste incinerators by chloridizing volatilization, PhD Thesis, Technical University of Clausthal, Clausthal, Germany, 2003. K. Naruse, S. Ozawa, Y. Kojima, N. Tanahashi, H. Matsuda, T. Yanase, Formation of HCl in NaCl–CO2 –H2 O and CaCl2 –CO2 –H2 O reaction systems under simulated conditions of municipal waste incineration, Kagaku Kogaku Ronbunshu 30 (2004) 726–731. S.A. Scott, J.S. Dennis, J.F. Davidson, A.N. Hayhurst, Thermogravimetric measurements of the kinetics of pyrolysis of dried sewage sludge, Fuel 85 (2006) 1248–1253. D.L. Hildenbrand, N.D. Potter, The vapour pressure and thermodynamics of vaporization of calcium chloride, The Journal of Physical Chemistry 67 (1963) 2231–2232.

favoured compared to reactions with hydrogen chloride in the relevant temperature range. Thus the determination of the favourable process depends on the reactivity of gaseous chlorine compounds with the heavy metal oxides. Acknowledgements The authors wish to acknowledge the support of the Austrian Bioenergy Centre, Graz (A), in collaboration with ASH DEC Umwelt AG, Vienna (A), BIOS BIOENERGIESYSTEME GmbH, Graz (A), ARP ¨ GmbH, Leoben (A), and Netzsch Geratebau GmbH, Selb (D). The Austrian Bioenergy Centre GmbH is a competence centre that is run in the frame of the Austrian “K plus” program. It receives ?nancial support from the federal government of Austria, the Austrian states Styria and Lower Austria, and the city of Graz. References

[5]

[6]

[7] [8] [9]

[10]

[11]

[12] ? [1] J.A. Gondalez, O.D. Quiroga, M.C. Ruiz, Kinetic study of the chlorination of gallium oxide, Metallurgical and Materials Transactions 36B (2005) 445–452. [2] P.K. Jena, E.A. Brocchi, Metal extraction through chlorine metallurgy, Mineral Processing and Extractive Metallurgy Review 16 (4) (1997) 211–237. ˇ ? [3] R.Z. Vraˇ ar, K.P. Cerovic, The study of chlorination kinetics of copper(I)sulphide c by calcium chloride in the presence of oxygen, Metallurgical and Materials Transactions 31B (2000) 723–731. [4] K. Nakayama, Y. Tajima, S. Mizuno, Y. Kojima, H. Matsuda, M. Takada, Effect of NaCl, KCl and CaCl2 contained in MSW molten ?y ashes on chlo-

[13]

[14]


相关文章:
更多相关标签: