Experimental Investigation of Hydrogen Chlorine Bonding with


Experimental Investigation of Hydrogen Chlorine Bonding with...

3 downloads 83 Views 1MB Size

Energy Fuels 2010, 24, 5851–5858 Published on Web 11/02/2010

: DOI:10.1021/ef101048k

Experimental Investigation of Hydrogen Chlorine Bonding with Limestone and Dolomite in the Furnace of a Stoker-Fired Boiler Szawomir Poskrobko,*,† Jan Łach,‡ and Danuta Kr ol§ † Biazystok University of Technology, Wiejska 45C, 15-351 Biazystok, Poland, ‡Technical University of Radom, Krasickiego 54, 26-600 Radom, Poland, and §Silesian University of Technology, Konarskiego 18, 44-101 Gliwice, Poland

Received August 9, 2010. Revised Manuscript Received October 17, 2010

The paper presents results of experimental semi-technical investigations to limit the mobility of chlorine released in the form of hydrogen chloride from the fuel in the stoker-fired boiler furnace. In the combustion process, limestone and dolomite were used as the sorbents. The background for the research was a fuel containing a low gram fraction of chlorine, namely, extracted rapeseed meal supplemented with polyvinyl chloride (PVC) recyclate granules, to diversify the quantity of the quick-release chlorine. The research results indicate the tendency of changes of the hydrogen chloride concentration in the flue gas in relation to the amount of PVC added to the fuel, i.e., gram fraction of chlorine, amount of limestone and dolomite added to the fuel, as well as the molar ratio Ca/Cl2 or (Ca þ Mg)/Cl2, respectively. It has been proven that the greatest dechlorination capacity is exhibited by hydrated lime, lesser by limestone, and the least by dolomite, which, nonetheless, can be used under certain conditions for HCl bonding in boiler furnaces.

operating conditions,3 and (ii) formation of acid and toxic gases, such as chlorinated hydrocarbons and polychlorinated dibenzodioxins (PCDDs) and dibenzofurans (PCDFs). There are, however, means for effective limitation of the presence of both chlorine and hydrogen chloride in the gas atmosphere of the furnace. They involve bonding these gases with calcium sorbents, which are added to the fuel before feeding it into the boiler. The undertaken research, whose partial results have already been offered in ref 1, constitutes an attempt at formulating an experimentally verifiable solution to this difficult technological problem arising from the practical needs. The bonding of HCl already in boiler furnaces and, at the same time, reducing its mobility significantly limits the possibilities of chlorination of hydrocarbon products of incomplete combustion. Polychlorinated hydrocarbon derivatives (PCDDs/PCDFs) occur when fuel contains chlorine, for example, organically bound in waste or in polyvinyl chloride (PVC), post-consumer plastic (PCP), or other recyclates,4 which is released during the thermal decomposition of the fuel. It has to be noted that, according to information contained in the literature,5 the intensity of HCl formation resulting from the decomposition of organic substances, e.g., PVC, during combustion is incomparably greater than in the case of biomass fuels, which contain chlorine usually in the form of inorganic bonds (KCl and NaCl). In ref 4, it has been proven that the efficiency of calcium sorbents in reducing PCDD/PCDF formation is noticeable already when the molar ratio of Ca/Cl2 is between 1 and 2. Thus, limiting the mobility of organically bound chlorine in boiler furnaces

1. Introduction The goal of the present study is to compare the potential of different calcium sorbents in terms of reducing the mobility of chlorine, which is released in the form of hydrogen chloride from the combustible substance in the furnace of a 225 kW water-underfeed stoker-fired boiler, which is in semi-technical conditions. It is generally accepted that a measure of chlorine mobility reduction in boiler furnaces is the actual decrease in the hydrogen chloride concentration in flue gas. Results of experimental research involving the use of hydrated lime as a bonding material have been presented in ref 1. This work presents the results obtained for the same model or basic fuels (i.e., extracted rapeseed meal with different mass amounts of PVC recyclate added) but with the use of other calcium sorbents, namely, calcium carbonate (limestone, CaCO3) and dolomite (CaCO3 3 MgCO3). The studies on limiting the mobility of chlorine in boiler furnaces are motivated by the need of reducing both the intensity of corrosion processes (functional requirement) and the emission of harmful substances into the atmosphere (ecological requirement). Chlorine, similar to other halogens, is released in thermal processes (e.g., during combustion of fuels containing chlorine) in the molecular form in the temperature of about 600 °C2 or as hydrogen chloride in the higher temperatures, causing (i) high-temperature chlorine and chloride corrosion of steel structural components in boiler furnaces, which has significant consequences for their *To whom correspondence should be addressed. Telephone: þ48857469205 and þ48857469200. Fax: þ48857469210. E-mail: drposkrobko@ wp.pl. (1) Poskrobko, S.; Łach, J.; Kr ol, D. Experimental investigation of hydrogen chloride bonding with calcium hydroxide in the furnace of a stoker-fired boiler. Energy Fuels 2010, 24 (3), 1948–1957. (2) Liu, K.; Pan, W. P.; Riley, J. T. A study of chlorine behavior in a simulated fluidized bed combustion system. Fuel 2000, 79, 1115–1124. (3) Nielsen, H. P.; Frandsen, F. J.; Dam-Johansen, K.; Baxter, L. L. The implications of chlorine-associated corrosion on the operation of biomass-fired boilers. Prog. Energy Combust. Sci. 2000, 26 (3), 283–298. r 2010 American Chemical Society

(4) Lu, S. Y.; Chen, T.; Yan, J. H.; Li, X. D.; Ni, Y. L. M. J.; Cen, K. F. Effects of calcium-based sorbents on PCDD/F formation from pentachlorophenol combustion process. J. Hazard. Mater. 2007, 147 (1-2), 663–671. (5) Wey, M. Y.; Chen, J. C.; Wu, H. Y.; Yu, W. J.; Tsai, T. H. Formations and controls of HCl and PAHs by different additives during waste incineration. Fuel 2006, 85 (5-6), 755–763.

5851

pubs.acs.org/EF

Energy Fuels 2010, 24, 5851–5858

: DOI:10.1021/ef101048k

Poskrobko et al.

seems to be indispensable for both operational and ecological reasons. This issue is discussed in scientific and technical literature, because the optimal solution in terms of dechlorination of furnace atmosphere in processes, such as combustion, co-combustion, or gasification of various fuels, is yet to be discovered. Considering the fact that the fuel investigated within this paper is formed from extracted rapeseed meal and PVC recyclate and contains, as assumed by the authors, a relatively great gram fraction of sulfur (S = 1.5%), another question arises: what is the influence of sulfur on the dechlorination process? This question was already dealt with in ref 6, and the presented results of the experimental research pointed to benefits stemming from the coexistence in the boiler furnace of SO2 and HCl with CaCO3 sorbent, which effectively bonds these acid products of combustion. It turns out that HCl precipitates bonding of SO2 into CaSO4, whereas SO2 does not adversely influence the process of bonding HCl by calcium sorbent (CaCO3) into CaCl2. The mechanism of chemisorption of HCl in a SO2 environment was investigated for various combustion temperatures, including 750 °C, as well as for various concentrations of H2O and HCl in exhaust gases. Simultaneously, the bonding of sulfur oxides by means of a calcium preparation into a poorly soluble in water CaSO4 sulfate form prevents emission of their aerosols into the atmosphere, which is elucidated in ref 7. The formation and emission of sulfate aerosols is a common phenomenon in the combustion of biomass with a great content of alkali metals (potassium and sodium). Results of HCl chemisorption analysis by means of CaCO3 sorbent in the temperatures of 650 and 850 °C in a laboratory thermogravimetric stand are contained in ref 8. In the case of these temperatures, the finally obtained product was CaCl 2 . In the initial (transitional) phase, however, a volatile compound, calcium hydroxychloride (CaOHCl), was produced, which during subsequent phases in a HCl environment was transformed into CaCl2. Relying on the outcomes of research conducted in several different research centers, it can be concluded that limestone, similar to hydrated lime, can be successfully employed as a substance that neutralizes harmful effects of chlorine activity in combustion processes. Understanding exact mechanisms of chemical and thermal transformations of this sorbent in an acid environment created by acid products of combustion (i.e., HCl and SO2) had a fundamental impact on discovery and practical use of such a fuel as extracted rapeseed meal combined with various PVC recyclate admixtures. Applicability of CaCO3 as an addition to fuels rich in chlorine encourages investigation of other widely available calcium sorbents, which can be mixed with the basic fuel without making additional outlays on its pretreatment before the mixing. One of such sorbents is dolomite, which is used as the cheapest addition in boiler furnaces, in which hard coal of great sulfur content is combusted. As far as the authors of the present study are concerned, the usage of dolomite in the processes of dechlorination is not very widespread, because

the sorbent contains MgCO3, which in combination with HCl forms MgCl2 characterized by a low softening temperature equal to 708 °C, whereas CaCl2 softens in boiler furnaces in the temperature of 772 °C. This phenomenon results in a dangerous process of boiler slagging. To prevent this undesirable effect, sulfur is introduced into the furnace, which in the reaction with magnesium forms infusible MgSO4. Because of operational difficulties, dolomite finds a wider applicator only in coal-based technologies. However, it is only of little importance for the purpose of chlorine mobility reduction in combustion/co-combustion of waste-derived fuels or biomass rich in potassium. One cannot fail to notice, however, that good sorption properties of dolomite, along with its wide availability and low cost, stand in favor of undertaking research work aiming at confirming the possibility of using dolomite in new power energy technologies. Accordingly, refs 9-12 analyze the properties of dolomite, focusing on its transformations in temperatures typical for combustion and gasification processes. Special attention is paid to its capability of neutralizing the acid gaseous products of combustion and undesirable gasification products, including H2S. In ref 13, dolomite was studied from the perspective of an active substance, with application in CO2 and NO2 adsorption. The adsorption characteristic in the temperatures of 750, 800, and 850 °C was presented by means of an empiric Freundlich isotherm, suitable for the adsorbents with an irregular active surface. In the present study, dolomite was used for chemisorption of HCl in a boiler furnace, in which test fuels were combusted in the temperature of 750 °C. The fuels were supplemented with various amounts of PVC, and at the same time, they were characterized by a low content of Ca, Na, and K but contained a great gram fraction of sulfur to protect the boiler against slagging. It should be added that compounds of sulfur, such as, for instance, (NH4)2SO4, are used in the ChlorOut method patented by the company Vattenfall and used in combustion of biomass, which contains chlorine in the form of KCl.14 This method reduces the spread of hightemperature corrosion centers in the heat-exchange zone of boilers, as well as reduces NOx emissions. The above overview implies that the evaluation of different calcium sorbents [Ca(OH)2, CaCO3, and CaCO3 3 MgCO3] from the point of view of their capability of reducing the mobility of chlorine released in the form of HCl during combustion of PVC-supplemented fuel with a great content of sulfur still remains a valid and significant research problem. The quoted reports from the literature, which are mostly based on results of laboratory investigations, suggest that the sorbents may be effectively used in dechlorination of the (9) Hartman, M.; Trnka, O.; Svoboda, K. Fluidization characteristics of dolomite and calcined dolomite particles. Chem. Eng. Sci. 2000, 55 (24), 6269–6274. (10) Fuertes, A. B.; Velasco, G.; Alvarez, T.; Fernandez, M. J. Sulfation of dolomite particles at high CO2 partial pressures. Thermochim. Acta 1995, 254, 63–78. (11) Yrjas, P.; Iisa, K.; Hupa, M. Limestone and dolomite as sulfur absorbents under pressurized gasification conditions. Fuel 1996, 75 (1), 89–95. (12) Sciazko, M.; Kubica, K. The effect of dolomite addition on sulphur, chlorine and hydrocarbons distribution in a fluid-bed mild gasification of coal. Fuel Process. Technol. 2002, 77-78, 95–102. (13) Duffy, A.; Walker, G. A.; Allen, S. J. Investigations on the adsorption of acidic gases using activated dolomite. Chem. Eng. J. 2006, 117 (3), 239–244. (14) Brostr€ om, M.; Kassman, H.; Helgesson, A.; Berg, M.; Andersson, C.; Backman, R.; Nordin, A. Sulfation of corrosive alkali chlorides by ammonium sulfate in a biomass fired CFB boiler. Fuel Process. Technol. 2007, 88 (11-12), 1171–1177.

(6) Matsukata, M.; Takeda, K.; Miyatani, T.; Ueyama, K. Simultaneous chlorination and sulphation of calcined limestone. Chem. Eng. Sci. 1996, 51 (11), 2529–2534. (7) Hindiyarti, L.; Frandsen, F.; Livbjerg, H.; Glarborg, P.; Marshall, P. An exploratory study of alkali sulfate aerosol formation during biomass combustion. Fuel 2008, 87 (8-9), 1591–1600. (8) Partanen, J.; Backman, P.; Backman, R.; Hupa, M. Absorption of HCl by limestone in hot flue gases. Part II: Importance of calcium hydroxychloride. Fuel 2005, 84 (12-13), 1674–1684.

5852

Energy Fuels 2010, 24, 5851–5858

: DOI:10.1021/ef101048k

Poskrobko et al. Table 1. Percentage Elemental Composition of the Fuels1 elemental compositiona (%) fuels

addition of PVC (%)

C

H

N

S

Cl

O

fuel I fuel II fuel III fuel IV

2 5 8 10

49.29 49.28 49.27 49.26

6.94 6.92 6.89 6.88

6.63 6.43 6.24 6.11

1.35 1.31 1.27 1.25

0.35 0.87 1.40 1.74

35.44 35.19 34.93 34.76

a

Expressed on a dry ash-free basis.

the petrochemical industry, where it is used as an addition to liquid fuels), as well as its higher and higher crop yields, causing increased supply on the feed market, extracted rapeseed meal can potentially serve as a component in the production of wastederived fuels and can be used as an addition to various species of agricultural or wood biomass. Additionally, it can constitute a biomass component of coal fuels, which means that the extracted rapeseed meal can be co-combusted with coal in power boilers. The magnitude of the whole matter may be illustrated by the fact that about 60% of the whole mass of rapeseed processed in the production of biodiesel, i.e., during the pressing and extraction of oil, is byproducts in the form of press-cakes and meal. It seems, then, that byproducts of oleaginous plant processing can make up for wood biomass deficits to a much greater degree than plants from energy crop cultivations. Second, the elemental composition of extracted rapeseed meal is, in comparison to other available biofuels, marked by only a minimal gram fraction of chlorine (∼0.01% in dry ash-free mass). Accordingly, in this case, one is dealing with a so-called “zero matrix”. Such a negligible gram fraction of chlorine in the elemental composition renders extracted rapeseed meal as a suitable component of fuels formed from waste, which contain even considerable amounts of this element. Let us mention that different combustible substances can contain diversified (sometimes up to substantial) gram fractions of chlorine, and the element can be bound in manifold chemical configurations (both mineral and organic),5 which means that the release of hydrogen chloride in the course of thermal decomposition of these substances can be uneven. Thus, the gram fraction of chlorine in extracted rapeseed meal (i.e., the fuel used here as the basic fuel) can be arbitrarily adjusted by increasing a desired percentage amount of PVC recyclate granule addition. The properties of extracted rapeseed meal with PVC recyclate are expounded in ref 1. It might be claimed that, in such a way as described, a whole gamut of new fuels (Table 1) with a diversified gram fraction of chlorine in the combustible substance is obtained. The form in which chlorine is present is in such circumstances strictly determined by the kind of organic chemical bond. Specifically, in the considered case, hydrogen chloride will be readily released during the combustion process into the furnace chamber, directly through the layer of fuel on the grate, in a uniform and regular manner as a result of PVC recyclate decomposition. Furthermore, extracted rapeseed meal, in contrast to other biomass materials, is characterized by a relatively low content of alkali metals, such as sodium and potassium, and alkaline earth metals, such as calcium. Hence, also in this respect, we practically deal with a so-called “zero matrix”. Actually, this also means that almost the entire amount of alkali components, bonding hydrogen chloride released from the fuel, will be provided by the added calcium sorbents, such as mineral calcium compounds (e.g., limestone and dolomite). The sorbents, having the form of fine powder of a good grinding quality, will be mixed with fuel before the feeding into the boiler. However, if a fuel contains great amounts of alkali elements, then a considerable part of the alkali component originates from the fuel itself (conversely to the case of extracted rapeseed meal). In such a case, analogically to the situation with chlorine, chemical bonds in which these elements

Figure 1. Schematic of the boiler furnace (test stand): 1, primary air distribution; 2, lower part of the stocker; 3, feeding scrow (delivering the fuel and sorbent); 4, primary and secondary air flow fan; 5, through the channel; 6, convection chamber; 7, water jacket; 8, water heat exchanger, 9, smoke conduit; 10, upper lid; 11, movable hole cleaning elements; 12, chimney; 13, flue gas exhaust fan; 14, flue gas monitoring.

boiler atmosphere in industrial processes. That is why our experiments were conducted in semi-technical conditions, with the use of a 225 kW water-underfeed stoker-fired boiler. 2. Formulation of the Problem Experimental research was conducted, similarly as in ref 1, with a fuel formed from industrial biomass residues, namely, extracted rapeseed meal supplemented with PVC recyclate granules. The experimental test stand is presented in Figure 1. Prior to determining the emission of hydrogen chloride in the PVC-extracted rapeseed meal mixture combustion, gas samples were taken by applying an exhaust aspirator. Both sampling and determination of the hydrogen chloride in the exhaust gas were performed according to Standard PN-EN 1911-1, 1911-2, and 1911-3. Thus, the heated probe was placed in the channel, and the gas, which was sucked by the aspirator, passed through a system of scrubbers filled with water, according to Standard EN ISO 3696:1995. Prior to each measurement series, the so-called blind trial of the system was performed. After the completion of the sampling process, the absorption fluid (i.e., the water with absorbed hydrogen chloride) was transferred to a measuring flask, to which wash water from the collecting system was added. After the absorption of the hydrogen chloride, the solution was analyzed via the spectrophotometric method to determine chlorides, using mercuric thiocyanate. For this purpose, first, a number of sodium chloride standard solutions were prepared and, then, reagent solutions for color were added. After dyeing the solutions, their absorbance was measured with a spectrophotocolorimeter at a wavelength of 460 nm. Thus, it was possible to determine the linear dependence of absorbance on the concentration of chlorides expressed in terms of hydrogen chloride. Next, the absorbance of both the blind sample and absorption fluid samples prepared analogically was measured spectrophotometrically at an analogous wavelength. The hydrogen chloride content in the investigated samples was calculated using the relationships established for reference solutions.1 Let us remind the reasons and provide some new arguments for the choice of extracted rapeseed meal as a particular model fuel. First of all, presently, a big emphasis is placed on successive increasing of the role of renewable energy in total energy consumption, which for many countries is tantamount to making use of biofuels or fuels formed from various sorts of waste and agricultural biomass that can partly substitute fossil fuels. Taking into account the growing rapeseed cultivation acreage (mostly because of the increased rapeseed oil production for the needs of 5853

Energy Fuels 2010, 24, 5851–5858

: DOI:10.1021/ef101048k

Poskrobko et al.

Table 2. Relative Hydrogen Chloride Concentration (φ) in Wet Flue Gas

Table 3. Basic Technological Dimensionless Process Parameters: Gram Ratios

φ

calcium hydroxidea

addition addition of PVC of sorbent calcium number weight (%) weight (%) hydroxidea limestone dolomite Fuel I 1 2 3 4

2

1 1.5 2 3

5.36 3.85 2.91 2.55

8.88 8.01 7.29 6.71

9.16 8.25 7.56 6.92

Fuel II 5 6 7 8

5

1 2 3 4

11.49 7.45 4.92 3.82

15.06 12.12 9.53 8.22

15.84 12.74 9.85 8.51

Fuel III 9 10 11 12

8

1,5 3 4.5 6

12.11 7.18 5.22 4.64

16.86 12.23 10.35 9.13

17.49 12.74 10.58 9.33

Fuel IV 13 14 15 16 a

10

2 4 6 8

13.91 6.1 5.69 4.73

15.96 12.06 10.21 9.18

17.10 12.82 10.68 9.48

limestone

dolomite

(Ca þ (CaO þ Ca/Cl CaO/Cl Mg)/Cl MgO)/Cl

number

Ca/Cl

CaO/Cl

1 2 3 4

1.60 2.37 3.14 4.69

2.25 3.32 4.40 6.56

Fuel I 1.21 1.78 2.35 3.50

1.69 2.49 3.29 4.89

1.24 1.82 2.41 3.59

1.80 2.65 3.51 5.22

5 6 7 8

0.65 1.27 1.90 2.52

0.91 1.79 2.66 3.54

Fuel II 0.49 0.96 1.47 1.88

0.68 1.33 2.05 2.63

0.50 0.98 1.45 1.93

0.73 1.42 2.12 2.81

9 10 11 12

0.60 1.19 1.78 2.37

0.85 1.70 2.49 3.32

Fuel III 0.45 0.89 1.36 1.77

0.63 1.24 1.90 2.47

0.46 0.91 1.36 1.81

0.67 1.33 1.98 2.63

13 14 15 16

0.64 1.27 1.90 2.52

0.90 1.78 2.66 3.54

Fuel IV 0.48 0.95 1.44 1.88

0.67 1.32 2.01 2.63

0.49 0.97 1.45 1.93

0.71 1.41 2.11 2.81

a

The calcium hydroxide conditions apply to the study in ref 1.

The calcium hydroxide conditions apply to the study in ref 1.

investigated Ca(OH)2 sorbent. Thus, primarily four model fuels were formed (fuel I, fuel II, fuel III, and fuel IV) with elemental compositions, as stipulated in Table 1. As mentioned earlier, varied gram fractions of chlorine were obtained by adding PVC recyclate to the base fuel (extracted rapeseed meal) in the desired percentage mass amount. Next, to ensure a possibly long contact between the sorbent and the reactive environment, to all of the four model fuels calcium carbonate (limestone, CaCO3) or dolomite (CaCO3 3 MgCO3) was added, retaining adequate percentage proportions, such as presented in Table 2. The magnesium-calcium sorbent contained 72% calcium carbonate (CaCO3) and 28% magnesium carbonate (MgCO3). Let us note here that dolomite rock, depending upon the mine from which it is obtained, can contain different proportions of CaCO3 and MgCO3, which have an impact on various mass fractions of Ca and Mg as well as CaO and MgO. Table 2 makes it clear that, by diversifying the amount of sorbent added to each of the model fuels, actually 32 different fuels were obtained [(4  4)  2 = 32], which were experimentally tested in a 225 kW water-underfeed stoker-fired boiler. The detailed description of the operating parameters of this boiler, including the furnace efficiency η, and the result measurements concerning the concentration of gaseous products of combustion (CO2, CO, and NOx) were provided in ref 1. Here, let it be just reminded that (i) fuel feed efficiency, which ensured the optimal velocity of its movement along the grate toward the ignition zones, was equal to 30 kg/h, and the furnace temperature was 750 °C; (ii) the furnace was cooled after each of the tests, and the remaining ash was removed; (iii) the furnace was heated for each single test using the charcoal; and (iv) one of the essential issues influencing the favorable outcome of the experiment was achieving, by trial and error approach, great furnace efficiency at a relatively low excess air coefficient λ. In the process of combustion of each of the 32 fuels, the concentration of basic gaseous products was measured, and relying on the HCl emission level, the efficiency

engage remain diversified and difficult to identify, which in turn effects their various reactivities. What should also be noted is that the investigated fuel is rich in sulfur; its dry ash-free combustible substance contains 1.38% of this element. Sulfur is released during the decomposition of fuel; in the gas phase, it prevails in the form of oxides, and in the furnace, it reacts with alkalis, bonding them partly and at the same time increasing the temperature of ash softening, which facilitates combustion in temperatures between 750 and 800 °C. In this way, the formation of ash conglomerates with unburnt PVC recyclate granules on the boiler grate is eliminated, which prevents the blocking (slagging) of nozzles, providing the air onto the furnace grate by plasticized ash. With regard to the choice of the mentioned calcium sorbents, their application, similarly as in the case of hydrated lime, can be explained by their wide availability and relatively low cost, as well as by the fact that they could be readily used in their common commercial form, i.e., requiring no initial processing, that is, usually, proper grinding. Overall, the concept behind the research on the limitation of chlorine mobility was as follows: by supplementation of the model fuel with the PVC recyclate, it was possible to model various gram fractions of chlorine, which, in turn, translated into specific levels of hydrogen chloride emission. The measurements, which were conducted during the combustion of each of the mentioned fuels with different percentage addition of the sorbents, concerned the concentration of main gaseous products, including hydrogen chloride. Their chief purpose was to identify the trend of HCl concentration changes in the gaseous products of combustion, dependent upon the percentage of PVC added to fuel and, in result, the amount of chloride contained in the fuel as well as the percentage of the calcium sorbents (CaCO3 and CaCO3 3 MgCO3) added to the analyzed fuels. Let us note here that the data concerning the Ca(OH)2 sorbent are from ref 1.

3. Description and Results of the Experimental Research The principles guiding the experimental research and the interpretation of the results are identical as in the case of the 5854

Energy Fuels 2010, 24, 5851–5858

: DOI:10.1021/ef101048k

Poskrobko et al.

Table 4. Basic Technological Dimensionless Process Parameters: Molar Ratios (μ) calcium hydroxidea

dolomite

dolomite

number

Ca/Cl2 and Ca(OH)2/Cl2

Ca/Cl2 and CaCO3/Cl2

(Ca þ Mg)/Cl2 and CaMg(CO3)2/Cl2

1 2 3 4

2.85 4.18 5.58 8.32

Fuel I 2.13 3.14 4.15 6.16

2.50 3.68 4.87 7.24

5 6 7 8

1.15 2.26 3.35 4.48

Fuel II 0,86 1.68 2.59 3.32

1.01 1.97 2.94 3.90

9 10 11 12

1.07 2.12 3.14 4.21

Fuel III 0.80 1.57 2.39 3.11

0.94 1.84 2.75 3.66

13 14 15 16

1.14 2.25 3.34 4.48

Fuel IV 0.84 1.67 2.53 3.31

0.99 1.96 2.93 3.89

a

Figure 2. Relative hydrogen chloride concentration (φ) in wet flue gas, relative to the amount of added PVC: (2) experimental data and () interpolation and extrapolation data (sorbent = limestone).

each of the sorbents (Table 4), because Ca2þ þ 2Cl- = CaCl2 and Mg2þ þ 2Cl- = MgCl2. In practice, gram ratios could be used as well (Table 3). Expressing the used calcium compounds by elemental calcium or calcium and magnesium and by their oxides enables the comparison of reactivity of dolomite to reactivity of other calcium minerals, in this case, limestone and hydrated lime. At the same time, it sheds light on the effectiveness of using oxides for the purposes of reduction of chlorine mobility in the furnace chamber of the analyzed boiler. Relying on the data contained in Tables 2-4, several graphs illustrating dependencies between relative hydrogen chlorine concentration (φ) and the above-mentioned dimensionless parameters can be drawn. The graphical presentation of the most important research results does not include cases in which, regardless of the percentage amount of PVC added to fuel, chlorine mobility in the boiler furnace was not reduced by means of the sorbent addition. In such situations, the increase in hydrogen chloride concentration in the flue gas, in the function of either the added PVC or the gram ratio Cl/Ca representing the content of PVC or Cl in the fuel expressed by the fraction of the Ca element, is just as presented in ref 1. In the presentation of the most important research results, particular attention is devoted to the comparison of potential of the analyzed sorbents to reduce chlorine concentration in flue gas, which translates into the capability of reducing chlorine mobility in the furnace chamber. Figure 2 illustrates the fact that the reduction of hydrogen chloride emission (φ), at a constant percentage amount of PVC added to the fuel, with the growing amount of the added sorbent in the form of limestone serving as a parameter, is significantly lower than in the case of Ca(OH)2, which is confirmed by the trendline equations. Just like with calcium hydroxide, one is practically dealing with a linear growth of φ along with the increase in the percentage amount of added PVC recyclate. Furthermore, Figure 3, having analogous significance to Figure 2, proves that the reduction of φ by means of the dolomite sorbent is slightly inferior when compared to the reduction by limestone. In consequence, it appears that calcium hydroxide (hydrated lime) is the most efficient calcium sorbent. The corresponding differences in the potential of chlorine mobility reduction in the furnace chamber are illustrated by Figures 4 and 5.

The calcium hydroxide conditions apply to the study in ref 1.

of chlorine mobility reduction was evaluated. To eliminate gross errors and to approximate an unknown factual value of the emission, just as in measuring any physical value, each of the 32 combustion tests was repeated 3 times. The denotation of the HCl concentration in the flue gas, then, involved taking the total of 96 samples, 3 samples for each of the 32 analyzed fuels. The results of measurements and calculations are presented in Tables 2-4. The first of the discussed tables, apart from the percentage mass fractions of PVC and sorbent additions to the particular fuel, contains essential results of the research in the form of a relative hydrogen chloride concentration (φ) in the flue gas. Let us clearly state that φ is defined as the ratio of average hydrogen chloride concentration (HCl) in wet flue gas obtained from three independent readings during the combustion of a given fuel (see Table 1) to the average hydrogen chloride concentration (HCl)0 in wet flue gas obtained from three independent readings during the combustion of extracted rapeseed meal. All of the readings were taken in the chimneysection, as shown in Figure 1. In Tables 3 and 4, each of the fuels are associated with such dimensionless process parameters as (i) gram ratios, i.e., Ca/Cl and CaO/Cl for calcium hydroxide and limestone and (Ca þ Mg)/Cl and (CaO þ MgO)/ Cl for dolomite (Table 3), and (ii) molar ratios, i.e., Ca/Cl2 and Ca(OH)2/Cl2 for calcium hydroxide, Ca/Cl2 and CaCO3/Cl2 for limestone, and (Ca þ Mg)/Cl2 and (CaCO3 3 MgCO3)/Cl2 for dolomite (Table 4). It is generally agreed that these are the following factors (which are, by the way, chlorine input data to the system) that determine the effectiveness of chlorine mobility reduction: (i) the amount of added sorbent, measured by molar ratios, Ca(OH)2/Cl2, CaCO3/Cl2, or (CaCO3 3 MgCO3)/Cl2, and (ii) the amount of added sorbent expressed by the amount of added Ca, measured by the Ca/Cl2 molar ratio for calcium hydroxide and limestone, or expressed by the amount of added calcium and magnesium (Ca þ Mg) for dolomite, measured by the molar ratio (Ca þ Mg)/Cl2. These molar ratios are equal for 5855

Energy Fuels 2010, 24, 5851–5858

: DOI:10.1021/ef101048k

Poskrobko et al.

Figure 3. Relative hydrogen chloride concentration (φ) in wet flue gas, relative to the amount of added PVC: ([) experimental data and () interpolation and extrapolation data (sorbent = dolomite).

Figure 5. Relative hydrogen chloride concentration (φ) in wet flue gas, relative to the amount of added CaCO3 3 MgCO3 sorbent: (curve 1) 10% PVC, (curve 2) 8% PVC, (curve 3) 5% PVC, and (curve 4) 2% PVC.

Figure 6. Relative hydrogen chloride concentration (φ) in wet flue gas within the range of amounts of the sorbent molar ratio ( μ): (curve 1) calcium hydroxide, (curve 2) calcium carbonate, and (curve 3) dolomite.

Figure 4. Relative hydrogen chloride concentration (φ) in wet flue gas, relative to the amount of added CaCO3 sorbent: (curve 1) 10% PVC, (curve 2) 8% PVC, (curve 3) 5% PVC, and (curve 4) 2% PVC.

functions already for μ > 1.5. Such an approximation is more accurate the greater the μ value. Actually, it can be proven that, for μ > 2, that is, for values meaningful from the point of view of the engineering practice, the relation R2 > 0.85 is true. Figure 8 enables identification of the percentage addition of calcium sorbent necessary to dechlorinate the flue gas to the φ = 10 level. Obviously, equations describing the trendlines corresponding to other, a priori defined dechlorination levels can be easily formulated, and then accordingly, the necessary amounts of each of the calcium sorbents can be determined. On the other hand, assuming φ = φad, it can be resolved whether it is possible to limit the chlorine mobility in the furnace chamber so effectively as to keep the concentration of hydrogen chloride in the flue gas below the specified permissible level or whether additional dechlorination of the flue gas in an external installation is needed and to what extent it should be carried out. In this context, it should be remembered that the possible amount of sorbent added to fuel is limited because of the ensuing increase of ash content in the fuel, which modifies elemental gram fractions of combustible substances, including elemental carbon, and that in turn, results

This shows the sorbent consumption for fuels with various gram fractions of chlorine. The figures considered above relate the efficiency of flue gas dechlorination to the amount of added calcium sorbents (i.e., limestone and dolomite). The efficiency is expressed by the molar fractions (μ) (Table 4): CaCO3/Cl2 and (CaCO3 3 MgCO3)/Cl2, In this case, a parameter is the percentage mass addition of PVC to the fuel. The parameter remains in a linear relation with the gram fraction of chlorine in the fuel, as from Table 1, we find Cl ¼ 0:1741ð% PVCÞ þ 0:0016

ð1Þ

The conclusion concerning the relationship between the reductive potentials of the discussed calcium sorbents is wellillustrated by Figures 6 and 7, representing the relative concentration of hydrogen chloride (φ) in relation to the molar fractions (μ) (Table 4): Ca/Cl2 and (Ca þ Mg)/Cl2. Moreover, in Figure 7, it can be clearly seen that the φ = φ(μ) curves from Figure 6 can be approximated with linear 5856

Energy Fuels 2010, 24, 5851–5858

: DOI:10.1021/ef101048k

Poskrobko et al.

as well as calcium and magnesium carbonate (dolomite, CaCO3 3 MgCO3). The obtained results prove that, similarly as in the case when calcium hydroxide [Ca(OH)2] was adopted for the purpose, the desired trend of hydrogen chloride concentration change in the flue gas can be achieved, i.e., the one confirming the reduction of chlorine mobility in the furnace chamber of the boiler. It occurs that the trend corresponds with the analogous trend of HCl concentration changes in exhaust gas produced in laboratory combustion of refuse-derived fuel (RDF) and hard coal with various PVC additions.15-19 Accordingly, if the dechlorination of a flue gas is conducted by means of a calcium sorbent (CaCO3), the relation between the HCl concentration in the flue gas and the amount of the sorbent in the base fuel will be as illustrated in Figures 4 and 5.19 In comparison, if an expensive non-calcium sorbent (LiCO3) is used for the purpose, the same relation can be approximated linearly. For Cl occurring in fuels in the form of organic compounds, the HCl concentration grows linearly,15 just as shown in Figures 2 and 3. On the other hand, nonlinear changes in the HCl concentration were observed in the cases when chlorine occurred in the form of inorganic compounds. The laboratory research was conducted for the temperatures of 700-800 °C. Moreover, the experimental data also testify to the fact that the greatest reactivity, i.e., the greatest potential for dechlorinating flue gas in the furnace chamber among the analyzed sorbents, is exhibited by hydrated lime, whereas limestone reveals a lesser dechlorination capacity and dolomite reveals the least dechlorination ability. The highest efficacy of calcium hydroxide stems from the greatest content of the Ca element in a fixed mass amount of sorbent. As for the dolomite, it has to be kept in mind that its use can lead to a local process of ash softening (because of the magnesium content), also affecting fly ash and potentially causing the formation of high-temperature corrosion centers in steel structural components of the boiler. Let us note here, however, that, when applying dolomite with a purpose of bonding sulfur in coal-fired boiler furnaces, the above adverse processes do not occur, because sulfates are characterized by higher melting temperatures than chlorides. If the relative concentrations of hydrogen chloride (φ) in flue gas for the cases with the hydrated lime sorbent (φ[Ca(OH)2]) and the limestone sorbent (φ(CaCO3)) are compared, then their ratio (φ[Ca(OH)2]/φ(CaCO3)), considering the important from the practical point of view values of molar ratio (μ) (Figure 7), oscillates between 59 and 63%. At the same time, the (φ[Ca(OH)2]/φ(CaCO3 3 MgCO3)) ratio reaches values between 47 and 58% and the (φ(CaCO3)/ φ(CaCO3 3 MgCO3)) ratio reaches values from about 79 to 93%. Using dolomite, however, may require lowering the

Figure 7. Linear approximation of relative hydrogen chloride concentration (φ) in wet flue gas within the range of practically applicable amounts of the sorbent molar ratio (μ): (curve 1) calcium hydroxide, (curve 2) calcium carbonate, and (curve 3) dolomite.

Figure 8. Amounts of added sorbents corresponding with φ = 10, relative to the amount of added PVC.

in the decrease of the calorific value of the fuel, having an adverse impact on the efficiency of the boiler. The excessive addition of sorbent might also cause a substantial growth of dust emission into the environment. In such circumstances, modernization of dust removal systems may be required. Once more, the amount of ash and fly ash collected by filters will rise, which is sure to effect a diminished furnace efficiency. Finally, apart from the mentioned negative consequences of an excessive sorbent addition, there are solid arguments of purely economic nature advocating careful admixture of the sorbent.

(15) Wey, M. Y.; Liu, K. Y.; Yu, W. J.; Lin, C. L.; Chang, F. Y. Influences of chlorine content on emission of HCl and organic compounds in waste incineration using fluidized beds. Waste Manage. 2008, 28 (2), 406–415. (16) Zhu, H. M.; Jiang, X. G.; Yan, J. H.; Chi, Y.; Cen, K. F. TG-FTIR analysis of PVC thermal degradation and HCl removal. J. Anal. Appl. Pyrolisis 2008, 82 (1), 1–9. (17) Chiang, K. Y.; Jih, J. C.; Lin, K. L. The effects of calcium hydroxide on hydrogen chloride emission characteristics during a simulated densified refuse-derived fuel combustion process. J. Hazard. Mater. 2008, 157 (1), 170–178. (18) Zhang, C.; Wang, Y.; Yang, Z.; Xu, M. Chlorine emission and dechlorination in co-firing coal and the residue from hydrochloric acid hydrolysis of Discorea zingiberensis. Fuel 2006, 85 (14-15), 2034–2040. (19) Zhu, S.; Zhang, Y.; Zhang, Y.; Zhang, C. Effect of CaCO3/ LiCO3 on the HCI generation of PVC during combustion. Polym. Test. 2003, 22, 539–543.

4. Remarks and Conclusions The limiting of chlorine mobility in the furnace chamber, resulting in the reduction of the hydrogen chloride concentration in the flue gas, was carried out by adding calcium sorbents to the basic fuel, that is, extracted rapeseed meal supplemented with PVC recyclate to increase the gram fraction of chlorine. The described experimental research, which was realized in semi-technical conditions with the use of a 225 kW water-underfeed stoker-fired boiler, employed such calcium sorbents as calcium carbonate (limestone, CaCO3) 5857

Energy Fuels 2010, 24, 5851–5858

: DOI:10.1021/ef101048k

Poskrobko et al.

temperature of the process, because of the possible formation of ash and slag conglomerates resulting from the presence of magnesium chloride on the grate of the boiler. The characteristics of flue gas dechlorination in the furnace chamber in relation to the percentage addition of analyzed calcium sorbents are of vital practical importance, because they allow for administering the optimum amount of the selected sorbent. Used in excessive amounts for dechlorination and desulfurization of flue gas, the sorbent changes the quality of ash and increases the volume of solid combustion remaining. An analogous situation, i.e., a change in the quality of ash, arises in the case of co-combustion of biomass with coal, at the same time causing considerable problems in high-power boilers. Naturally, also economic issues are of significance, i.e., the cost of dechlorination and desulfurization of flue gas. It seems that the same sorbents as for chlorine may be used for other halogens (e.g., bromine and fluorine). In the cases when φ > φad, the necessity arises to dechlorinate flue gas in an installation outside the boiler. In practice, this aim can be achieved by such means as (i) dry cleaning of flue gas with the use of Ca(OH)2 and powdered activated carbon, (ii) semi-dry cleaning of flue gas with the use of Ca(OH)2 and water vapor (processes taking place on a wet cake of a bag filter), (iii) wet cleaning of flue gas with the use of scrubbers having milk of lime as a working medium, (iv) wet cleaning of flue gas with the use of scrubbers having NaOH solution as a working medium, (v) semi-dry with Ca(OH)2 combined with wet NaOH cleaning of flue gas, and (vi) wet cleaning of flue gas with the use of scrubbers having sodium carbonate solution as a working medium. Whenever there is a need, there is also a possibility of coupling these methods, remembering that, in such situations, the calcium-based method should be applied in the first order. Accordingly, the HCl emission can be reduced to a φ < φad value. The bonding of hydrogen chloride in the furnace chamber of a boiler effectively prevents the formation of highly toxic compounds, PCDDs and PCDFs (in secondary reactions). Hydrogen chloride is a precursor of the creation of these congeners, which is of special significance for the processes conducted in temperatures below 1100 °C restraining the decomposition of the compounds. This fact testifies to the significance of the presented results of experimental research conducted in semi-technical conditions. Ultimately, it also means that high-temperature corrosion processes can be effectively kept at bay and, at the same time, the lowering of the availability of the boilers can be prevented. Thus, the conducted study can be perceived as a part of activities aimed at adapting power boilers for combustion of nonforest biomass or for co-combustion of such biomass with fossil fuels or industrial and municipal wastes. As a result, expansion of the biomass assortment available for power industry is becoming more and more real. As for the dechlorination of flue gas by means of calcium sorbents directly on the furnace grate, it protects the steel structural elements of the grate from chloride corrosion. Such protection has its importance, for instance, when firing fuels formed from waste in stoker-fired waste-heat boilers, because their convective section is usually built from ceramic lining. Exchanger steel elements are situated in the upper part of the first pass and in the second pass. Accordingly, their corrosion

protection may be realized by the blowing of powdered calcium sorbents or ammonium sulfate (the ChlorOut method) into the convective section. In these cases, combined methods of dechlorination can be used for the purpose, namely, on the grate with, for example, such sorbents as limestone or dolomite and in the convective section with the use of adequately lower amounts of more expensive sorbents, such as hydrated lime. It has to be emphasized that during the experiments there have not been observed any cases of slag conglomerate formation, which is a certain proof that, when fuel contains a requisite gram fraction of sulfur that forms high-melting MgSO4, then to dechlorinate flue gas in the furnace, it is usually sufficient to use dolomite. In the described circumstances, the gram fraction of sulfur in the fuel without the PVC addition was relatively high, having amounted to 1.5%. All in all, the achieved results suggest that calcium sorbents can be successfully used for the purpose of reducing the mobility of chlorine released in the form of hydrogen chloride in the processes of combustion of fuels formed from waste. Possible complications in boilers operating arise in the cases when the fuels in question (or their components) are formed from biomass of agricultural origin. If such biomass contains substantial amounts of potassium and chlorine, the formation of high-temperature corrosion centers is facilitated. In this situation, calcium preparations are of limited use and the application of dolomite is not recommended at all (because of the presence of magnesium, which lowers the temperature of ash softening). This is still one of the biggest issues in the area of research aimed at increasing the scale of efficient use of biomass in the power industry. Acknowledgment. This work is supported by the Ministry of Science and Higher Education, Poland (Grant R06 018 02).

Nomenclature C = gram fraction of carbon (%) Cl = gram fraction of chlorine (%) H = gram fraction of hydrogen (%) (HCl) = average hydrogen chloride concentration in wet flue gas obtained from three independent readings during the combustion of a given fuel (mg/dm3) (HCl)ad = admissible hydrogen chloride concentration in wet flue gas (mg/dm3) (HCl)0 = average hydrogen chloride concentration in wet flue gas obtained from three independent readings during the combustion of extracted rapeseed meal (mg/dm3) LHV = lower heating value (kJ/kg) O = gram fraction of oxygen (closing balance of gram fractions of elements) (%) N = gram fraction of nitrogen (%) R2 = correlation coefficient S = gram fraction of sulfur φ = (HCl)/(HCl)0, relative hydrogen chloride concentration in wet flue gas φad = (HCl)ad/(HCl)0, admissible relative hydrogen chloride concentration in wet flue gas η = boiler furnace efficiency λ = excess air number μ = molar ratio

5858