Environ. Sci. Technol. 2002, 36, 5498-5503
Nitric Oxide Reduction in Coal Combustion: Role of Char Surface Complexes in Heterogeneous Reactions ANA ARENILLAS, F E R N A N D O R U B I E R A , A N D J O S EÄ J . P I S * Instituto Nacional del Carbo´n, CSIC, Apartado 73, 33080 Oviedo, Spain
Nitrogen oxides are one of the major environmental problems arising from fossil fuel combustion. Coal char is relatively rich in nitrogen, and so this is an important source of nitrogen oxides during coal combustion. However, due to its carbonaceous nature, char can also reduce NO through heterogeneous reduction. The objectives of this work were on one hand to compare NO emissions from coal combustion in two different types of equipment and on the other hand to study the influence of char surface chemistry on NO reduction. A series of combustion tests were carried out in two different scale devices: a thermogravimetric analyzer coupled to a mass spectrometer and an FTIR (TG-MS-FTIR) and a fluidized bed reactor with an on line battery of analyzers. The TG-MS-FTIR system was also used to perform a specific study on NO heterogeneous reduction reactions using chars with different surface chemistry. According to the results obtained, it can be said that the TG-MS-FTIR system provides valuable information about NO heterogeneous reduction and it can give good trends of the behavior in other combustion equipments (i.e., fluidized bed combustors). It has been also pointed out that NO-char interaction depends to a large extent on temperature. In the low-temperature range (800 °C), a different mechanism is involved in NO heterogeneous reduction, the nature of the carbon matrix being a key factor.
Introduction Although there are important natural sources of NOx formation, fossil fuel combustion in industry and the transport sector are major sources of NOx emission (1). It is generally accepted that the amount of NO and N2O emitted from coal combustion is the result of homogeneous and heterogeneous formation and in situ destruction reactions. The main source of the final NO and N2O emissions is fuel-bound nitrogen (2). Previous studies on coal devolatilization at different final temperatures and heating rates have shown that a high proportion of the fuel nitrogen remains in the char (3, 4). Thus, coal char is rich in nitrogen and is the main contributor to NO in low-NOx burners and under fluidized bed combustion conditions (2, 5). However, coal char is not only a source * Corresponding author. Telephone: +34 985 11 89 71. Fax: +34 985 29 76 62. E-mail:
[email protected]. 5498
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 24, 2002
of NO. Due to its carbonaceous nature, coal char can also reduce NO through heterogeneous reduction reactions on its surface. During the devolatilization step in the overall coal combustion process, the nitrogen in coal is split into volatile N and char N. Afterward, the formation and reduction of NO and N2O occur through homogeneous reactions in the gas phase and heterogeneous reactions involving the char. The homogeneous reactions have been the subject of a greater number of studies and these days they are well controlled in new combustion technologies. The heterogeneous mechanisms, however, are still not completely understood, and further work is needed in order to minimize current NO emissions. The reactions of NO with carbonaceous materials have been studied by different authors (6-8). However, these reactions are not as well understood as other gas-carbon reactions, such as O2, CO2, or steam with carbon. The overall reduction process can be represented by reaction 1, but the
C* + NO f N2 + COx
and/or
N2O + COx
(1)
detailed mechanism and the nature of the active sites involved have still not been completely elucidated. It is generally agreed that the first step in the heterogeneous reduction of NO by carbon is chemisorption on the carbon surface. Following the chemisorption step, a series of reactions occur, where different surface compounds intervene to produce N2, N2O, CO, and CO2. Although it is now recognized that carbon plays a significant role in the reduction of NO during combustion, the nature of the surface compounds involved is still unclear. Several mechanisms have been proposed to explain the heterogeneous reduction of NO. Some authors (9) have suggested that NO is adsorbed on the char surface to form NO complexes, C(NO). This process may be reversible or irreversible. According to other works, this adsorption is mainly associative (1, 4):
C(NO) + C(N) f N2O + 2C*
(2)
C(NO) + C(NO) f N2 + CO2 + C*
(3)
At high temperature, the mobility of surface nitrogen species increases and the probability of two nitrogen atoms forming an NtN bond is greater. A dissociative adsorption of NO has also been proposed (10), and this process results in the formation of surface complexes that form N2 and CO2 after the desorption step:
4C* + 2NO f 2C(N) + 2C(O) f N2 + CO2 + 3C*
(4)
There have also been several studies on the effectiveness of carbon as a support for different metals in the catalytic heterogeneous reduction of NO (11-13). However, this kind of NO reduction should be catalogued as flue gas cleaning (i.e., SCR). Traditionally, the nitrogen released during devolatilization has received more attention than that released during char oxidation, since it was reported to be the major contributor to NO and to be more amenable to control by modification of combustion zone aerodynamics and by air staging (5, 14). Therefore, as the NO from volatiles has been controlled through the adoption of low-NOx burners and staged combustion, the role of char in overall NO emissions should be examined in greater detail. Accordingly, knowledge of the reactions that take place in the formation/reduction of NO 10.1021/es0208198 CCC: $22.00
2002 American Chemical Society Published on Web 11/16/2002
TABLE 1. Proximate and Ultimate Analyses of the Coals Studied (C, W, G) and Char W150 proximate analysis (wt % db)
ultimate analysis (wt % daf)
sample
volatile matter
ash
C
H
N
S
Oa
C W G W150
36.7 17.0 4.2 1.7
7.1 6.2 9.3 8.0
84.5 89.2 94.4 95.8
5.5 4.6 2.1 0.9
1.7 1.2 1.0 1.4
1.5 0.7 0.9 0.5
6.8 4.3 1.6 1.4
a
Calculated by difference.
involving coal char is of paramount importance. In this work, a study of NO reduction was performed from two different approaches. On one hand, the combustion experiments in the TG-MS-FTIR system provided valuable information about NO reduction during combustion and give good trends of behavior in other combustion systems at a different scale (i.e., fluidized bed combustors). On the other hand, and due to the scarce information of the factors that affect the NO heterogeneous reduction reactions, specific tests were also performed in the TG-MS-FTIR system, to study the influence of char surface chemistry on the NO heterogeneous reduction.
Experimental Section This work can be divided into two different sections. The first one dealing with coal combustion reactions and how the tests performed in the TG-MS-FTIR system can give good trends of the behavior in other combustion devices (i.e., fluidized bed combustors). Thus, it would be possible to compare the results of the TG-MS-FTIR system under more realistic conditionssin terms of fluid dynamics and heating ratesoperating in fluidized bed combustion, a technology in which heterogeneous reactions involving coal char play an important role. The second section of the paper deals with the specific NO heterogeneous reduction reactions, which are the ones less clarified and thus less controlled. This implies the necessity of further studies on this topic. Thus, in this work, a study of the influence of the char surface chemistry in these kinds of reactions is presented. Materials. For the combustion tests, either in the thermogravimetric analyzer or in the fluidized bed reactor, three coals of different rank and origin were used. These included a high-volatile bituminous coal (Camocha, C), from the north of Spain, a low-volatile bituminous coal (Welch, W), from the United States, and a Spanish anthracite (Gillon, G). The ultimate and proximate analyses of these coals are presented in Table 1. The chars used in the specific study of the NO heterogeneous reactions were all prepared from the low-volatile bituminous coal (W). Pyrolysis was carried out in a fixed bed quartz reactor (i.d. 30 mm), under nitrogen flow, up to a final temperature of 850 °C, and a soaking time of 1 h. The char obtained was denoted as W150. The main characteristics of char W150 are also given in Table 1. The textural characterization of the W150 char was carried out by measuring true (He) and apparent (Hg) densities and mercury porosimetry. The following pore volume distribution of char W150 was obtained: 60% macropores, 5% mesopores, and 35% micropores. Adsorption isotherms in CO2 at 0 °C were also performed, and char W150 presented a total surface area of 243 m2 g-1. Active surface area values (ASA) were determined by a gravimetric method (15, 16), and an ASA value of 8 m2 g-1 was obtained for char W150. Procedure. A series of experiments were carried out in a Setaram TGA 92 thermogravimetric analyzer (TG). For the coal combustion tests, 25 mg of sample was heated from
room temperature to 800, 850, and 900 °C at 15 °C min-1, in 20% oxygen in argon. On the other hand, for the temperatureprogrammed reduction (TPR) experiments, 25 mg of sample was placed over a bed of powdered alumina, to facilitate contact between the sample and the reactive gas. The temperature was raised from room temperature to 1100 °C at 15 °C min-1 under a flow (50 mL min-1) of 400 ppm NO in Ar. A quadrupole mass spectrometer (MS) and an FTIR placed in parallel and linked to the thermobalance were used to analyze the gases evolved during both the combustion and TPR experiments. The ionization of the analyzed gas in the MS was performed by an axial beam ion source (100 eV). A Faraday collector was used to detect the ions, separated according to their mass-to-charge ratio (m/z). In the collection of the IR spectra, the purge gas was dry, CO2-free air. The resolution of the FTIR was set to 4 cm-1 on a continuous basis. Gas evolved from the TG entered the MS and FTIR through a heated (200 °C) stainless steel capillary. The gas cell in the FTIR was also heated to 225 °C. To avoid undesirable secondary reactions, the probe was placed very close (∼3 mm) to the sample crucible of the thermobalance, in the direction of the gas flow. To eliminate cold points and gas condensation in the connecting line, the bottom of the thermoanalyzer was also heated to 200 °C. Although all the analyses performed in this work were qualitative, the gaseous compounds’ evolution profiles should be repeatable for an effective comparison of intensity peak areas between different samples, i.e., semiquantitative analysis. Optimization of the coupling system and the normalization procedure of the MS profiles have been described in a previous work (17). The species were monitored in the MS at m/z 14 (N+, N22+), 28 (N2+ and 12CO+), 29 (13CO+), 30 (NO+ and 12C18O+), 44 (CO2+ and N2O+), and 46 (C16O18O+) (3, 4, 18), among other species. The N22+ and NO+ profiles in the MS were corrected for contributions from CO2+ and 12C18O+, respectively, using a method described in the literature (18). A series of combustion tests were also performed in a fluidized bed reactor, to check the applicability of the results obtained in the TG-MS-FTIR tests. The experimental procedure involved the injection of batches (0.5 g) of the three coals into a 52-mm-internal diameter stainless steel fluidized bed reactor, electrically heated (23). Silica sand (d ) 0.3 mm) was used as bed material and air as fluidizing gas (U ) 0.16 m s-1). The air was preheated before it entered the reactor. The exhaust gases were followed by a battery of specific analyzers for CO, CO2, N2O, and NOx (NO + NO2). Different combustion temperatures were used (800, 850, and 900 °C) in order to study the influence of operating temperature on the formation of nitrogen compounds. During the combustion tests in the TG-MS-FTIR system, the m/z 44 signal from the MS can mainly be assigned to CO2, which is the main gaseous product due to the oxidant atmosphere. However, in the temperature-programmed reduction experiments, no oxygen was present in the reacting gases. Figure 1 shows the N2O+ (m/z 44) and C+ (m/z 12) profiles of char W150 during a temperature-programmed reduction experiment (400 ppm NO in Ar) and during a blank experiment (Ar). As can be seen in Figure 1, during the TPR test, signal m/z 44 presents different peaks. The m/z 12 signal assigned to C+ does not present any appreciable peak unlike N2O+. This implies that m/z 44 can be assigned to N2O as the main product, although some CO2 may be formed as a subproduct of the heterogeneous reduction (6, 19, 20). These contributions, however, can be neglected according to the results displayed in Figure 1. VOL. 36, NO. 24, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5499
FIGURE 1. Comparison of the evolution profiles of m/z 44 (N2O+) and 12 (C+) followed by the mass spectrometer during the temperatureprogrammed reduction tests (400 ppm NO in Ar) and the blank experiment (Ar). The TPR experiments were focused in the temperature range 20-1000 °C, because at higher temperatures gasification of the char becomes significant, and a high intensity in the m/z 12 and 44 signals was then detected. In this study, special attention was paid to the formation of N2O, which is especially important in fluidized bed coal combustion due to the relatively low operation temperatures (800-950 °C) used in this technology. As can be seen in Figure 1, the main peak temperature in the N2O profile occurs at ∼850 °C, under the experimental conditions used in the TPR tests of this work. Several researchers have reported that the presence of O2 is essential for N2O formation during coal char-NO reactions (2, 21). This has been explained by assigning to oxygen the role of freeing the internal C(N) species so that they can react with NO to form N2O. Taking into account the molecular similarity between O2 and NO, it is doubtful whether the ability to break-up char structure can be attributed just to O2 (22). It seems that the higher reactivity of O2 in comparison with that of NO, and thus its greater facility to break-up structures, has to be the reason for the reaction of coal charNO being enhanced in the presence of oxygen. However, this reaction also occurs when there is no oxygen in the reactive gases. This has been corroborated with the results obtained in this work.
Results and Discussion Interpretation of the Evolved Gas Analysis Data. Figure 2 shows the m/z 30 (NO+), 28 (N2+ + CO+) and 44 (N2O+ + CO2+) evolution profiles from MS and those of CO and N2O followed by FTIR for a TPR experiment involving char W150. These profiles indicate the formation or destruction of a gaseous compound from the initial reactive gas (400 ppm NO in Ar). This means that NO disappears (negative units) while the other compounds are formed (positive units). Nevertheless, it has to be pointed out that the relative intensity of one compound cannot be compared with the relative intensity of another one, as each compound has their own response factor in MS analysis. However, a comparison between the relative intensity of one compound in one sample with that of the same compound in another sample can be made; this is the so-called semiquantitative analysis (3, 4, 17, 18, 23). As was mentioned before, there are some m/z signals with possible contributions from more than one compound (i.e., m/z 28 and 44). The incorporation of an FTIR in parallel to the MS improves the detection of the gaseous compounds evolved, as the FTIR can distinguish N2O from CO2 and can 5500
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 24, 2002
FIGURE 2. Evolution profiles of the main compounds followed during the TPR test involving char W150. detect CO without any interference of N2. The results obtained from FTIR are also presented in Figure 2. It can be observed that the N2O profile presents a maximum peak at ∼350 °C, coinciding with the disappearance of NO. From the comparison of FTIR and MS profiles, it can be concluded that only between 450 and 550 °C is the m/z 44 profile in the MS mainly due to CO2 evolution, which coincides with the no disappearance of the nitric oxide. Figure 2 shows that the maximum peak temperatures for N2O emission, in both MS and FTIR profiles, coincide with the minimum peak temperatures for NO. This is clear evidence that the main species of the m/z 44 signal is N2O+. It can also be seen that the N2O+ profile presents different peaks at around 350, 650, and 850 °C. The NO+ profile presents an extra minimum at ∼950 °C, while no maximum was detected for N2O+. This indicates that there is no N2O at such a high temperature and the only nitrogen compound produced is N2. No other N species were detected (i.e., HCN) during the TPR experiments. The m/z 28 signal was detected with a maximum peak around this high temperature (i.e., 950 °C). The FTIR detected the formation of CO around these temperatures. However, although the shape of the NO profile from MS (m/z 30) and FTIR coincides, the shape of CO from FTIR and m/z 28 from the MS differs in the sense that the maximum production of CO is achieved at ∼850 °C (see FTIR profile in Figure 2) and the maximum of m/z 28 from the MS is reached at 950 °C with a shoulder at 850 °C. This shows that although CO is formed, there is a clear contribution of N2 formation in the MS profile. It can also be seen in Figure 2 that the N2O profile retrieves the baseline at ∼1000 °C, while this does not occur in the case of the NO and N2 profile. Other authors have observed that the reaction between nitrogen oxides and carbon generates nitrogen-containing complexes on the carbon surface. The use of X-ray photoelectron spectroscopy (XPS) has made the characterization of some of these surface complexes possible. Thus, XPS analysis revealed the formation of pyridinic, pyrrolic, quaternary nitrogen and C(NO) species on the carbon surface after C-NO reaction (1, 6). Therefore, the possibility that the NO profile does not retrieve the baseline as a consequence of some nitrogen being retained in the char after the C-NO reaction is highly probable. The NO profile in Figure 2 also presents a peak (RI positive) at low temperature. This could be due to the reversible adsorption of NO that takes place at room temperature, the
FIGURE 4. N2O emission during combustion tests in the fluidized bed reactor for the coals studied (C, W, G), expressed in area units (a.u.) of the evolution profiles. FIGURE 3. NO emission during combustion tests in TG-MS-FTIR and in the fluidized bed reactor (FBR) for the coals studied (C, W, G), expressed in area units (a.u.) of the evolution profiles. NO being desorbed as the temperature increases and reaches a maximum at ∼200 °C. Coal Combustion Tests. The coal combustion tests were performed in two different devices, at different scale: a thermogravimetric analyzer and a fluidized bed combustor. The NO and N2O emissions detected during the experiments come from the overall combustion reaction, i.e., volatiles and char combustion. With regard to the combustion tests performed in the thermogravimetric analyzer, the NO emission values for the three coals studied in this work (C, W, G) were obtained. It needs to be pointed out that although no quantitative analysis was performed in this work, the repeatability of the analysis and the normalization procedure applied in this work allow semiquantitative analysis of the gaseous compound emissions to be made by calculating peak area values (17). Figure 3 shows the trend of NO emissions in the combustion tests, the values being expressed in area units (a.u.). The nitrogen content and the amount of sample were included in the NO conversion values. The results obtained from the TG-MS-FTIR tests provide valuable information to study nitrogen oxide formation in other combustion devices (i.e., fluidized bed combustor). The combustion tests performed in the fluidized bed reactor used in this study give similar trends for NO emission than the combustion tests in the TG-MS-FTIR system, as can be seen in Figure 3. An increasing trend of NO emission with temperature can be observed in this figure, although it is not a great increase, especially when considering coal C. A significant dependence of NO emission with coal rank can be also observed. The lower the coal rank, the higher the nitrogen conversion to NO (C > W > G). On the other hand, Figure 4 shows the N2O emission for the fluidized bed combustion tests of the three selected coals. It can be observed that between 800 and 900 °C there is a decreasing trend in N2O formation. The effect of coal rank can be observed on the amount of N2O released, following the same trend already mentioned for the NO emission: as coal rank increases, the reactivity decreases, and thus the nitrogen conversion to nitrogen oxides also decreases. According to the results obtained, the basic studies performed in the TG-MS-FTIR system provides valuable information and good trends for other kinds of reactors at a different scale (i.e., fluidized bed reactor). Role of Char Surface Chemistry in NO Heterogeneous Reduction. According to the results presented above, char
properties and especially surface chemistry are key factors in the NO heterogeneous reduction. Due to the advantages of the TG-MS-FTIR system, a series of experiments were programmed in order to perform a specific study on the influence of different char surface chemical properties on the NO reduction. Char W150 was used to obtain a series of samples with different surface chemistry: char without surface complexes (W150*), char with oxygen surface complexes (W150-O), and char with free carbon active sites (W150-C). It has to be pointed out that these four samples (i.e., W150, W150*, W150O, and W150-C) have the same carbon matrix and the treatments they suffered in order to modify the chars surface chemistry do not alter their physical properties (texture or structure) appreciably, or at least not to mask the influence of the surface chemistry on the NO heterogeneous reduction. Char W150* was obtained by treating the W150 char at 850 °C for 6 h under Ar flow in order to clean the char surface; then the TPR test was performed. Char W150-O was obtained by treating the initial W150 char at 850 °C for 6 h under Ar atmosphere; the temperature was then lowered to 150 °C, and the char was exposed to an oxygen flow for 17 h in order to increase the amount of oxygen surface complexes. The oxygen was swept from the chamber before the TPR test. Char W150-C was obtained by cleaning the char surface (850 °C and 6 h under Ar flow), oxygen chemisorption was then conducted (17 h under O2 flow at 150 °C), and finally, desorption of the oxygen complexes was accomplished by heating the char sample under Ar for 6 h at 850 °C. The N2O evolution profiles obtained after the TPR tests for chars W150, W150*, W150-O, and W150-C are presented in Figure 5. This figure shows a shift of the main maximum peak of the N2O+ profiles of the modified chars (i.e., W150*, W150-O, and W150-C) to higher temperatures in comparison with char W150. This behavior could be due to the initial cleaning step performed during the char production process. It seems that, during TPR tests, surface complexes have an effect similar to that of ignition due to the presence of volatile matter in coal combustion, enabling the process to take place at lower temperatures. The maximum N2O peak in the profiles of the three modified chars appears at the same temperature (900 °C), possibly due to a common mechanism of N2O formation. The main differences between the profiles are found in the low-temperature range (
