**3. Review of studies on formation of nitrogen compounds during coal combustion**

Reduction of carbon dioxide and nitrogen oxide emissions in pressurised combustion methods is a result of intensive research (Gajewski, 1996). Release of fuel-nitrogen from coal during pressurised combustion, the rate and degree of its conversion to NO and N2O are the data necessary for understanding of formation and destruction mechanisms for these pollutants.

The published studies on the mechanism of fuel-nitrogen release show a lot of discrepancies and contradictions. In case of pressurised combustion, there are no comprehensive kinetic models of NO and N2O formation available. Few papers, mentioned below, require further investigations on modelling of nitrogen conversion in pressurised combustion processes.

Weiszrock et al. (1997) conducted studies on pressurised coal combustion in a laser reactor. At various O2 levels, the increase in pressure resulted in NO emission reduction, but N2O emissions increased.

Aho et al. (1995) as well as Aho and Pirkonen (1995) demonstrated reduction of conversion into N2O, which is contradictory to the findings of Weiszrock et al. (1997), while Lu et al. (1992) did not find any effects of the combustion pressure on nitrogen conversion into N2O.

Croiset et al. (1996, 1998) proposed a model of NO and N2O formation during pressurised combustion of char. The model was based on eight surface reactions with [C], [CO], [CN] and [CNO] active centres. Modelling of nitrogen oxide emissions using kinetic constants calculated by the authors demonstrates reduction of NO levels with enhanced pressure. Croiset et al. noted that at increased temperature and pressure, reduction of NO to N2O is a more important reaction than reduction of NO to N2. In their calculations, the authors of the model omitted the inner structure and made the reaction kinetic constants dependant on pressure, which arouses some doubts.

For 0.2 MPa, De Soete et al. (1999) complemented the model with the efficiency of particle volume application in the reaction of [CO] and {CNO] active centre formation. They also compared kinetic coefficients of individual reactions published by Croiset et al. (1996, 1998) for "Westerholt" coal (at 0.2 MPa) and by de Soete (1990) for "Prosper" coal (at 0.1 MPa), but they did not make any comments on kinetic coefficients for higher pressures presented by Croiset et al. (1998).

Tomeczek and Gil (2000, 2001) as well as Gil (2000) proposed their model of NO and N2O formation and reduction in pressurised char combustion based on chemical reactions presented by de Soete (1990) and Croiset et al. (1998). They determined their own kinetic coefficient of NO reduction over char where the activation energy of char was 79 kJ/mol within the pressure range of 0.2 MPa to 1.0 MPa. In case of the other coefficients of heterogeneous reactions, they considered the efficiency of particle volume application and made the coefficients non-pressure dependant, which resulted in a good agreement of the model and the conducted experiments within a large pressure range (0.2–1.5 MPa). Moreover, they estimated a total concentration of active centres (not determined yet) and included it in kinetic equations. However, the model requires further investigations and experimental studies.

In case of NO and N2O modelling, the initial pathway of fuel-nitrogen conversion into NO during coal combustion is extremely difficult to explain. During char coal combustion, the

Fuel-N Conversion to NO, N2O and N2 During Coal Combustion 43

boundary active centre [CN] to form [N] which subsequently may react with O2 to produce NO or with NO to produce N2O, which was also suggested by Krammer and Sarofim (1994). A disadvantage of this model is that the experiments with one coal type were subsequently

In available literature, there are also some other models of nitrogen oxide formation and

Aria et al. (1986, 1986) presented a model where fuel-nitrogen was assumed to release as NH and NO in a heterogeneous reaction with O2, and NO reduction was assumed to occur in a homogenous reaction in a boundary layer of a particle. Although the combustion process was assumed to occur on the external layer of a particle and in its pores, porosity changes were not made dependant on a particle burn-out level, which is suggested by Visona and Stanmore (1996, 1999). Also, the preparation method of char which was subsequently chilled

Tullin et al. (1993) assumed that NO is produced as a result of a heterogeneous reaction of oxygen with char-nitrogen, while N2O is formed in a heterogeneous reaction of NO with fuel-nitrogen, and both compounds are reduced during a heterogeneous reaction with elemental carbon C. The authors found that N2O decreased with the increase in the particle burn-out level, which is contradictory to the observations made by de Soete (1990) and Croiset et al. (1998) who showed that the increase in N2O emissions was proportional to the coal burn-out level. Tullin et al. (1993) as well as Goel et al. (1994) verified the proposed

Sarofim et al. (1995) described the porosity effects on NO and N2O release from fuelnitrogen. They observed a dependency on the surface area of reaction and diffusion in pores during conversion of fuel-nitrogen into NO and N2O. Moreover, they showed that the conversion degree of fuel-nitrogen is proportional to the particle burn-out. In their investigations, the authors ignored a large fraction of mesopores involved in the reaction and they entirely neglected macropores whose role in diffusion of gases is significant.

Despite considerable progress, understanding of the mechanisms of nitrogen compound conversions during coal combustion is still unsatisfactory (Benson, 1968; Anthony el al., 1976; Pottigisser, 1980; Bliek el al., 1985). There are no reliable data on the kinetics of nitrogen conversion at increased pressures. In view of the development of pressurised combustion techniques, there is an urgent need of detailed studies on the issue. In the published studies on the mechanism of fuel-nitrogen release and its conversion in a particle, measuring points (mostly recorded under isothermal conditions when the final heating temperature was reached) were applied. Experiments where samples are analysed during heating allow for more precise determination of kinetic constants because for individual reactions, it is possible to separate temperature and time dependencies. A majority of researches agree that formation of nitrogen oxides should be described on the basis of internal diffusion and a chemical reaction in pores. In case of pressurised combustion, there are no comprehensive analyses in the form of kinetic equations (Mallet, 1995) so the issue requires systematic studies on nitrogen conversion in pressurised combustion processes. A literature review shows that while determining kinetic constants of fuel-nitrogen conversion into nitrogen oxides, individual

characteristics of investigated char porous structures should be considered.

reduction described that were not compared in the paper by Molin et al. (2000).

used for determination of kinetic constants.

arouses some doubts.

model for one coal type.

problem seems less difficult, but the basic questions remain. It should be noted that experimental results of NO emissions during char combustion may be modelled in various ways, which means a substantial uncertainty of understanding of NO formation from char-nitrogen.

Visona and Stanmore (1996, 1999) showed that while modelling NO emissions, comparable results may be obtained through various pathways of fuel-nitrogen conversion:


The researchers found that both pathways of fuel-nitrogen release are possible at about 1750 K.

Molina et al. (2000) compared models of coal particle combustion and pathways of fuelnitrogen conversion into nitrogen oxides presented by Wendt and Schulze (1976), Shimizu et al. (1992), Goel et al. (1994), Visona and Stanmore (1996) and Soete et al. (1999).

The models proposed by Visona and Stanmore (1996) as well as by de Soete et al. (1999) are described above. A feature of the model presented by Wendt and Schulze (1976) is carbon oxidation to CO and char-nitrogen oxidation to NO in pores of heterogeneous particles. During diffusion towards the outer surface of the particle, CO undergoes homogenous oxidation to CO2, and NO homogenous reduction occurs which was assumed according to the Zeldowicz opposing reaction (1946). The mechanism is based on a reburning phenomenon that occurs in particle pores. An important factor for the mechanism is kinetics of CO oxidation adopted from Howard et al. (1973). The diameterlength ratio of the pores was demonstrated to determine a degree of char-N conversion into NO at about 1000 K. At temperatures far above 1000 K, oxygen is basically consumed at pore entrances so diffusion into pores becomes less significant and nitrogen conversion depends on phenomena occurring in the pore boundary layer. In this case, it was demonstrated that CO concentration in the pore neighbourhood is a key factor in the nitrogen conversion into NO. The authors did not perform any experimental verification of the developed mathematical model. Within the investigated combustion range, char strongly depends on initial coal; thus, in a mathematical model aimed at predicting NO concentrations, such factors as individual char characteristics and its changes during particle burning should be considedred.

Shimizu et al. (1992) presented the simplest model of nitric oxide emissions comprising one combustion reaction and two NO formation and reduction reactions. In coal particles, oxygen directly reacts with fuel nitrogen to produce NO; then it is reduced to N2 in a homogenous reaction with CO. Chan et al. (1983) and Goel et al. (1994) are the only researchers to include also the reaction of NO with CO catalysed by char. Studies by Chan et al. (1983) revealed that in the presence of CO, the rate of heterogeneous NO reduction is enhanced, particularly up to a CO/NO ratio of about 3; if the ratio is higher, a poor stabilisation occurs.

Goel et al. (1994) extended the model presented by de Soete (1990). They proposed a mechanism of N2O formation through a reaction with an active centre [NCO], and N2O reduction in a heterogeneous reduction reaction with elemental carbon C. They described NO decomposition through heterogeneous reduction over char and a homogenous reaction of NO with CO. Moreover, the authors proposed a mechanism where oxygen breaks a

problem seems less difficult, but the basic questions remain. It should be noted that experimental results of NO emissions during char combustion may be modelled in various ways, which means a substantial uncertainty of understanding of NO formation

Visona and Stanmore (1996, 1999) showed that while modelling NO emissions, comparable

nitrogen released as HCN diffuses to a particle surface and is converted into NO in the

The researchers found that both pathways of fuel-nitrogen release are possible at about 1750 K. Molina et al. (2000) compared models of coal particle combustion and pathways of fuelnitrogen conversion into nitrogen oxides presented by Wendt and Schulze (1976), Shimizu

The models proposed by Visona and Stanmore (1996) as well as by de Soete et al. (1999) are described above. A feature of the model presented by Wendt and Schulze (1976) is carbon oxidation to CO and char-nitrogen oxidation to NO in pores of heterogeneous particles. During diffusion towards the outer surface of the particle, CO undergoes homogenous oxidation to CO2, and NO homogenous reduction occurs which was assumed according to the Zeldowicz opposing reaction (1946). The mechanism is based on a reburning phenomenon that occurs in particle pores. An important factor for the mechanism is kinetics of CO oxidation adopted from Howard et al. (1973). The diameterlength ratio of the pores was demonstrated to determine a degree of char-N conversion into NO at about 1000 K. At temperatures far above 1000 K, oxygen is basically consumed at pore entrances so diffusion into pores becomes less significant and nitrogen conversion depends on phenomena occurring in the pore boundary layer. In this case, it was demonstrated that CO concentration in the pore neighbourhood is a key factor in the nitrogen conversion into NO. The authors did not perform any experimental verification of the developed mathematical model. Within the investigated combustion range, char strongly depends on initial coal; thus, in a mathematical model aimed at predicting NO concentrations, such factors as individual char characteristics and its changes during

Shimizu et al. (1992) presented the simplest model of nitric oxide emissions comprising one combustion reaction and two NO formation and reduction reactions. In coal particles, oxygen directly reacts with fuel nitrogen to produce NO; then it is reduced to N2 in a homogenous reaction with CO. Chan et al. (1983) and Goel et al. (1994) are the only researchers to include also the reaction of NO with CO catalysed by char. Studies by Chan et al. (1983) revealed that in the presence of CO, the rate of heterogeneous NO reduction is enhanced, particularly up to

Goel et al. (1994) extended the model presented by de Soete (1990). They proposed a mechanism of N2O formation through a reaction with an active centre [NCO], and N2O reduction in a heterogeneous reduction reaction with elemental carbon C. They described NO decomposition through heterogeneous reduction over char and a homogenous reaction of NO with CO. Moreover, the authors proposed a mechanism where oxygen breaks a

a CO/NO ratio of about 3; if the ratio is higher, a poor stabilisation occurs.

results may be obtained through various pathways of fuel-nitrogen conversion:

et al. (1992), Goel et al. (1994), Visona and Stanmore (1996) and Soete et al. (1999).

in reactions with O2 in particle pores, fuel-nitrogen is converted to NO

from char-nitrogen.

particle boundary layer.

particle burning should be considedred.

boundary active centre [CN] to form [N] which subsequently may react with O2 to produce NO or with NO to produce N2O, which was also suggested by Krammer and Sarofim (1994). A disadvantage of this model is that the experiments with one coal type were subsequently used for determination of kinetic constants.

In available literature, there are also some other models of nitrogen oxide formation and reduction described that were not compared in the paper by Molin et al. (2000).

Aria et al. (1986, 1986) presented a model where fuel-nitrogen was assumed to release as NH and NO in a heterogeneous reaction with O2, and NO reduction was assumed to occur in a homogenous reaction in a boundary layer of a particle. Although the combustion process was assumed to occur on the external layer of a particle and in its pores, porosity changes were not made dependant on a particle burn-out level, which is suggested by Visona and Stanmore (1996, 1999). Also, the preparation method of char which was subsequently chilled arouses some doubts.

Tullin et al. (1993) assumed that NO is produced as a result of a heterogeneous reaction of oxygen with char-nitrogen, while N2O is formed in a heterogeneous reaction of NO with fuel-nitrogen, and both compounds are reduced during a heterogeneous reaction with elemental carbon C. The authors found that N2O decreased with the increase in the particle burn-out level, which is contradictory to the observations made by de Soete (1990) and Croiset et al. (1998) who showed that the increase in N2O emissions was proportional to the coal burn-out level. Tullin et al. (1993) as well as Goel et al. (1994) verified the proposed model for one coal type.

Sarofim et al. (1995) described the porosity effects on NO and N2O release from fuelnitrogen. They observed a dependency on the surface area of reaction and diffusion in pores during conversion of fuel-nitrogen into NO and N2O. Moreover, they showed that the conversion degree of fuel-nitrogen is proportional to the particle burn-out. In their investigations, the authors ignored a large fraction of mesopores involved in the reaction and they entirely neglected macropores whose role in diffusion of gases is significant.

Despite considerable progress, understanding of the mechanisms of nitrogen compound conversions during coal combustion is still unsatisfactory (Benson, 1968; Anthony el al., 1976; Pottigisser, 1980; Bliek el al., 1985). There are no reliable data on the kinetics of nitrogen conversion at increased pressures. In view of the development of pressurised combustion techniques, there is an urgent need of detailed studies on the issue. In the published studies on the mechanism of fuel-nitrogen release and its conversion in a particle, measuring points (mostly recorded under isothermal conditions when the final heating temperature was reached) were applied. Experiments where samples are analysed during heating allow for more precise determination of kinetic constants because for individual reactions, it is possible to separate temperature and time dependencies. A majority of researches agree that formation of nitrogen oxides should be described on the basis of internal diffusion and a chemical reaction in pores. In case of pressurised combustion, there are no comprehensive analyses in the form of kinetic equations (Mallet, 1995) so the issue requires systematic studies on nitrogen conversion in pressurised combustion processes. A literature review shows that while determining kinetic constants of fuel-nitrogen conversion into nitrogen oxides, individual characteristics of investigated char porous structures should be considered.

Fuel-N Conversion to NO, N2O and N2 During Coal Combustion 45

According to the above processes, the amounts of forming oxides depend on combustion conditions. High oxygen concentrations during the initial phase of coal combustion promote

Historically, the first homogeneous mechanism of NO formation is a thermal mechanism. The thermal mechanism of nitrogen oxide formation from molecular nitrogen is a sequence

During his studies on gaseous flames at high temperatures, Zeldowicz (1946) observed that formation of thermal NOs cannot result from direct collisions of N2 and O2 molecules

N2 + O2 NO + NO (1)

N2 + O NO + N (2)

 O2 + M O + O + M (4) As the mechanism led to lower calculated results than the experimental ones, it was

The rate of NO formation resulting from the Zeldowicz mechanism (1946) mainly depends on the reaction kinetics (5) as well as concentrations of oxygen atoms and nitrogen molecules in the particle neighbourhood. Kinetic constants of the above reactions are presented in scientific papers (Arai et al., 1986; Miller & Bowman, 1989). A detailed review of the kinetic constants of thermal NOs formation are presented in the Ph.D. dissertation by Gradoń (2003), while the issues of the thermal mechanism are included in the papers by, among others: Fenimore, 1971; Johnsson et al., 1992; Bowman, 1973; Sarofim & Pohl, 1973; de Soete, 1975; Malte & Pratt, 1977; Fenimore & Fraenkel, 1981; Glarborg et al., 1986; Miller & Bowman, 1989; Muzio et al., 1990; Glarborg et al., 1992;

After a series of experiments conducted in a flow reactor at 1653 K to 1798 K, Tomeczek and Gradoń (2003) proposed a concept of extended thermal mechanism. Measured NO concentrations were far higher than those calculated according to the Zeldowicz mechanism (1946). The authors proposed the extended thermal mechanism based on five reactions: two reactions [(2) and (3)] were adopted from the Zeldowicz mechanism (1946)

N + O2 NO + O (3)

N + OH NO + H (5)

formation of fuel-nitrogen oxides (Attar and Hendrickson, 1982).

which is triggered by a reaction of molecular oxygen dissociation:

supplemented by a reaction (Fenimore & Jones, 1957; Fenimore, 1971):

of chemical reactions that occur independently of the combustion process.

**4.1 Homogeneous mechanism**

**Zeldowicz mechanism** 

according to a global reaction:

He proposed a double-reaction mechanism:

Williams et al., 1992; Tomeczek & Gradoń, 1997.

**Extended thermal mechanism** 

From the conducted analysis, it may be concluded that a significant part of experimentally confirmed differences in kinetic constants of char-nitrogen conversion into nitrogen oxides result from omission of individual characteristics of investigated char porous structures that strongly affect the mechanism of nitrogen conversion.
