1 Introduction

CO2 is considered to be one of the most important anthropogenic gas emission species involved in global warming (Habib et al., 2011). Oxy-fuel combustion is one of the advanced technologies used to capture CO2 when burning fossil fuels (Buhre et al., 2005). In oxy-fuel combustion, oxygen or a mixture of oxygen and flue gas, is used as an oxidizer instead of air. As a result, the volume fraction of CO2 is much higher than that of typical air-fuel combustion conditions. With more than 95% concentrated in the product gas, through a series of relatively simple purification processes, CO2 can be compressed, transported, and stored (Buhre et al., 2005).

Compared with typical air-fuel systems, the combustion process (as well as pollutant formation) will be different in the oxy-fuel combustion process due to the high CO2 level (Chui et al., 2003; Wang et al., 2008). Undiluted, the temperature of oxy-fuel combustion will be much higher than that of air combustion because of the absence of nitrogen. To apply oxy-fuel combustion in ordinary boilers, flue gas needs to be recycled to maintain the temperature at an acceptable level (Chui et al., 2004). The flame temperature decreases when the amount of oxygen in the inlet is reduced and the amount of recycled flue gas is increased (Bhuiyan and Naser, 2015a). The temperature profile in the boiler may be changed due to the differences in heat capacity between CO2 and N2. The flame propagation speed also differs (Normann et al., 2009) due to differences in heat capacity and chemical properties between CO2 and N2. Significant variation has been found in the temperature distribution and CO2 concentration when comparing air-firing with oxy-firing under different recycling ratios in a computational model of co-firing of biomass with coal under oxy-fuel conditions (Bhuiyan and Naser, 2015b).

The reduction of NO x due to recycled flue gas is a known benefit of oxy-fuel combustion (Kiga et al., 1997; Chui et al., 2004; Buhre et al., 2005; Ahn et al., 2009). Part of the NO in the flue gas is reduced by precursors, such as HCN and NH3, and carbonaceous reducing radicals in fuel-rich regions (Buhre et al., 2005).

However, the formation of NO x during combustion will be indirectly affected by the occurrence of high concentrations of carbon dioxide. As is well known, the production of NO x can be dramatically reduced through staging combustion technology, even without flue gas recirculation (Boushaki et al., 2008; 2009). Much more NO x reduction can be achieved by staging compared with the air-fuel combustion process (Cao et al., 2010; Watanabe et al., 2011). Also, thermal NO x would be dramatically reduced because of the absence of nitrogen from the system (Normann et al., 2008).

CO2 is not 100% chemically inert (Hecht et al., 2011), although it is usually the final combustion product of elemental carbon. Reaction (1) illustrates how OH is formed from CO2 and H (Glarborg and Bentzen, 2008; Mendiara and Glarborg, 2009; Watanabe et al., 2011). A chain carrier of H of a branched chain reaction during combustion of fuels with H is involved in this reaction (Fig. 1). The pathways from nitrogen containing species to NO x are linked closely to chain carriers. The formation of NO x , fuel NO, is then affected indirectly by the chemical effect of CO2. The effect of oxy-fuel combustion on NO formation is also directly increased by high CO2 concentrations through Reaction (2), which is also believed to be the reason for the CO2 chemical effect on NO (Feng et al., 1998).

$${\rm{C}}{{\rm{O}}_2} + {\rm{H}} \rightarrow {\rm{CO}} + {\rm{OH}},$$
(1)
$${\rm{C}}{{\rm{O}}_2} + {\rm{N}} \rightarrow {\rm{NO}} + {\rm{CO}}.$$
(2)
Fig. 1
figure 1

Branched/Unbranched chain reactions in oxy-fuel combustion

Full size image

Mendiara and Glarborg (2009) studied the oxidization of methane and ammonia under highly diluted conditions in CO2 and N2 via experiment and numerical simulation. They found that, compared with nitrogen-diluted conditions, CO2 facilitates NO formation under fuel-rich conditions but inhibits it under fuel-lean or stoichiometric conditions. It is believed that high levels of CO2 facilitate the formation of NO by increasing the OH/H ratio. Furthermore, the formation of HNCO through NH2+CO is increased due to high CO levels. However, some reactions of N containing species with O or H are inhibited due to decreased concentrations of O and H, such as NH2+O to form HNO and NH2+H to form NH. These two reactions are important pathways for NH3 to form NO, especially NH2+O to form HNO.

In a study of staged oxy-fuel combustion, Watanabe and Ichiro et al. (2011) observed higher OH levels via planar laser induced florescence (PLIF) in a fuel/O2/CO2 flame than in a fuel/air flame at an O2/CH4 ratio of 0.7 (equivalence ratio of 1.42). They believed that Reaction (1) does not compete with the branched chain reaction for the H atom since the radical pool of chain carriers is large. The oxidation of species containing nitrogen increases due to the increased OH radicals, and NO x tends to be reduced in fuel-rich conditions. Because precursors of NO x are prevented from being oxidized by supplementary oxygen in the secondary stream, staged combustion in oxy-fuel combustion performs better than air-fuel combustion in reducing nitrogen oxidation.

We conclude that, apart from the equivalence ratio under high levels of CO2, the formation of NO x may be affected by the level of fuel in the mixture. But little is known about how variation in the level of fuel affects NO x formation in an O2/CO2 atmosphere. The fuel level limits the speed of the branched chain reaction.

In this study, variation in the chemical effects of CO2 on NO x formation from NH3 due to changes in the fuel level in mixtures with high CO2 concentrations was investigated via numerical simulation.

2 Methods

Combustion of CO2 diluted CH4/NH3 fuel with O2/CO2 as an oxidizer was simulated using a Chemkin Pro plug flow reactor. The reactor was 90 cm long and 26 mm in diameter, and was set in the isothermal range of 1073–1773 K. The inlet velocity was set to 30 cm/s. The initial temperature of the mixture was set to the same temperature as that of the reactor to maintain a residence time of 3 s, which is similar to that found in boilers in thermal plants burning pulverized coal.

The composition of the mixtures is listed in Table 1. The simulated fuel streams were doped with NH3 as the NO x precursor in volatiles from coal. The volume fraction of NH3 was set to 2% of the total CH4 fuel stream according to the ratio of N to C of lignite from Yunnan Province in China. O2 occupied 21% of the volume of the O2/CO2 oxidizer. The equivalence ratios of the mixture of CH4/NH3 and O2/CO2 were 0.6, 1.05, and 1.5 in Cases A1, A2, and A3, respectively. Each mixture in group A was diluted 10 times with CO2 in group B and 100 times in group C.

Table 1 Gas composition at the inlet of the reactor
Full size table

The numerical simulation was carried out using Chemkin Pro commercial software. Two reaction mechanisms were adopted: the GRI-Mech mechanism 3.0 (GRI30), and the mechanism developed by Mendiara and Glarborg (2009), referred to as M09 in this study. To study the effect of Reaction (1), the cases in Table 1 were also simulated with mechanisms without Reaction (1), namely the reverse reaction of CO+OH to form H+CO2. These mechanisms are referred to as IRGRI and IRM09.

The conversion ratio from NH3 to exhaust NO x (NO x CR) was defined as:

$${\rm{N}}{{\rm{O}}_x}\;{\rm{CR}} = {{{\rm{Exhaust\;N}}{{\rm{O}}_x}\;{\rm{ flow\;rate}}} \over {{\rm{Inflow\;N}}{{\rm{O}}_x}\;{\rm{ flow\;rate}}}}.$$

3 Results and discussion

An unbranched chain reaction mechanism is proposed to illustrate the chemical effect of CO2 on the radical pool (Fig. 1). The reactions CO2+H to form OH and CO, OH+H2 to form H2O and H can be taken as chain reactions without branching. The availability of O/H/OH, as well as the oxidation and reduction of N-containing species, can be affected by the relative dominance of the two types of chain reaction.

Results from the GRI30 and M09 mechanisms (Fig. 2) show coherence to a certain extent in predicting the conversion ratio of nitrogen element in methane flame to NO x (NO x conversion ratio) in an undiluted O2/CO2 atmosphere (Watanabe et al., 2011) and in simulating NH3 oxidation in an O2/CO2 atmosphere with diluted CH4 as fuel (Mendiara and Glarborg, 2009), and both mechanisms performed very well. The M09 mechanism was a little better than GRI30 in the latter simulation. Thus, simulated results for conditions in groups A and C were used only for mechanisms based on GRI30 and M09, respectively. The effects of Reaction (1) for NH3 oxidation in group B simulated by GRI30 and M09 were similar, so only results from GRI30 were used in this study.

Fig. 2
figure 2

Comparison of simulated values of NO x conversion ratio (CR) with experimental values from Watanabe et al. (2011) and Mendiara and Glarborg (2009) (Circles denote conditions with CO 2 levels of about 90% and diluted fuel; triangles denote conditions with about 70% CO 2 ; the values calculated using the GRI mechanism are shown as closed symbols while those calculated using the M09 mechanism are shown by open symbols)

Full size image

Figs. 3–5 compare the conversion ratios of NH3 to NO x at the end of the plug flow reactors using both mechanisms, under fuel-rich, stoichiometric, and fuel-lean conditions, respectively.

Fig. 3
figure 3

Conversion ratio of N to NO x at the end of the reactor vs. temperature under fuel-rich conditions ( Φ =1.5; A3, B3, and C3 correspond to conditions described in Table 1 )

Full size image
Fig. 4
figure 4

Conversion ratio of N to NO x at the end of the reactor vs. temperature under stoichiometric conditions ( Φ =1.05; A2, B2, and C2 correspond to conditions described in Table 1 )

Full size image
Fig. 5
figure 5

NO x conversion ratio at the end of the reactor vs. temperature under fuel-lean conditions ( Φ =0.6; A1, B1, and C1 correspond to conditions described in Table 1 )

Full size image

Fig. 3 shows the results for fuel-rich conditions. The NO x conversion ratio remained very low under undiluted and fuel-rich conditions (A3). NO x formation was increased by Reaction (1) under A3 and B3 conditions. For C3 conditions, NO x formation was inhibited at temperatures between 1273 K and 1373 K, and increased above 1573 K.

Fig. 4 shows the results for stoichiometric conditions. The NO x conversion ratio increased as the temperature increased. The conversion ratios of NO x from the mechanisms were nearly the same under 10 times dilution conditions. Reactions to form NO were facilitated without CO2 dilution, but inhibited under 100 times CO2 dilution conditions.

Fig. 5 shows the results for fuel-lean conditions. NO x formation was inhibited regardless of dilution, and the effect increased with CO2 dilution.

We can conclude from Figs. 3–5 that the chemical effect of CO2 on the conversion of NH3 to NO through Reaction (1) is impacted by both the equivalence ratio and CO2 level. NO formation is facilitated under fuel-rich conditions with undiluted oxy-fuel CO2 concentrations, but is inhibited under fuel-lean conditions and very high CO2 concentrations. When the inlet mixture was over 99% CO2, the reaction of CO2+H to form CO and OH led to a reduced conversion ratio of NH3 to NO, under all equivalence ratios. NO formation was slightly increased (or decreased) under near stoichiometric (or fuel-lean) conditions.

Fig. 6 shows the volume fractions of NO along the axis of the reactor at 1523 K, and equivalence ratios of 0.6, 1.05, and 1.5 without inlet CO2 dilution. NO is first produced in the reactor, and then part of the NO is reduced along the reactor under fuel-rich and stoichiometric conditions. Formation is facilitated under fuel-rich conditions due to Reaction (1), and NO destruction is inhibited due to Reaction (1) under fuel-rich and stoichiometric conditions, while NO is gradually reduced. The conversion ratio of NO was slightly higher when Reaction (1) was applied to the mechanism. There were no significant differences in the peak volume fraction of NO between the results from the two mechanisms under fuel-lean and stoichiometric conditions. NO is gradually reduced under stoichiometric conditions and this process was inhibited due to Reaction (1), while NO is not reduced under fuel-lean conditions. The trends of the NO volume fraction along the reactor at 1523 K are similar to those shown in Fig. 6 (A2) under condition B3 (fuel-rich), and similar to those shown in Fig. 6 (A1) under conditions B1 (fuel-lean) and B2 (near stoichiometric).

Fig. 6
figure 6

Profile of NO along the reactor under conditions A1, A2, and A3 at T =1523 K (A1, A2, and A3 correspond to conditions described in Table 1 )

Full size image

Fig. 7 shows mole fractions of radicals OH and H along the axis of the reactor at 1523 K, and equivalence ratios of 0.6, 1.05, and 1.5 without inlet CO2 dilution. The concentration of H was not always reduced by the reaction of CO2+H forming CO and OH. This result differs from that found by Mendiara and Glarborg (2009). As with the fuel-rich case, more OH and H were produced due to this reaction, while less H was produced under near stoichiometric conditions. H and OH are important intermediate species for the conversion of NH3 to NO, though the paths may differ under fuel-rich and stoichiometric conditions. Under fuel-rich conditions, the increase in OH and H resulted in a higher peak of NO, while under near stoichiometric conditions, Reaction (1) resulted in more OH but less H. The effect on the peak value of NO and its reduction after the peak was then weakened compared with the effects under fuel-rich conditions.

Fig. 7
figure 7

Profile of OH and H radicals along the reactor under conditions A1, A2, and A3 at T =1523 K. Solid lines denote the GRI30 mechanism, while dashed lines denote the IRGRI mechanism. A1, A2, and A3 correspond to conditions described in Table 1

Full size image

Fig. 8 shows the main pathways through which NO was reduced to N2 after its peak. The main path was NO+H to form HNO (Reaction (3)), then HNO was reduced to NH by CO and H2 (Reactions (4) and (5)). At the same time, HNO was reduced to NO by H (Reaction (6)) and NO was reduced to N2O by NH (Reaction (7)), and at the end N2O was reduced to N2 by H (Reaction (8)) and M (Reaction (9)). In an inferior path, NO was directly reduced to N2 by NH (Reaction (10)) and N. The availability of H is the major reason for the change in the reaction rate of NO reduction induced by changing the existence of Reaction (1). Reactions (3) and (8) from NO to HNO were facilitated by the increased level of H, resulting in a greater formation of NH. Hence, the reactions from NO to N2O and N2, and the reaction from N2O to N2, were facilitated.

$${\rm{H}} + {\rm{NO}} + {\rm{M}} \leftrightarrow {\rm{HNO}} + {\rm{M}},$$
(3)
$${\rm{NH}} + {\rm{C}}{{\rm{O}}_2} \leftrightarrow {\rm{HNO}} + {\rm{CO}},$$
(4)
$${\rm{NH}} + {{\rm{H}}_2}{\rm{O}} \leftrightarrow {\rm{HNO}} + {{\rm{H}}_2},$$
(5)
$${\rm{HNO}} + {\rm{H}} \leftrightarrow {{\rm{H}}_2} + {\rm{NO}},$$
(6)
$${\rm{NH}} + {\rm{NO}} \leftrightarrow {{\rm{N}}_2}{\rm{O}} + {\rm{H}},$$
(7)
$${{\rm{N}}_2}{\rm{O}} + {\rm{H}} \leftrightarrow {{\rm{N}}_2} + {\rm{OH}},$$
(8)
$${{\rm{N}}_2}{\rm{O}}\left({ + {\rm{M}}} \right) \leftrightarrow {{\rm{N}}_2} + {\rm{O}}\left({ + {\rm{M}}} \right),$$
(9)
$${\rm{NH}} + {\rm{NO}} \leftrightarrow {{\rm{N}}_2} + {\rm{OH}}.$$
(10)
Fig. 8
figure 8

Main reaction path for NO reduction (the bolded one is for GRI30 only, and the grey ones are for IRGRI only)

Full size image

The reaction CO2+H to form OH and CO resulted in a higher H level at the peak NO concentration and eliminated the consumption of H during NO reduction after the peak. The higher H level was the main factor contributing to the reduction of NO under conditions without CO2 dilution. In a reducing atmosphere, the reduction of NO was facilitated due to the higher H level resulting from Reaction (1). Lower H levels were obtained under stoichiometric conditions. Under lean conditions, NO would not be reduced once it has been produced in the reaction zone.

Fig. 9 shows the volume fraction of NO along the reactor at 1523 K, 100 times CO2 dilution, and an equivalence ratio of 1.05. The curves with an equivalence ratio of 0.6 and 1.5 are similar to those shown in Fig. 9. This shows that the formation of NO is inhibited due to Reaction (1) regardless of the equivalence ratio, once the reactant is diluted 100 times in CO2. Also, under these conditions none of the NO formed is reduced after it has reached its maximum concentration. The start of NO formation is delayed due to Reaction (1). With a high degree of dilution in CO2, concentrations of all the species are low, except for CO2. The H radical tends to be caught and consumed by CO2, so the reaction is led by the unbranched chain reaction in the initial stage of the reaction. The branched chain reactions become important only when active chain carriers of H are accumulated at sufficiently high levels. With the IRGRI mechanism, the branched chain reactions lead the reaction at the beginning. The delay of the branched chain reaction results in reduced H, O, and OH levels, which are vital to the oxidation of nitrogen species to NO. Therefore, the formation of NO decreases due to the chemical effect of CO2. NO is not reduced even after its peak under high CO2 dilution fuel-rich conditions because of low H availability in the reaction zone.

Fig. 9
figure 9

Profile of NO along the reactor under condition C2 at T =1523 K

Full size image

Table 2 shows the indirect effect of Reaction (1) on the formation and reduction of NO in terms of the conversion ratio. The promotion effect of Reaction (1) on NO during the formation period is weakened gradually with CO2 dilution. NO formation is inhibited under high CO2 dilution and fuel-lean conditions by Reaction (1) indirectly (Fig. 9). The trends of the indirect effect of CO2 dilution on NO reduction are similar to those on NO formation. But under high dilution conditions, regardless of the equivalence ratio, the promotion effect on NO reduction resulting from Reaction (1) dissipates. Under the condition of 100 times CO2 dilution, Reaction (1) has no effect on the reduction process of NO.

Table 2 Effect of Reaction (1) on formation and reduction of NO at T =1523 K
Full size table

Under fuel-rich conditions without CO2 dilution, NO is rapidly produced and then most of it is subsequently reduced. The conversion from NH3 to NO is facilitated due to the increased radical pool resulting from Reaction (1). On the other hand, the process of reduction of NO to N2 relies on the residual radical pool of O/H/OH after the NO peak, especially the pool of H. These two opposite effects result in only a slight change to the NO conversion rate at the end of the reaction. Under the conditions with 10 times CO2 dilution, the branched chain reactions are weakened, and the radical pool is eliminated. The influence of Reaction (1) on the radical pool is then weakened and the increase in NO is reduced. After the peak of NO, the consumption of H due to Reaction (1) inhibits the reduction of NO, which results in a higher NO x concentration at the end of the reaction. Under conditions with even higher CO2 dilution, the increase in NO x formation is even smaller, and because of the lower level of reactants, the relative concentration of H at the peak of NO is lower. NO is hardly reduced after the peak. So at the end of the reaction, the level of NO is higher when Reaction (1) is adopted in the mechanism.

Under stoichiometric conditions, NO formation before its peak is slightly facilitated without CO2 dilution due to Reaction (1). As the dilution with CO2 increases, this effect on NO is severely inhibited and the reduction effect gradually weakens (Table 2). The indirect effects of Reaction (1) on the NO x conversion ratio at the end of the reaction are changed from facilitating without CO2 dilution, to insignificant with 10 times CO2 dilution, and then to inhibition with 100 times CO2 dilution.

Under fuel-lean conditions, formation of NO is inhibited by Reaction (1) regardless of the extent of dilution with CO2. NO formed is not reduced under fuel-lean conditions.

Fig. 10 shows the ratio of residual NH3 at the end of the reaction under fuel-rich conditions at different temperatures. There are temperature ranges where the conversion of NH3 is initiated but not completed. In these temperature regions, the conversion of NH3 in the plug flow reactor simulated with the M09 mechanism is slower than that with the IRM09 mechanisms, under conditions of 100 times CO2 dilution. Simulations with GRI30 and IRGRI yield the same conclusion. Under conditions of 10 times dilution, the conversion of NH3 is faster with the GRI30 mechanism. The conversion of ammonia under fuel-rich conditions is a process in which H atoms are gradually lost. The N-H bonds are attacked by active radicals such as O, H, and OH. As described above, the radical pool of O/H/OH is affected by Reaction (1) and the effects differ with different CO2 dilution times. The NH3 conversion is faster with a denser radical pool. The role of Reaction (1) becomes more important as the dilution with CO2 increases. The weakening of the branched chain reaction results in relatively light radical pools, and the conversion of NH3 is consequently inhibited.

Fig. 10
figure 10

Residual ratio of NH3 at the end of the reactor under fuel-rich conditions ( Φ =1.5; A3, B3, and C3 correspond to conditions described in Table 1 )

Full size image

Fig. 11 shows the NH3 residual ratio at the end of the reaction at different temperatures under fuel-lean conditions. The conversion of NH3 is inhibited due to the existence of Reaction (1) under conditions with 100 times CO2 dilution. Under conditions with 10 times CO2 dilution or no dilution, the conversion of NH3 is not influenced by the absence of Reaction (1) in the mechanism. The results are similar under conditions with an equivalence ratio of 1.05. Under conditions with 10 times CO2 dilution or no dilution, H in the reaction space is consumed rapidly by the branched chain reaction. The influence of the reaction on the radical pool and the whole reaction space is then limited. Therefore, Reaction (1) has almost no influence on NH3 conversion. Under conditions with 100 times CO2 dilution, the dominance of the branched chain reaction diminishes and the radical pools decrease. The conversion of NH3 is then inhibited (Fig. 10).

Fig. 11
figure 11

Residual ratio of NH3 at the end of the reactor under fuel-lean conditions ( Φ =0.6)

Full size image

Fig. 12 shows the conversion ratio of nitrogen to main products at the end of the reaction under an equivalence ratio of 1.5 with 10 times CO2 dilution. The final products of nitrogen comprise mainly N2 and NO x . N2O and HNCO would be formed under relatively low temperatures, between 1100 K and 1400 K. N2O tended to be produced under conditions of low temperatures, low equivalence ratios, and high CO2 dilution levels, while HNCO was produced at low temperatures and high CO2 dilution levels. The chemical effect of CO2 played a more important role on HNCO than on N2O. Formation of HNCO was facilitated due to the chemical effect.

Fig. 12
figure 12

Conversion ratio of nitrogen to main products in condition B3 ( Φ =1.5, solid lines denote the GRI30 mechanism, while dashed lines denote the IRGRI mechanism)

Full size image

4 Conclusions

In this paper, the indirect relationship between the fuel level in mixtures of high CO2 level and the effect of CO2 via the reaction CO2+H to form OH and CO on NO x formation from NH3 were studied through numerical simulation.

Generally, the reaction CO2+H→CO+OH facilitates NO formation under fuel-rich conditions without CO2 dilution, but inhibits NO formation under fuel-lean conditions with high CO2 dilution. The radical pool of the chain carriers O/H/OH is affected, thereby influencing the formation and reduction of NO. The reduction of NO is affected mostly by H.

An unbranched chain reaction mechanism was proposed to illustrate the chemical effect of CO2 on the radical pool and NO x formation and reduction.

As the CO2 concentration increases, the mutual promotion of the unbranched chain reaction and the branched reaction gradually changes to mutual competition for H, which gradually weakens the promotion of NO formation. Under conditions with CO2 levels higher than 90%, however, the formation of NO is inhibited.

The competition for H as CO2 concentration increases influences the conversion of NH3, changing from facilitation to inhibition of NH3 conversion under fuel-rich conditions, and from no influence to inhibiting conversion under near stoichiometric and fuel-lean conditions.