N2O Using Ab Initio

CrN2O reaction, so far no one has analyzed this reaction from a theoretical point of view. Here, by using an ab initio MO theory, we attempted to anal...
6 downloads 0 Views 121KB Size
Download PDF
3434

J. Phys. Chem. B 1999, 103, 3434-3441

Analysis of the Reaction of Carbon with NO/N2O Using Ab Initio Molecular Orbital Theory Takashi Kyotani* and Akira Tomita Institute for Chemical Reaction Science, Tohoku UniVersity, 2-1-1 Katahira, Sendai 980-8577 Japan ReceiVed: December 2, 1998; In Final Form: February 26, 1999

An attempt was made to analyze the reaction of carbon with NO or N2O by using an ab initio molecular orbital theory. This method allows the simulation of the chemisorption process of these gas molecules on zigzag and armchair edge sites of carbon. It was found that NO adsorption with the N atom down is more thermally favorable than the adsorption with the O atom down, while the O down mode is more favorable than the N down one for N2O adsorption because the former process releases a stable N2 molecule to form a surface oxygen complex. The chemisorption of NO or N2O with its bond axis parallel to the edge line gave the most stable chemisorbed species. The presence of surface oxygen complexes (quinone-type carbonyl group) on the edge decreases the strength of some bonds in NO- and N2O-chemisorbed species and consequently lowers their thermal stability. Furthermore, the N2 formation process in the C-NO/N2O reaction was analyzed and possible N2 formation routes were proposed. The routes predicted by the molecular orbital theory were compared with the previous experimental results.

Introduction The reduction of nitrogen oxide by coal char or pure carbon has been widely investigated by many researchers in relation to the NOx and/or N2O emission during coal combustion.1,2 We have also carried out a series of studies3-9 on this issue using a polymer resin char and attempted to clarify the reaction mechanism with several experimental techniques such as a temperature-programmed reaction and step response experiments using isotopically labeled reactants. It was found that a significant amount of nitrogen is trapped on carbon during the C-NO reaction.4 The formation of N2 from this reaction can be mainly ascribed to the reaction between surface nitrogen species, C(N), and gaseous NO.6,7 The reaction scheme can be expressed by the following equations:

2C() + NO f C(N) + C(O) or CO

(1)

C(N) + NO f N2 + C(O) or CO

(2)

where C() and C(O) are the carbon free site and surface oxygen complex, respectively. Whether the O atom from NO molecule remains as C(O) on the carbon surface or desorbs as CO depends on the severity of reaction conditions. In the case of the C-N2O reaction, the following reaction has been generally accepted as the origin of N2 formation:10-12

C() + N2O f N2 + C(O)

(3)

Very recently, we have experimentally confirmed that N2 is formed without breakage of the N-N bond in the N2O molecule during the C-N2O reaction,8 which is consistent with reaction 3. In addition to the experimental approach, the molecular orbital (MO) theory is very helpful for the understanding of the reaction mechanism. The usefulness of MO theory in the analysis of catalyzed and uncatalyzed carbon gasification has been demonstrated by many researchers.13-20 The following two works were dedicated to the analysis of the C-NO system using MO theory. Iaconis et al. analyzed the adsorption of the NO molecule

on the carbon basal plane by using MO theory at both semiempirical and ab initio levels.21 They simulated physical adsorption of NO and predicted this adsorption to be an endothermic process. Their purpose was not the theoretical analysis of the chemical reaction of NO with the carbon edge site, but the analysis of the physical adsorption of NO on the carbon basal plane. Ninomiya et al. applied a semiempirical MO theory to understand the mechanism of N2O formation from NO and C(N).22 It was found that the NO molecule chemisorbs preferentially on the N atom in the pyridine molecule (as a model C(N) compound) to form NNO species. Thus, they regarded N2O desorption from such NNO species as a possible origin of N2O formation during coal combustion. However, no theoretical analysis on the C-NO reaction and N2 formation from this reaction was done in their report. In the case of the C-N2O reaction, so far no one has analyzed this reaction from a theoretical point of view. Here, by using an ab initio MO theory, we attempted to analyze and simulate the reaction processes such as the chemisorption of NO and N2O molecules on the carbon surface and the subsequent desorption of N2 molecules from the chemisorbed species. In addition, we compared these results with the previous experimental results. The objects of this study are to explain the experimental results of C-NO and C-N2O reactions from a theoretical point of view and to give a concrete picture of how the NO or N2O molecule reacts with carbon to form N2. Calculation Method The Gaussian 94 package23 was used to calculate molecular orbitals of model compounds for the carbon-NO, N2O system. For simplicity, several single layers of polyaromatic compounds with different sizes were employed as models for carbon. Figure 1 illustrates the structure of two types of carbon models; one type (models A, B, and C) has zigzag and the other one (models D and E) has armchair edge site on the upper side of each model. Some of carbon atoms are numbered for convenience. The edge atoms on the upper side are unsaturated and the rest of them

10.1021/jp9845928 CCC: $18.00 © 1999 American Chemical Society Published on Web 04/08/1999

Reaction of Carbon with NO/N2O

J. Phys. Chem. B, Vol. 103, No. 17, 1999 3435 TABLE 1: Average C-C Bond Length and Bond Angle for Carbon Models

Figure 1. Selected models for the carbon edge plane. Large and small circles stand for carbon and hydrogen atoms, respectively.

model

av C-C bond length (pm)

av bond angle (deg)

A B C D F

141.3 141.4 141.6 140.1 140.6

120.30 120.31 120.31 120.10 120.11

(d) for single point energy calculation (the approach using these two methods sequentially is generally symbolized by B3LYP/ 6-31G(d)//UHF/3-21G(d)). The applicability of this approach to a polyaromatic molecule was rationalized by the pioneer work of Chen et al.24 The spin multiplicity of each compound including the chemisorbed species was determined in the following way: we always made B3LYP/6-31G(d)//UHF/3-21G(d) calculations for a given structure under different values of spin multiplicity, and we chose the multiplicity which led to a chemically reasonable molecular structure and spin densities of edge carbon atoms. Furthermore, we confirmed that the selected multiplicity gives the smallest difference between the expectation value of the total spin operator 〈S2〉 and the expected value S(S + 1)/2 (where S stands for the total spin). When these three conditions are satisfied among different values of multiplicity, we select the one that gave the most energetically stable state. The heat of adsorption, ∆H, was determined as a difference between the total energy of the optimized system and the sum of the energies of the corresponding carbon model and gas molecule. We evaluated the thermal stability of the resultant chemisorbed species from the calculated heat of adsorption. Furthermore, we discussed a possible path of N2 formation from the nitrogen chemisorbed species, C(N), on the carbon edge site. Results and Discussion

Figure 2. N-down, O-down, and side-on modes of NO approach toward zigzag and armchair sites.

are terminated with hydrogen atoms. A single NO or N2O molecule was put at a distance of 0.13-0.15 nm from the unsaturated edge (zigzag or armchair site) in three different ways, as shown in Figure 2. For N-down and O-down modes, the NO bond axis was perpendicular to the edge line with the N and O atoms down, respectively. In the case of the side-on mode, the bond axis is parallel with the edge line. The three ways of approaches (N- and O-down and side-on) were adopted also for the chemisorption of the N2O molecule. In the present study, we only consider the approaches in the plane of the carbon model layer. The whole system including the model carbon and the gas molecule was subjected to an ab initio MO calculation with all the geometrical parameters optimized. We employed the UHF (unrestricted Hartree-Fock) method with the basis set of 3-21G(d) for geometric optimization, followed by density functional method using the B3LYP functional with the basis set of 6-31G-

Ab Initio MO Calculation of Carbon Models. Before the simulation of the reaction of carbon with NO or N2O, we made the ab initio MO calculations of the carbon models selected in the present study. The results are summarized in Table 1, where average values of all the C-C bond lengths and bond angles in each model are given. It was found that the zigzag type models (A-C) have a longer average C-C length and a larger angle than the armchair models (D and F). For both models, the average bond length is increased with increasing model size. Chen et al. have already determined the optimized structures of models A and B by using the B3LYP/6-31G(d)//UHF/3-21G(d) method,24 and the results of our calculation are almost the same as theirs. The values of length and atomic bond population for the bonds between the numbered carbon atoms are tabulated in Table 2. The latter parameter can be used as a measure of bond strength and it was determined by the Mulliken population analysis, which is available in the Gaussian 94 package. In the following sections, we will discuss how these parameters are changed upon NO or N2O chemisorption on the edge site. Chemisorption of NO on the Carbon Edge Site. Figure 3 show the structure of the surface nitrogen complexes created on zigzag site, and Table 3 indicates the bond length and the atomic bond population for each chemisorbed species. Despite the full optimization, all the resultant structures were found to be in one plane; i.e., all dihedral angles are either 0 or 180° in each model (this was the case also for all other chemisorbed species in the present study). When NO was in the N-down mode above the edge carbon atom (C(2)) of model A, a linear C(NO) species was formed (structure a). The N-down approach to C(2) of models B and C gave a similar linear species

3436 J. Phys. Chem. B, Vol. 103, No. 17, 1999

Kyotani and Tomita

TABLE 2: Bond Length and Atomic Bond Population for Some Bonds in Carbon Models model

bond

length (pm)

atomic bond population

model

bond

length (pm)

atomic bond population

A

C(1)-C(5) C(2)-C(5) C(5)-C(8) C(1)-C(4) C(1)-C(5) C(2)-C(5) C(2)-C(6) C(5)-C(8) C(1)-C(4) C(1)-C(5) C(2)-C(5) C(2)-C(6) C(3)-C(6) C(3)-C(7) C(5)-C(8) C(6)-C(9)

141 139 145 141 140 140 141 145 140 141 139 139 141 140 145 145

0.76 0.66 0.86 0.63 0.64 0.64 0.63 0.84 0.65 0.65 0.66 0.66 0.65 0.65 0.81 0.81

D

C(1)-C(2) C(1)-C(3) C(2)-C(4) C(1)-C(2) C(2)-C(5) C(3)-C(4) C(3)-C(6) C(5)-C(6)

134 144 144 141 136 141 136 147

0.84 0.72 0.72 0.74 1.00 0.74 1.00 0.36

B

C

Figure 3. Surface nitrogen complexes formed by NO adsorption on the zigzag site in N- and O-down and side-on modes.

(structures b and c, respectively). It can be seen from Table 3 that structures a and c have almost the same geometry and distribution of atomic bond population. In the case of the O-down mode, a linear C(ON) species was formed (structure d). The atomic bond population of N-O bond is much smaller than that in a, b, and c, suggesting weak bond strength of the N-O in d. For all the cases, the production of the linear species reduced the bond strength of the most neighboring C-C bonds (C(2)-C(5) and C(2)-C(6)). The side-on approach to C(1) and C(2) of models B and C gave species e and f, respectively, where both N and O atoms are chemically bound to the edge carbon atoms to form a five-membered ring. As a result of such chemisorption of NO, the framework of the substrate (models B and C) was distorted to a greater extent; the bond length of C(5)-C(8) decreased from 145 pm to 138 or 139 pm. Such forcible bond length reduction results in the decrease of bond strength (from 0.84 or 0.86 to 0.13 in atomic bond population). The topology and bond strength of species e are almost the same as those of f. The observed geometrical similarities between a and c and between e and f imply that a large size carbon structure like model C is not always necessary for the simulation of NO chemisorption on the carbon edge site.

E

The calculated value of ∆H for each species is tabulated in Table 4. It was found that all the six chemisorption processes toward the zigzag site are exothermic, i.e., a negative value of ∆H. Especially, structures e and f are the most stable species. The values of ∆H for the N-down approaches (a, b, and c) are in the range of -150 to -210 kJ/mol, but the O-down approach gave a small absolute value of ∆H (-10 kJ/mol). The N-down approach is, therefore, a more thermally favorable process than the O-down one. A similar MO calculation was done toward the armchair site using models D and E. The resultant NO chemisorbed structures are illustrated in Figure 4, and the values of bond length and atomic bond population are indicated in Table 5. The N-down approach to the C(1)-C(2) bond of model D gave structure g with a ∆H of -100 kJ/mol (Table 4), which is somewhat higher than the ∆H’s for NO chemisorption on the zigzag site in the N-down mode. The O-down mode produced a C(ON) species (structure h) with a N-O bond length of 187 pm and atomic bond population of 0.08. This very small value suggests that the N atom has only a small chemical interaction with the O atom, i.e., the release of a single N atom from NO during its O-down approach. As can be expected, the process that produces an unstable N atom has a large positive value of ∆H (520 kJ/ mol). In other words, the O-down approach to the armchair site is not energetically favorable. On the other hand, the side-on approach to model E produced very stable species that consists of a six-membered ring (structure i), whose ∆H is close to that of the stable species formed on the zigzag site by the side-on approach (structures e and f). The calculated values of ∆H for the chemisorbed species are not far from the experimental value reported by Xia et al.25 They pretreated activated carbon under N2 flow at 1173 K, and then without air exposure, this sample was subjected to NO adsorption at 300 K. The measured ∆H at very low coverage was in the range of 300-500 kJ/mol. Since the surface of such a heattreated sample must be rich in the unsaturated site, the observed values of ∆H were attributed to NO chemisorption on the unsaturated sites, which is what we simulated in the present study. Taking it into consideration that we employed only a single layer of simple polyaromatic compound as a model of carbon structure, the rough agreement between the calculated and experimental ∆H implies such simplification to be tolerable for the simulation of the NO chemisorption phenomenon by the ab initio MO calculation. Effect of C(O) on NO Chemisorption. As coal combustion occurs in the presence of oxygen, it would be worth elucidating the effect of oxygen on the C-NO reaction.9 It is generally accepted that surface oxygen complexes, C(O), function as an

Reaction of Carbon with NO/N2O

J. Phys. Chem. B, Vol. 103, No. 17, 1999 3437

TABLE 3: Bond Length and Atomic Bond Population of NO-Chemisorbed Species Formed on the Zigzag Site structure

bond

length (pm)

atomic bond population

structure

bond

length (pm)

atomic bond population

a

N-O C(2)-N C(1)-C(5) C(2)-C(5) C(5)-C(8) N-O C(2)-N C(1)-C(5) C(2)-C(5) C(2)-C(6) C(5)-C(8) N-O C(2)-N C(1)-C(5) C(2)-C(5) C(5)-C(8) N-O C(2)-O C(1)-C(5) C(2)-C(5) C(5)-C(8)

124 131 139 144 143 131 136 144 142 142 144 125 130 139 145 143 135 135 140 141 143

0.57 0.68 0.89 0.30 0.75 0.52 0.67 0.79 0.28 0.35 0.66 0.56 0.67 0.78 0.33 0.67 0.25 0.54 0.80 0.38 0.79

e

N-O C(1)-O C(2)-N C(1)-C(4) C(1)-C(5) C(2)-C(5) C(2)-C(6) C(5)-C(8) N-O C(1)-O C(2)-N C(1)-C(4) C(1)-C(5) C(2)-C(5) C(2)-C(6) C(5)-C(8)

147 139 136 142 140 143 144 139 147 139 136 141 137 140 144 138

0.27 0.41 0.72 0.35 0.60 0.47 0.26 0.13 0.27 0.40 0.72 0.36 0.61 0.48 0.26 0.13

b

c

d

f

TABLE 5: Bond Length and Atomic Bond Population of NO-Chemisorbed Species Formed on the Armchair Site structure

bond

length (pm)

atomic bond population

g

N-O C(1)-N C(2)-N C(1)-C(2) C(1)-C(3) N-O C(1)-O C(2)-O C(1)-C(2) C(1)-C(3) N-O C(3)-N C(2)-O C(1)-C(2) C(2)-C(5) C(3)-C(4) C(3)-C(6) C(5)-C(6)

129 146 146 131 141 187 145 145 137 135 144 136 139 138 140 141 143 141

0.50 0.22 0.22 0.32 0.55 0.08 0.22 0.22 0.39 0.53 0.11 0.56 0.27 0.84 0.83 0.75 0.75 0.37

h

i

Figure 4. Surface nitrogen complexes formed by NO adsorption on the armchair site.

TABLE 4: Heats of Adsorption of NO-Chemisorbed Species structure

∆H (kJ/mol)

structure

∆H (kJ/mol)

a b c d e

-210 -150 -200 -10 -570

f g h i

-570 -100 520 -530

intermediate during carbon gasification with O2. Here, we have attempted to investigate how the presence of C(O) influences the chemisorption of NO on the carbon edge site. As a basic model of the edge site, we employed model C and bound a single O atom to the C(3) atom as a quinone-type carbonyl group. After the optimization calculation of this oxygencontaining model, we put a single NO molecule in either the N-down or side-on mode on the edge sites. The final optimized structures of two chemisorbed species are shown in Figure 5, together with the values of atomic bond population for some of the bonds. The N-down approach gave structure j, where the

C(NO) species is not linear, as observed in structure c, but the N-O bond is tilted toward the other O atom bound to the C(3) atom. There is, however, no chemical interaction between these two O atoms (its atomic bond population is a negative value). It was found that the atomic bond population of the N-O bond (0.30) is smaller than that of the N-O bond (0.56) in structure c. In the case of the side-on mode, structure k was formed. The geometry of this five-membered ring species was almost the same as that of structure f, but the values of atomic bond population for the N-O and C(2)-N bonds are somewhat lower than the corresponding bonds in structure f, respectively. For both cases, the resultant absolute value of ∆H (-160 and -490 kJ/mol for j and k, respectively) was found to be 10-20% less than that of the model without C(O) (c and f). These findings indicate that the presence of C(O) decreases the thermal stability of the C(NO) species. N2 Formation Mechanism in C-NO Reaction. To elucidate the N2 formation route from surface nitrogen complexes, we tried to put one more molecule of NO on the N atom of a chemisorbed species similar to structure f (the left side of Figure 6). The ab initio optimized calculation for such a system resulted in the formation of C(NNO) surface species (structure l in Figure 6), where the value of the atomic bond population for each bond is indicated. It was found that the bond population of N-O is

3438 J. Phys. Chem. B, Vol. 103, No. 17, 1999

Kyotani and Tomita

Figure 5. Surface nitrogen complexes formed by NO adsorption on edge sites having a surface oxygen complex. The value of the atomic bond population for each bond is indicated. Unless stated otherwise, numerical values in the following figures correspond to the atomic bond population.

Figure 6. N2 formation process in the reaction of the surface nitrogen complex with the NO molecule.

Figure 7. N2 formation process from the C(NN) species.

much lower than those of the other bonds. The breakage of the N-O bond would be, therefore, easy. To estimate the bond strength of the C(NN) species after such bond breakage, we made the optimized calculation for a linear C(NN) species between two C(O)’s using model A as a basic carbon structure (a self-consistent field could not be obtained for a similar calculation using model C). The resultant optimized structure and the values of atomic bond population are shown in Figure 7 (structure m). The linearity of C(NN) was kept even after the optimization. It should be noted that the N-N bond has a large bond population (1.39), but the value for the C-N bond is very small (0.15). Such a small atomic bond population would lead to easy desorption of N2 from structure m due to the bond breakage of C-N. Thus, we can expect N2 formation from the C(NNO) species once the species has a structure such as m due to the bond breakage of N-O. From our step response experiments using isotopes,6,7 we concluded that N2 is mainly formed by the reaction between surface nitrogen complex and NO gas, as expressed in eq 2. The mechanism predicted above is consistent with the mechanism determined by the experimental approach. As a chemical form of N in carbon, many researchers have pointed out so far the presence of pyridinic, pyrrolic, and quaternary nitrogen from the analysis with X-ray photoelectron spectroscopy.4,5,26,27 Here, we evaluate the possibility of N2

Figure 8. Structures of the carbon model with pyridinic nitrogen and its NO-chemisorbed species.

formation from the reaction of pyridinic nitrogen with NO by using a model illustrated in the upper side of Figure 8 (structure n). The values of the atomic bond population for some of the bonds are given in Figure 8. The distribution of the atomic bond population for structure n is not so different from that of model B (Table 2), although the values of the bond length of C(5)-N and C(6)-N (136 pm) are much shorter than those of the corresponding bonds in model B. NO chemisorption in the sideon mode weakened the strength of the two C-N bonds, as shown in the right structure in Figure 8 (structure o). These values are, however, still larger than the bond population of N-N, suggesting difficulty of N2 desorption from structure o. Furthermore, the value of ∆H for such NO chemisorption was -190 kJ/mol, which is much higher than that of the side-on

Reaction of Carbon with NO/N2O

J. Phys. Chem. B, Vol. 103, No. 17, 1999 3439

Figure 10. Formation of the N-N bond from two surface nitrogen complexes.

C(N) + C(N) f 2C() + N2

Figure 9. Structures of the carbon model with the pyridone type nitrogen and its NO-chemisorbed species.

chemisorption onto model B (-570 kJ/mol). It means that the side-on approach toward pyridinic nitrogen is less favorable process than the approach toward the nitrogen-free carbon edge site. As another type of surface nitrogen complex, Zhu et al. have experimentally verified the formation of a pyridone functional group when carbons with pyridinic and pyrrolic groups were partially combusted.28 Hence, it is worth investigating the possibility of N2 formation from the pyridone functional group. We have employed structure p (the upper side of Figure 9) as a pyridone type nitrogen complex. The structure was already optimized, and the resultant values of the atomic bond population were indicated in this figure. Zhu et al. have suggested that the decomposition of the pyridone structure by cleavage of the C(4)-N bond leads to the production of C(O) and C(N) species with the latter being subsequently subjected to another reaction such as oxidation. Such selective bond breakage is not very probable from the theoretical point of view, because the difference in bond population between the C(4)-N bond and the other bonds is rather small, even though the bond population for the C(4)-N is lower than the others. For the analysis of N2 formation from structure p, we put a NO molecule toward the nitrogen atom of the pyridone structure in the side-on mode. The optimized structure is illustrated in the right side of Figure 9 as structure q together with the values of the atomic bond population. The NO approach results in the formation of a sixmembered ring with two nitrogen and one oxygen atoms. This adsorption process was an exothermic reaction with ∆H ) -250 kJ/mol and it weakened the strength of C(4)-N and C(6)-N bonds very much. Furthermore, the bond population for the N-O bond was reduced to only 0.10, indicating easy breakage of this bond. Since the bond population for the N-N bond is still smaller than those for C(4)-N and C(6)-N bonds, the explicit prediction whether N2 desorbs or not is not possible. However, we can conclude at least that N2 formation from structure q is more probable than that from structure o in Figure 8. As another mechanism of N2 formation in the C-NO reaction, the following one has also been proposed by many researchers:1,11,12

(4)

In the previous study, we have experimentally confirmed the presence of this reaction, even though it is not a major route for N2 formation, compared with eq 2.7 Here, we theoretically evaluate the possibility of reaction 4 by using the structure of model B, whose C(1) and C(2) atoms are bound to single N atoms, respectively. The resultant optimized structure is shown in Figure 10 together with the values of atomic bond population. A five-membered ring including two nitrogen atoms was formed on the edge plane, and the two nitrogen atoms are chemically bound each other. The population of this N-N bond is, however, lower than those of two C-N bonds. The breakage of the C-N bond would be, therefore, less plausible than that of the N-N bond. Thus, we can conclude that reaction 4 is very difficult to take place on the zigzag edge, as in Figure 10. Reaction of N2O with Carbon. The MO calculation of the C-N2O system was carried out in a manner similar to the case of the C-NO system; N-down, O-down, and side-on approaches on both zigzag and armchair edges using models A, B, D, and E. The N-down mode produced the C(NNO) surface species, while the O-down one led to the release of the N2 molecule to leave the O atom bound to the edge carbon atom. In the case of the side-on mode, six- and seven-membered rings are formed with ∆H of around -400 kJ/mol on zigzag and armchair sites, respectively. These final optimized structures on the carbon surface are illustrated in Figure 11, and the values of the bond length and atomic bond population are given in Table 6. The value of ∆H for each chemisorption process is indicated in Table 7. The very low values of the C-N bond population in structures s and v implies easy desorption of N2O from the edge site. The instability of the C(NNO) species is confirmed by the values of ∆H (-20 and -40 kJ/mol for s and v, respectively). The large negative values of ∆H for structures t and w can be ascribed to the production of a stable N2 molecule upon the chemisorption process. Since the C(NNO) species in structure u has a geometry and bond population similar to those of the C(NNO) species in structure l (Figure 6), N2 release from u would be possible by the same token as the case of l. For the structure x, the N-N bond is stronger than C-N and N-O bonds. Thus, N2 desorption from x is also possible. These results suggest that except for the N-down mode, the reaction of C with N2O can produce N2 and the surface oxygen complex, C(O). The reaction scheme is exactly the same as eq 3. Conclusions The chemisorption process of NO and N2O molecules on carbon edge sites and the subsequent reduction of these gases to N2 were simulated by using an ab initio MO method. The

3440 J. Phys. Chem. B, Vol. 103, No. 17, 1999

Kyotani and Tomita

Figure 11. Surface nitrogen complexes formed by N2O adsorption.

TABLE 6: Bond Length and Atomic Bond Population of N2O-Chemisorbed Species structure

bond

length (pm)

atomic bond population

structure

bond

length (pm)

atomic bond population

s

N-O N-N C(2)-N C(1)-C(5) C(2)-C(5) C(5)-C(8) C(2)-O C(1)-C(5) C(2)-C(5) C(5)-C(8) N-O N-N C(1)-O C(2)-N C(1)-C(4) C(1)-C(5) C(2)-C(5) C(2)-C(6) C(5)-C(8)

131 114 132 139 143 143 128 140 144 143 147 132 139 136 141 139 142 143 142

0.19 1.02 0.10 0.84 0.61 0.76 1.23 0.84 0.26 0.76 0.06 0.50 0.47 0.64 0.49 0.45 0.24 0.42 0.67

v

N-O N-N C(1)-N C(2)-N C(1)-C(2) C(1)-C(3) C(2)-C(4) C(1)-O C(1)-C(2) C(1)-C(3) N-O N-N C(2)-O C(3)-N C(1)-C(2) C(2)-C(5) C(3)-C(4) C(3)-C(6) C(5)-C(6)

126 130 145 145 130 140 140 154 126 141 142 128 138 137 138 142 140 145 145

0.48 0.34 0.03 0.03 0.66 0.51 0.52 0.21 0.81 0.62 0.01 0.51 0.27 0.38 0.90 0.87 0.38 0.82 0.43

t

u

TABLE 7: Heats of Adsorption of N2O-Chemisorbed Species structure

∆H (kJ/mol)

s t u v w x

-20 -500 -440 -40 -270 -390

adsorption of NO and N2O on carbon edge sites resulted in the formation of several types of nitrogen-containing complexes, C(N). From the thermal stability of these complexes, it was concluded that for NO adsorption the N-down approach toward the edge site was more thermally favorable than the O-down one and the chemisorption of the NO or N2O molecule with its bond axis parallel to the edge plane gave the most stable chemisorbed species. The presence of surface oxygen complexes, C(O), on edge sites reduced the heat of NO chemisorption to some extent. The MO calculation of the system including both NO and a nitrogen chemisorbed species led to the formation of a sixmembered ring complex including NNO bonding. The bond

w x

population analysis predicted that N2 desorption from such a complex is very probable. This reaction scheme is exactly the same as eq 2, which was experimentally confirmed. A similar analysis was done about N2 formation either from the reaction of pyridinic or pyridone type nitrogen with NO or from two neighboring C(N) species (eq 4) on a zigzag site. In the case of the C-N2O reaction, the MO calculation predicted the release of N2 and the formation of C(O) (eq 3) when N2O was put on the edge site in both the O-down and side-on modes. In conclusion, the present study demonstrated how an ab initio MO theory is useful for the understanding of the reaction mechanism even for a very complicated system such as C-NO and C-N2O reactions. Although we investigated the effect of surface oxygen complexes on the C-NO reaction, the situation simulated here is still far from the real one in combustion, because it involves the presence of low levels of NO/N2O and a high concentration of O2. The presence of O2 should be incorporated in the next step of the simulation. Acknowledgment. The authors thank Dr. H. Gotoh of Toyohashi University of Technology, Japan, for his helpful discussion.

Reaction of Carbon with NO/N2O References and Notes (1) Thomas, K. M. Fuel 1997, 76, 457-473. (2) Aarna, I.; Suuberg, E. M. Fuel 1997, 76, 475-491. (3) Yamashita, H.; Tomita, A.; Yamada, H.; Kyotani, T.; Radovic, L. R. Energy Fuels 1993, 7, 85-89. (4) Suzuki, T.; Kyotani, T.; Tomita, A. Ind. Eng. Chem. Res. 1994, 33, 2840-2845. (5) Chambrion, Ph.; Suzuki, T.; Zhang, Z.-G.; Kyotani, T.; Tomita, A. Energy Fuels 1997, 11, 681-685. (6) Chambrion, Ph.; Orikasa, H.; Kyotani, T.; Tomita, A. Fuel 1997, 76, 493-498. (7) Chambrion, Ph.; Kyotani, T.; Tomita, A. Energy Fuels 1998, 12, 416-421. (8) Noda, K.; Chambrion, Ph.; Kyotani, T.; Tomita, A. Eurocarbon ’98, Extend Abstracts and Programme - Vol. II; The French and German Carbon Groups: Strasbourg, 1998; pp 597-598. (9) Chambrion, Ph.; Kyotani, T.; Tomita, A. 27th International Symposium on Combustion; The Combustion Institute: Pittsburgh, PA, 1998; pp 3053-3063. (10) Smith, R. N.; Lesnini, D.; Mooi, J. J. Phys. Chem. 1957, 61, 8186. (11) De Soete, G. G, 23rd International Symposium on Combustion; The Combustion Institute: Pittsburgh, PA, 1990; pp 1257-1264. (12) Rodriguez-Mirasol, J.; Ooms, A. C.; Pels, J. R.; Kapteijn, F.; Moulijn, J. A. Combustion Flame 1994, 99, 499-507. (13) Pan, Z. J.; Yang, R. T. J. Catal. 1990, 123, 206-214. (14) Chen, S. G.; Yang, R. T. J. Catal. 1993, 141, 102-113. (15) Chen, S. G.; Yang, R. T.; Kapteijn, F.; Moulijn, J. A. Ind. Eng. Chem. Res. 1993, 32, 2835-2840. (16) Janiak, C.; Hoffmann, R.; Sjo¨vall, P.; Kasemo, B. Langmuir 1993, 9, 3427-3440.

J. Phys. Chem. B, Vol. 103, No. 17, 1999 3441 (17) Kyotani, T.; Ito, K.; Tomita, A.; Radovic, L. R. AIChE J. 1996, 42, 2303-2307. (18) Kyotani, T.; Ogasawara, S.; Radovic, L. R.; Tomita, A. Proceedings of Carbon ’96; The British Carbon Group: Newcastle upon Tyne, 1996; pp 276-277. (19) Skokova, K.; Radovic, L. R. Prepr.-Am. Chem. Soc. DiV. Fuel. Chem. 1996, 41, 143-147. (20) Chen, S. G.; Yang, R. T. Energy Fuels 1997, 11, 421-427. (21) Iaconis, E.; Illas, F.; Russo, N.; Toscano, M. Gazz. Chim, Ital. 1988, 118, 603-605. (22) Ninomiya, Y.; Kato, H.; Koketsu, J. In Coal Science Vol. I, Proceeding of the Eighth International Conference on Coal Science; Pajares, J. A., Tascon, J. M. D., Eds.; Elsevier: Amsterdam, 1995; pp 747-750. (23) Frisch M. J.; Trucks, G. W.; Schlegel, H. B.; W. Gill, P. M.; Johnson, B. G.; Robb, M. A.; Cheeseman, J. R.; Keith, T.; Petersson, G. A.; Montgomery, J. A.; Raghavachari, K.; Al-Laham, M. A.; Zakrzewski, V. G.; Ortiz, J. V.; Foresman, J. B.; Cioslowski, J.; Stefanov, B. B.; Nanayakkara, A.; Challacombe, M.; Peng, C. Y.; Ayala, P. Y.; Chen, W.; Wong, M. W.; Andres, J. L.; Replogle, E. S.; Gomperts, R.; Martin, R. L.; Fox, D. J.; Binkley, J. S.; Defrees, D. J.; Baker, J.; Stewart, J. P.; HeadGordon, M.; Gonzalez, C.; Pople, J. A. Gaussian 94, Revision E. 2; Gaussian, Inc.: Pittsburgh, PA, 1995. (24) Chen, N.; Yang, R. T. Carbon 1998, 36, 1061-1070. (25) Xia, B.; Phillips, J.; Menendez, J. A.; Radovic, L. R. Extended Abstracts of 23rd Biennial Conference on Carbon; The American Carbon Society: University Park, PA, 1997; pp 418-419. (26) Pels, J. R.; Kapteijn, F.; Moulijn, J. A.; Zhu, Q.; Thomas, K. M. Carbon 1995, 33, 1641-1653. (27) Kelemen, S. R.; Gorbaty, M. L., Kwiatek, P. J.; Fletcher, T. H.; Watt, M.; Solum, M. S.; Pugmire, R. J. Energy Fuels 1998, 12, 159-173. (28) Zhu, Q.; Money, S. L.; Russel, A. E.; Thomas, K. M. Langmuir 1997, 13, 2149-2157.