The experimental samples were obtained from the fast-growing grass planting base in Yangshi Town, Loudi, Hunan. The industrial and elemental analyses of the grass are shown in Table 1. Analytical- and reagent-grade Al₂O₃, CaO, Ca(OH)₂, and SiO₂ with the purities of over 99, 98, 95, and 99%, respectively; all these are typical industrial additives that are applied to solid fuels. In this experiment, the following samples were tested: 1) samples of pure fast-growing grass; 2) mixture of fast-growing grass samples and additives at concentrations of 3, 5, 10, and 15%. All samples were pre-treated by drying in an electrothermal drying oven at 106°C for 12 h, followed by grinding and then sieving the particles through an 80-mesh sieve to obtain particles smaller than 200 µm in diameter. Finally, the samples were labelled and stored in dry containers for subsequent use.

The experimental system is illustrated in Figure 1. The combustion test was performed in a quartz tube reactor of a tube furnace with an inner diameter of 43 mm and a length of 600 mm (OTF-1200X, Hefei Kejing Material Technology Co., Ltd.). The temperature in the furnace was adjustable between room temperature and 1,200°C. NO _x emission was measured using a gas analyzer (Testo 350, Testo in Germany) connected to a computer that monitors and saves the experimental data in real time.

The experiment was performed using the following steps: 1) enter the target temperature into the temperature-control tube furnace and start heating; 2) open the gas cylinder valve and adjust N₂ and O₂ flowmeters such that their flows are 0.8 and 0.2 L/min, respectively; 3) after reaching the target temperature, weigh 0.2 g of each sample and spread it out in a porcelain boat; then, push the boat to the center of the tube furnace quickly; 4) place the probe of the gas analyzer at the gas outlet and observe the collected data on the computer. The combustion reaction was considered complete when the volume concentration of NO _x decreases to 2 ppm. Each set of experiments was repeated 3 times and average values were calculated.

In this experiment, combustion is considered to be complete when the NO _x emission concentration falls to 2 ppm; the elapsed time is the burnout time t. Calculation formula references (Liu et al., 2018; Xu et al., 2021).

① The average concentration (AC) is calculated as follows: AC = ∫ 0 t Cdt t (1) Where, the numerator on the right-hand side refers to the integral of the gas concentration for the reaction time, t represents the burnout time, and the AC is measured in ppm.

② Calculation of NO _x volume (V)

Considering that the V is negligible compared to the input air volume, the V is approximately calculated as follows: V = Q × t × AC 10 6 (2) Where, the V is measured in L; Q represents the input air flow in L/s.

③ The mass (M) of produced NO _x is calculated as follows: M = V × 273.15 T × 22.4 × M g (3) Where, the unit of M is g; T(K) represents the temperature of the gas; M _g (g/mol) represents the molar mass of NO _x .

④ The denitration rate (η) is calculated as follows: η = M 0 − M 1 M 0 × 100 % (4) Where, M ₀ represents the mass of NO _x produced by combustion of samples without additives; M ₁ is the mass of NO _x produced by combustion of samples with additives.

The reaction pathway for NO _x generation by the oxidation of fuel-N (Vermeulen et al., 2012; Shah et al., 2019) shown in Figure 2 displays that the following two main sources of NO _x in the flue gas: the conversion of volatile-N into volatile-NO _x ; and the conversion of char-N into char-NO _x . The conversion of fuel-N mainly occurs at temperatures below 900°C; therefore, the NO _x emission levels were measured and analyzed in the flue gases from combustion of fast-growing grass at different temperatures below 900°C.

As shown in Figure 3, NO _x generated from volatile combustion appeared late (at 60s) in the NO _x emission curve at 600°C; the NO _x concentration (peak intensity) was rather low at 19.50 ppm and the combustion reaction lasted for 179 s. At a relatively low temperature of 600°C, not only is the volatilization slow, but the chemical reaction that generates volatile-NO _x from oxidation of volatile-N through intermediate products HCN and NH₃ is also extremely slow; therefore, NO _x concentration is relatively low and the time of NO _x emission is rather long (Li et al., 2018; Shah et al., 2018). This is similar to that of microalgae combustion at 600°C (Chen et al., 2017). Unlike at 600°C, two peaks were observed in the NO _x emission curves at 700, 750, 800, and 850°C, corresponding to the volatile and char combustion stages. At temperatures of 700, 750, 800, and 850°C, the first peak was observed much sooner at around 13 s; the peak intensities increased gradually as the temperature increased, reaching 47.50, 70.50, 76.50, and 91.50, respectively. This is similar to results observed for microalgae combustion (Sun et al., 2019; ZhaoSu, 2019), solid biomass combustion (LiChyang, 2020) and the mixed combustion of sewage sludge and rice husk (Xu et al., 2021b). The second peaks were observed in the NO _x emission curves at 77, 58, 62, and 67 s with intensities of 9.00, 11.50, 6.50, and 10.00 ppm, for 700, 750, 800, and 850°C, respectively; the intensities of these peaks were much lower than those of the first peaks. Similar observations have also been reported in previous studies (Lane et al., 2013; Sun et al., 2019). These phenomena were attributed to the generation of char-NO _x from combustion of char-N of the fast-growing grass. The mechanism is shown in Eq. 5 (Thomas et al., 1993; Molina et al., 2000; LI et al., 2007). char − N + 1 2 O 2 → NO (5)

Currently, NO _x removal is mainly achieved by adding additives to urea or ammonia solutions to introduce SNCR. However, these methods are damaging to the equipment and have high operating costs (Pudasainee et al., 2012); therefore, they are not applicable to small combustion equipment. Consequently, this study focused on mixed combustion of fast-growing grass along with additives to reduce NO _x emission at different combustion temperatures using varying mass fractions of four different additives.

As shown in Table 2, the additives have an inhibitory effect on the generation of volatile-NO _x . At 600°C, the first peak of NO _x emission from combustion of fast-growing grass was significantly reduced after the introduction of additives. The peaks were reduced to average values of 34.40, 25.64, 16.01, 21.80% upon addition of SiO₂, Al₂O₃, CaO, and Ca(OH)₂, respectively. There are different reasons for the peak reduction for different additives. SiO₂ inhibits N-conversion to NO _x precursors (HCN, HNCO, and NH₃) during combustion (Ren et al., 2010); similarly, Al₂O₃ inhibits NO _x precursors generation (Ren et al., 2010) and is often used as the carrier of the denitration catalyst (Ma. et al., 2015; Li et al., 2013; Yao et al., 2017; QiShan et al., 2020). On one hand, CaO can catalyze HCN to N₂, as shown in Eq. 6 (Tan et al., 2009); on the other hand, it can also react with HCN and NH₃ to form N₂ during the pyrolysis stage, as shown in Eqs. 7–9 (Fu et al., 2014; Liu et al., 2015; Zhang et al., 2017). Similarly, Ca(OH)₂ decomposes into CaO and H₂O at high temperatures exhibiting a peak-reducing effect, as shown in Eq. 10 (Liu et al., 2015). 2 C i H j + 2 HCN → N 2 + ( j + 1 − k ) H 2 + 2 C i + 1 H k (6) CaO + 2 HCN → CaC N 2 + CO + H 2 (7) Ca C x N y → Ca C x + y 2 N 2 (8) Ca C x + yN H 3 → Ca C x N y + 3 2 y H 2 (9) C a ( OH ) 2 → CaO + H 2 O (10)

Surprisingly, just adding more additives did not have a better effect on reducing the peak. Optimal mass fractions were determined for the additives. For example, SiO₂, Ca(OH)₂, and CaO exhibited the best NO _x removal at a mass fraction of 5% with peak-reducing rates of 43.59, 35.90, and 19.64%, respectively. However, Al₂O₃ obtained the best reduction effect at a mass fraction of 3%, decreasing the peak intensity by 30.77%. These phenomena result from the agglomeration or blanketing of active components on the additive surface when the amount of additives reaches a certain value, which in -turn reduces of the number of active sites. Similar observations have also been reported by previous studies (Yu et al., 2017a; Sun et al., 2019). In order to further study the effects of additives on NO _x emission from the combustion of fast-growing grass at 600°C, the emission curves of the four additives at their optimal mass fractions (see Figure 4) were determined.

As shown in Figure 4, the additives affected not only the first peak of NO _x emission, but also its appearance time. The peak appeared 13 s sooner upon adding 5% SiO₂, whereas the peak appeared 16, 5, and 9 s later upon adding 5% Ca(OH)₂, 3% Al₂O₃, and 3% Al₂O₃, respectively. Similar phenomena have also been reported by Sun’s research group (Sun et al., 2019).

As the temperature changes, the effect of additives on lowering the first peak of NO _x emission varied, as shown in Table 3. Overall, at 700, 750, and 800°C, Ca(OH)₂ casued the most decrease in the first peak, whereas Al₂O₃ exhibited the same effect at 850°C. This is because of the relationship between the catalytic activity of the additives and the temperature (Sun et al., 2019), which caused the efficiency of NH₃ conversion to N₂ to differ at different temperatures.

Similar to the situation at 600°C, additives exhibited different rules to efficiently reduce the first peak at the same temperature, with varying mass fractions. For instance, at 700°C, Al₂O₃ and Ca(OH)₂ worked best at 3%, reducing the peak by 44.21 and 43.16%, respectively. At 800°C, however, Al₂O₃ and Ca(OH)₂ exhibited the best results at 5%, with the peak intensities dropping by 24.84 and 28.11%, respectively.

As shown in Figure 5, unlike the situation for 600°C, additives had little effect on the appearance time for the first peaks; the time for the peaks to appear was around 15 s after the addition of the four additives. However, the effect of different additives on the second peak varies, indicating that the additives have different effects on the production of char-NO _x . Al₂O₃ and SiO₂ reduced the second peak down to a certain extent, whose effect is unapparent in the high temperature zone (800°C and 850°C); however, it is quite obvious in the low temperature zone (700°C and 750°C). Unlike the two aforementioned additives, CaO increased the second peak to a certain extent, which is the most obvious at 700°C, and raised the second peak by 61.11% compared to that without additives. This may be because CaO can facilitate the conversion of char-N to char-NO _x (Liu et al., 2018b) and function most efficiently at 700°C (HayhurstLawrence, 1996). These phenomena will be further investigated in subsequent studies. Interestingly, the situation of Ca(OH)₂ is opposite to that of CaO. It all depends on whether the H₂O generated after decomposition of Ca(OH)₂ reacts with the HCN and char-C to produce CO, the reaction mechanism for which is shown in Eq. 11 (Liu et al., 2017; SchäferBonn, 2002) and Eq. 12 (Heikkilä et al., 2021; HermannHüttinger, 1986; MoulijnKapteijn, 1995); it also matters whether CO can convert NO to N₂, whose reaction mechanism is shown in Eq. 13, 14 (Chan et al., 1983; Thomas et al., 1993; AarnaSuuberg, 1997; Dong et al., 2007; Gong et al., 2021). HCN + H 2 O → N H 3 + CO (11) char − C + H 2 O → CO + H 2 (12) CO+C ( O ) → C O 2 + C f (13) C f + NO → C ( O ) + 1 2 N 2 (14)

Figure 6 presents the denitration rate for the combustion of fast-growing grass with four mass fractions of four different additives at different temperatures. As shown in Figure 6, at different temperatures, four additives exhibited different NO _x removal effects. At 600 and 750°C, SiO₂ and CaO displayed the best denitration effects with average denitration rates of 39.80 and 42.69%, respectively. At 700, 800, and 850°C, Al₂O₃ demonstrated the best denitration effect with average denitration rates of 34.56, 17.27, and 22.71%, respectively, The denitrification effect is better than CaCO₃ (Qi et al., 2017)and calcium and Ca-Fe oxides additives (Yu et al., 2017b; Qi et al., 2017). The addition of CaO at 700°C showed no denitrification effect and the average yield of NO _x increased by 10.86%, this is because it promotes the production of char-NO _x .

As shown in Figure 6, at the same temperature, the same additive displayed different NO _x removal effects (η) for different mass fractions. At 600 °C, Ca(OH)₂ worked best (η = 63.28%) at a mass fraction of 5%; at 700°C, Al₂O₃ showed the best effect (η = 50.34%) at 3%, whereas CaO failed to make any difference but rather increased NO _x emission; at 750°C, Ca(OH)₂ worked (η = 57.44%) the best at a mass fraction of 3%; at 800°C, Al₂O₃ worked best (η = 27.05%) at a mass fraction of 15%; at 850°C, Al₂O₃ worked best (η = 27.28%) at a mass fraction of 3%.

According to the peak-reducing effect and NO _x removal effect of the additives on NO _x emission from the combustion of fast-growing grass, the optimal control conditions at different temperatures are as follows: 1) at 600°C, 5% SiO₂ is the best, with the highest peak-reducing rate and NO _x removal efficiency; 2) at 700 and 750°C, 3% Al₂O₃ and 3% Ca(OH)₂ are the best, respectively; 3) at 800°C, 5% Ca(OH)₂ has the best peak-reducing effect; however, it has a much lower (9.89% lower) NO _x removal efficiency than that of 15% Al₂O₃; 15% Al₂O₃ has the highest NO _x removal efficiency with a slightly lower (3.92% lower) peak-reducing effect than that of 5% Ca(OH)₂; therefore, at 800°C, 15% Al₂O₃ is the best; 4) at 850°C, 3% SiO₂ and 3% Al₂O₃ are better and acceptable, with a gap of only 0.91% in the peak-reducing effect and only 0.06% in the NO _x removal efficiency.