A Strategic NH₃-Dosing Approach for the Minimization of N₂O Production During NH₃-SCR Reactions over Cu-SSZ-13 Catalysts

Abstract

The undesired production of N₂O during NH₃-SCR reactions is investigated over a reference commercial Cu-CHA catalyst. Steady-state experiments performed in the 150–500 °C temperature range exhibit a bimodal trend in the N₂O formation profile, confirming the existence of two different reaction mechanisms occurring at low and high temperatures. Focusing on a low-to-medium T-range, N₂O production, usually ascribed to NH₄NO₃ formation and decomposition, increases with the NO₂/NO_x ratio. However, an excess of NO₂ leads to a decrease in the N₂O release due to ammonium nitrate deposition and catalyst clogging phenomena. Steady-state and dynamics experiments show the promoting effect of both NH₃ feed concentration and NH₃ storage on N₂O production at T > 200 °C. Surprisingly, N₂O decreases with increasing NH₃/NO_x ratio at lower temperature. A novel approach based on the strategic injection of NH₃ is also applied to mitigate the N₂O formation while maintaining high deNO_x activity. Remarkably, complete NO_x conversion and ~ 11% N₂O saving are achieved (with inlet NO₂/NO_x = 0–0.5) at temperatures exceeding 200 °C; in addition, a peculiar behavior is observed in the N₂O profile, which increases and decreases when adding and removing NH₃ from the feed, respectively. Notably, the opposite trend is observed in the N₂O profile at 200 °C. When under Standard SCR conditions, this so far unreported observation challenges the NH₄NO₃ formation route for N₂O and suggests the existence of different controlling phenomena at different temperature regimes: i) the Cu/redox chemistry at T ≤ 200 °C and ii) the NH₃ storage at higher temperature, ideally up to 300 °C.

1 Introduction

In order to meet the environmental regulations aimed at reducing greenhouse gases and polluting emissions, diesel vehicles have been recently required to drastically reduce their engine-out harmful exhausts. the selective catalytic reduction of nitrogen oxides by injection of urea (NH₃-SCR) is currently considered the best available aftertreatment technology for NO_x (NO + NO₂) emission control in diesel or other lean-burn engine vehicles [1]. Depending on the NO₂/NO_x feed ratio, the reduction of NO_x typically proceeds according to the following stoichiometries, named Standard (1), Fast (2), and NO₂ (3) SCR reactions [1]:

$$4\text{ NO }+ 4\text{ NH}_{3} +\text{ O}_{2} \to 4\text{ N}_{2} + 6\text{ H}_{2}\text{O}$$

(1)

$$\text{NO }+\text{ NO}_{2} + 2\text{ NH}_{3} \to 2\text{ N}_{2} + 3\text{ H}_{2}\text{O}$$

(2)

$$6\text{ NO}_{2} + 8\text{ NH}_{3} \to 7\text{ N}_{2} + 12\text{ H}_{2}\text{O}$$

(3)

Typical SCR catalysts, such as Cu- or Fe-exchanged zeolites, achieve very high deNO_x performances above 200 °C [2,3,4]: as a consequence, over the years, the NH₃-SCR concept has been implemented by several automotive companies [1, 5]. Despite its effectiveness, however, some aspects still represent a concern. In particular, several side reactions can take place over the entire operating temperature window of the SCR reactions [6, 7], leading to the undesired production of nitrous oxide (N₂O), especially over Cu-based zeolite catalysts [7,8,9,10]. The generated amount of N₂O may differ with temperature and with the zeolite framework (e.g., Cu-CHA, Cu-MFI, and Cu-BEA), with small pore zeolites generally producing less amounts of nitrous oxide [8, 11]. This feature, in addition to the high deNO_x activity and high hydrothermal stability, makes Cu-CHA the most efficient and most used catalysts for NH₃-SCR automotive applications [11,12,13]. Nevertheless, N₂O being a powerful greenhouse gas (GHG) with a global warming potential (GWP) that is ~ 300 times that of CO₂ [6, 7, 9], it is highly desirable to reduce its emissions, even if small, in view of future regulation guidelines. In this sense, a good grasp of the N₂O formation mechanism during NH₃-SCR reactions, which is still highly debated, is urgently needed. Generally, N₂O production can occur through several side SCR reactions [6, 7, 11, 14,15,16], as listed below:

$$4\text{ NO }+ 4\text{ NH}_{3} + 3\text{ O}_{2} \to 4\text{ N}_{2}\text{O }+ 6\text{ H}_{2}\text{O}$$

(4)

$$4\text{ NO}_{2} + 4\text{ NH}_{3} +\text{ O}_{2} \to 4\text{ N}_{2}\text{O }+ 6\text{ H}_{2}\text{O}$$

(5)

$$8\text{ NO}_{2} + 6\text{ NH}_{3} \to 7\text{ N}_{2}\text{O }+ 9\text{ H}_{2}\text{O}$$

(6)

$$4\text{ NO }+ 2\text{ NH}_{3} \to 2\text{N}_{2} +\text{ N}_{2}\text{O }+ 3\text{ H}_{2}\text{O}$$

(7)

$$2\text{ NH}_{3} + 2\text{ O}_{2}\to \text{ N}_{2}\text{O }+ 3\text{ H}_{2}\text{O}$$

(8)

On Cu-based catalysts, Zhang et al. [7] suggested the non-selective ammonia oxidation (NSNO) and non-selective catalytic reduction (NSCR) reactions to be the main N₂O formation routes. Noteworthy, the characteristic bimodal behavior (i.e., two different peaks detected in the 150–600 °C T-range) observed by many authors in the steady-state N₂O production profile [6, 17] pointed out the existence of at least two different formation mechanisms depending on the operating temperature. At high temperature, besides NSNO, N₂O formation is ascribed to the reaction between NH₃ and gaseous NO through an Eley–Rideal mechanism [7]. The involvement of different sites in the low-T N₂O formation is also proposed [18]. At low temperature, Langmuir–Hinshelwood reactions (occurring between NH₃ and NO_x species both adsorbed on the catalyst surface) are widely acknowledged to take part in the N₂O formation, but the precise nature of the involved intermediates is still a topic of discussion [7, 11, 17]. Several authors suggested the NO₂ concentration (the NO₂/NO_x feed ratio can vary in the presence of a diesel oxidation catalyst, DOC [19, 20]), as well as the NH₃ concentration, to promote the N₂O process, ascribing this to the formation and decomposition of nitrate intermediates [6, 7, 21]. Olsson et al. [22] suggested ammonium nitrate precursors to react with oxygen to form N₂O. Other works suggested the low-T N₂O formation to derive from NH₄NO₃ decomposition [6, 23, 24]. More recently, the ammonium nitrate route when under standard SCR conditions has been challenged in the literature, ascribing instead the N₂O formation to the active participation of Cu sites [11, 17], which are in turn involved in a SCR redox mechanism (Cu^II ↔ Cu^I) [25,26,27,28]. Some authors proposed the low-temperature N₂O formation pathway to involve H₂NNO decomposition (by hydrogen transfer to a O₂ molecule adsorbed on a [Cu(NH₃)₂]⁺ complex) rather than NH₄NO₃, hence explaining the promoting effect given by the Cu loading [11, 29]. The role of different copper sites in N₂O formation, such as ZCu^IIOH and Z₂Cu^II, has also been investigated [30]. In addition, thanks to NH₃ + NO titration experiments, Xi et al. [17] observed more oxidized catalysts (i.e., with a higher fraction of Cu^II sites) to result in higher N₂O release.

Despite the many observations reported so far, there are still open points and a clear assessment of the N₂O formation route is still missing. Herein, we evaluate the N₂O production under different operating conditions over a reference Cu-CHA catalyst. For this purpose, a series of experimental runs, both under steady-state and dynamic conditions, has been performed by varying temperature and NO₂/NO_x and NH₃/NO_x ratios. In addition, a novel approach based on the strategic dynamic injection of NH₃ is shown as a possible method to mitigate the N₂O production without affecting the deNO_x activity under selected operating conditions. As a long-term goal, we want to provide additional information for a deeper understanding of the N₂O formation mechanism, thus supplying guidelines on how to optimize the SCR process so as to meet future environmental regulations.

2 Materials and Methods

The experimental study is conducted over a commercial copper exchanged chabazite catalyst in the form of washcoated honeycomb monolith with cpsi (cells per square inch) equal to 600. A core monolith sample with dimensions of approximately 1.1 cm² and 7.3 cm for square cross-sectional area and length, respectively, is drilled out for testing (V_monolith ≈ 8 cm³, Cu loading ≈ 4 g_Cu/L). As reported in the Supplementary Information (SI, see Sections SI.1 and SI.2), the Cu content and the Cu speciation (i.e., ZCu^IIOH/Z₂Cu^II ratio) are determined by NO + NH₃ titration and NO₂ adsorption/desorption experiments, in line with protocols described in previous publications [31, 32]: these reveal the catalyst to be characterized by around 510 µmol of total Cu sites and ZCu^IIOH and Z₂Cu^II fractions of 53 and 47%, respectively. Notice that, before testing, the sample underwent a hydrothermal aging pretreatment for 5 h at 600 °C (10% v/v H₂O in air).

The core catalyst sample, after being inserted in a tubular reactor with sample holder, is placed in an electric furnace. The lab-scale rig employed for this study is custom-designed to simulate the industrial NH₃-SCR processes: as a result, nitrogen is selected as the inert carrier gas. Nevertheless, this selection limits our ability to monitor the dynamic evolution of the N₂ production resulting from SCR reactions. By means of a peristaltic pump, liquid water is injected in a heated line (180 °C), vaporized and mixed with the gaseous feed stream. The outlet concentrations of gaseous species such as NH₃, NO, NO₂, N₂O, and H₂O are continuously recorded by a Matrix MG-5 FT-IR analyzer provided by Bruker. For additional information regarding the system set-up, please consult [2, 33, 34].

The different protocols adopted to investigate the side N₂O production over Cu-CHA during SCR reactions are described in the following. All the experiments are performed at a gas hourly space velocity (GHSV) of 40 kh⁻¹ and in the presence of water (8% v/v), to be more representative of realistic conditions.

2.1 Steady-State Activity Protocols

Steady-state NH₃-SCR conditions are investigated evaluating NO_x, NH₃ conversions, and N₂O formation by varying the temperature stepwise in a 150–500 °C range (steps of 25 °C from 150 to 250 °C; steps of 50 °C from 250 to 500 °C). First, NH₃ oxidation conditions are examined by feeding 500 ppm of NH₃ in 8% v/v of O₂. Then, 500 ppm of NO are added to the feed stream to estimate the N₂O production under Standard SCR conditions. NO₂ is also included: indeed, the steady-state procedure is replicated by varying the NO₂/NO_x feed molar ratio (NO_x = NO + NO₂) from 0.25 to 1. For certain conditions, the effect of NH₃/NO_x ratio (or “α parameter”) is also investigated by changing the NH₃ feed content from 500 to 1000 ppm (i.e., from α = 1 to α = 2), with fixed NO_x feed concentration (500 ppm).

2.2 Dynamic Protocols

Additional dynamic protocols are carried out to evaluate the effects of both pre-adsorbed and gaseous ammonia. First, an isothermal run with co-feed of NH₃ + NO_x (500 ppm + 500 ppm) is performed at 250 °C (here selected as reference reaction temperature) in a stream of O₂ + H₂O (both 8% v/v) to start the SCR reaction on an empty sample, i.e., not pre-saturated with NH₃ (Protocol A, see Figure SI-5A).

The same procedure is then replicated in the presence of a NH₃-preloaded catalyst (Protocol B, see Figure SI-5B). To better mimic realistic conditions, where some residual ammonia is always left onto the catalyst at the end of each engine cycle [35], half-storage NH₃ conditions (i.e., ϑ_NH3 is equal to 50% of the total storage achievable at 250 °C) are investigated. Cu-based catalysts can notoriously adsorb less amounts of ammonia at higher temperatures: the intended storage condition is then accomplished by pre-adsorbing NH₃ at a higher temperature (T_preads = 320 °C) with respect to the one selected for the following SCR step (T_reaction = 250 °C). The same result could also be achieved by varying the NH₃ adsorption time (t_preads) [36]. More details are reported in the SI (see Section SI.3). Notably, by comparing Protocol A and Protocol B, we can evaluate the effect of the ammonia coverage on both deNO_x activity and N₂O production.

Finally, according to Protocol C (see Figure SI-5C), 500 ppm of NO_x only are added to the O₂ + H₂O feed stream in the presence of a NH₃-preloaded catalyst (ϑ_NH3 = 50%). This time, by direct comparison of Protocol B and Protocol C, the effect of gaseous ammonia can be readily assessed.

2.3 NH₃ Strategic Injection Protocol

From a more practical point of view, a preliminary experimental procedure based on the strategic dynamic injection of NH₃ is performed to explore the possibility of minimizing N₂O production over Cu-SSZ-13 during NH₃-SCR processes while maintaining high NO_x conversions. This new approach is mainly based on pulsed NH₃ feed phases (Protocol D). In particular, starting from an ammonia preloaded catalyst (ϑ_NH3 = 50%), consecutive NH₃ pulses are performed at isothermal conditions in a constant feed stream of 500 ppm of NO_x in O₂ + H₂O (both 8% v/v). In particular, we can distinguish between two main phases: i) consumption of a certain amount of stored ammonia by 500 ppm of NO_x (with NH_3,in = 0 ppm, α₁ = 0) for a time interval equal to t₁ and ii) restoring of the consumed stored ammonia by pulsing an excess of gaseous NH₃ (= 1000 ppm, α₂ = 2) for a time interval equal to t₂. Particularly, t₁ and t₂ intervals are selected according to the results from Protocol C (described in Sect. 2.2) and to a balancing equation (9) based on the following constraints and assumptions: i) the reference NH₃ loading is always restored and ii) constant NO_x conversion ($\eta$). The described procedure (see Fig. 1) is replicated both under standard and fast SCR conditions (NO₂/NO_x = 0–0.5).

Fig. 1

Example of NH₃ pulsing strategy at 250 °C, starting from NH₃ preadsorbed at 320 °C over Cu-CHA. Phase 1: NO_x = 500 ppm, NH₃ = 0 ppm (α₁ = 0) for t₁. Phase 2: NO_x = 500 ppm, NH₃ = 1000 ppm (α₂ = 2) for t₂. Feed: O₂ = H₂O = 8% v/v. GHSV = 40 kh⁻¹

$${\alpha }_{1}{t}_{1}+{\alpha }_{2}{t}_{2}-\eta \left({t}_{1}+{t}_{2}\right)=0$$

(9)

3 Results and Discussion

The N₂O formation from SCR reactions is first investigated under steady-state conditions, mainly focusing on the effects of NO₂/NO_x and NH₃/NO_x ratios. Once identified, the temperature range corresponding to the highest N₂O production (Sect. 3.1), dynamic experiments, more interesting from a practical application point of view, are also performed (Sections. 3.2 and 3.3).

3.1 Steady-State NH₃-SCR Activity

3.1.1 Steady-State N₂O Formation

Steady-state results in the 150–500 °C T-range are shown in Fig. 2 in terms of NO_x/NH₃ conversions (see Figure SI-6 for the outlet concentrations) and N₂O production. High deNO_x activity, with NO and NH₃ conversions above 90% already at 200 °C, is achieved under standard SCR conditions (Fig. 2A). In line with the literature [6], an interesting shape is observed for the N₂O profile, which first increases up to ~ 10 ppm at 200 °C, then drops to ~ 3 ppm when the temperature is 350 °C, and increases again back to ~ 10–11 ppm at 500 °C. This specific bimodal behavior is in line with the theory according to which two different formation mechanisms may exist for N₂O depending on the reaction temperature range [6, 17, 21, 37]. Specifically, up to 250 °C ammonia and nitric oxide appear to be consumed according to an equimolar stoichiometry (NO:NH₃ = 1:1), and, in line with reaction (4), lead to the undesired production of N₂O. Above 250 °C, in line with literature findings [6, 18], NO and NH₃ conversions show divergent profiles: while NH₃ displays complete conversion in the whole temperature range, a non-negligible decrease is observed for NO (NO conversion ~ 75% at 500 °C). The high-temperature NH₃ overconsumption is typically ascribed to ammonia oxidation reactions, either selective (to N₂) or non-selective (to N₂O), which, together with non-selective catalytic reduction (NSCR) reactions may explain the decrease in NO conversion and the rise in the N₂O curve [6, 7, 38,39,40]. Indeed, as reported in Section SI.7, the contribution of the sole non-selective NH₃ oxidation reaction (NSNO) is not enough to describe the whole N₂O production observed at high temperature during standard SCR.

Fig. 2

Steady-state NH₃-SCR experiments on Cu-CHA versus temperature: NO_x and NH₃ conversions (%) and N₂O production (ppm) at GHSV = 40 kh⁻¹. Feed: NO_x = NH₃ = 500 ppm, O₂ = H₂O = 8% v/v in N₂. T = 150–500 °C. A Standard SCR (NO₂/NO_x = 0). B Fast SCR (NO₂/NO_x = 0.5)

The steady-state Fast SCR reaction (NO₂/NO_x = 0.5) is analyzed in Fig. 2B. Once again, ammonia is completely converted in the whole 150–550 °C temperature range; however, NO_x (NO + NO₂) conversion is higher with respect to that reached under standard SCR conditions, being ~ 100% and equimolar with NH₃ up to around 400 °C. The same bimodal trend is observed for the N₂O formation profile, which however is higher in concentration: this time, a maximum peak of ~ 16 ppm is detected between 200 and 250 °C (versus ~ 10 ppm under Standard SCR). This outcome is commonly ascribed to the formation (10) and decomposition (11–13) of NH₄NO₃, intermediate species coming from the interaction of NO₂ and NH₃ [15, 17, 21, 41]:

$$2\text{ NO}_{2} + 2\text{ NH}_{3} \to \text{ N}_{2} +\text{ NH}_{4}\text{NO}_{3} +\text{ H}_{2}\text{O}$$

(10)

$$\text{NH}_{4}\text{NO}_{3} +\text{ NO }\to \text{ N}_{2} +\text{ NO}_{2} + 2\text{ H}_{2}\text{O}$$

(11)

$$\text{NH}_{4}\text{NO}_{3} \to \text{ NH}_{3} +\text{ NO}_{2} + 1/2\text{ H}_{2}\text{O }+ 1/4\text{ O}_{2}$$

(12)

$$\text{NH}_{4}\text{NO}_{3} \to \text{ N}_{2}\text{O }+ 2\text{ H}_{2}\text{O}$$

(13)

Notice that the formation of surface nitrate groups is essential in order to lead to the NH₄NO₃ intermediate [17, 21]. In this regard, while under Fast SCR reaction NO₂ (which is present in the feed) can be directly adsorbed on the catalyst leading to the formation of surface nitrate intermediates, under Standard SCR conditions NO needs to be first oxidized to NO₂. In principle, this should explain the difference observed between Fig. 2A, B in terms of low-temperature N₂O production [6, 42]. However, while the NO + O₂ reaction step should play a key role in the Cu-nitrates formation so as to lead to the consequent release of nitrous oxide [42,43,44], no evidence of nitrate formation was detected at Standard SCR conditions according to recent publications [17, 45, 46]: accordingly, through both spectroscopic and experimental analysis, some authors [11, 17] have ascribed the low-temperature Standard SCR N₂O formation to Cu-redox processes rather than to the NH₄NO₃ route, as observed instead for Fast SCR. This aspect will be recalled in Section 3.3.

3.1.2 Effect of NO₂/NO_x Feed Ratio

The effect of NO₂ is also investigated by further varying the NO₂ feed concentration. The N₂O production profiles obtained under NH₃-SCR conditions with NO₂/NO_x equal to 0, 0.5, 0.75, and 1, respectively, and fixed equimolar NH₃, are shown in Fig. 3. N₂O selectivities are also displayed in the SI (Figure SI-7). As a general effect, in line with previous literature findings [6, 19], the higher the NO₂ feed content, the higher the N₂O production. In addition, a shift to higher temperatures is also observed: the ~ 10 ppm peak centred at 200–250 °C for Standard SCR conditions (purple squares, NO₂/NO_x = 0) grows up to ~ 60 ppm and shifts to 300–350 °C for NO₂/NO_x = 1 (yellow triangles). Beside this, from Fig. 3, we can observe that the mentioned correlation (i.e., higher NO₂ content and higher N₂O production) does not always hold in the low-to-medium temperature range (up to 300 °C), with output data that may appear confusing. To address this, both NO_x conversion and N₂O formation (y-axes) are plotted in Fig. 4 versus the NO_x/NO₂ feed ratio (x-axes) at 200, 250, and 300 °C only.

Fig. 3

N₂O production (ppm) versus temperature from steady state NH₃-SCR experiments on Cu-CHA with variable NO₂/NO_x ratios. GHSV = 40 kh⁻¹. Feed: NO_x = NH₃ = 500 ppm, O₂ = H₂O = 8% v/v in N₂. T = 150–500 °C. Legend: ■ NO₂/NO_x = 0; ♦ NO₂/NO_x = 0.5; ●) NO₂/NO_x = 0.75; ▲) NO₂/NO_x = 1

Fig. 4

Steady-state NH₃-SCR experiments on Cu-CHA vs. NO₂/NO_x ratio. GHSV = 40 kh⁻¹. Feed: NO_x = NH₃ = 500 ppm, O₂ = H₂O = 8% v/v in N₂, with variable NO₂/NO_x ratio. A NO_x conversion (%). B N₂O production (ppm). Legend: ♦ T = 200 °C; ■ T = 250 °C; ●) T = 300 °C

A clear increasing trend in the N₂O steady-state profile is only observed when enhancing the NO₂/NO_x ratio from 0 to 0.5 (Fig. 4B), with almost complete NO_x conversion reached in the same range (Fig. 4A): different effects on NO_x conversion and N₂O production are obtained at the three considered temperatures when further increasing the NO₂/NO_x feed ratio up to 1.

Similar to the behavior reported by Xi et al. [17], at 200 °C, NO_x conversion drops from 99 to 52% on increasing NO₂/NO_x from 0.5 to 1. Correspondingly, the N₂O make drops from ~ 19 to 5 ppm. This evidence is directly related to an inhibition phenomenon due to the deposition of ammonium nitrate into the chabazite pores, which leads to the blocking of the Cu active sites [17, 21, 41]. Specifically, NH₄NO₃ is usually formed on the catalyst surface at 200 °C and most of it decomposes around 300 °C according to reactions (10–13) [6]. When the operating temperature is too low, however, the NH₄NO₃ removal rate is limited, while its formation is still promoted by increasing the concentration of NO₂ in the gas feed composition (here NO₂/NO_x ≥ 0.5). This would lead to the accumulation of nitrates in the zeolite pores, with consequent severe mass-transfer limitations causing additional confinement to nitrates and enhancing their thermal stability [6, 17, 47, 48]. Accordingly, as observed in Fig. 4A, not only strong self-inhibition of N₂O formation occurs, but also NO_x conversion is limited because of the blocking of Cu-active sites [6, 49, 50].

To further rationalize this aspect, Fig. 5 shows the temperature programmed desorption (TPD) experiment (T-ramp = 200–550 °C, in N₂ only) realized at the end of the SCR reaction step with NO₂/NO_x = 0.75. Notably, a significant release of N₂O, NO₂, and NH₃ is observed with a peak centered at ~ 300 °C, in line with NH₄NO₃ formation (10), accumulation, and decomposition (12–13) due to the temperature increase [17]. The same TPD experiment was replicated after SCR reactions performed at 200 °C with different NO₂/NO_x ratios. The released N₂O profiles are compared in Figure SI-10: in line with what previously discussed and with refs. [6, 17], N₂O release during the temperature ramp is significant only for NO₂/NO_x > 0.5, being almost negligible for NO₂/NO_x ≤ 0.5.

Fig. 5

NO₂ (green), NH₃ (black), and N₂O (pink) released profiled during TPD experiment after NH₃-SCR at T = 200 °C (NO₂/NO_x = 0.75) on Cu-CHA. GHSV = 40 kh⁻¹. T-ramp: 200–550 °C at 15 °C/min, in N₂ only

At 300 °C (see Fig. 4), which is high enough for the NH₄NO₃ decomposition to occur, the steady-state N₂O production monotonically increases in the whole NO₂/NO_x range investigated, hence suggesting no accumulation of ammonium nitrate in the catalyst pores. Accordingly, high NO_x conversion values (> 94%) are always detected. As expected, an intermediate behavior is instead observed when running the steady-state SCR reactions at 250 °C, with outputs right in between the results achieved at 200 and 300 °C.

3.1.3 Effect of NH₃/NO_x Feed Ratio

To investigate the effect of NH₃ on the N₂O formation, steady-state SCR experiments were replicated by feeding 1000 ppm of NH₃ along with 500 ppm of NO_x (α = 2). The results, here collected in terms of steady-state N₂O production (light blue circles), are displayed in Fig. 6 and compared with that obtained for α = 1 (pink diamonds, discussed in Sect. 3.1.2). Again, different NO₂/NO_x ratios (0–0.25–0.5–0.75) and operating temperatures (T = 200–250–300 °C) are examined.

Fig. 6

Steady-state N₂O production during NH₃-SCR reactions on Cu-CHA vs. NO₂/NOx ratio—effect of NH₃/NOx ratio variation. GHSV = 40 kh⁻¹. Feed: NOx = 500 ppm, NH₃ = 500 ppm (α = 1) or 1000 ppm (α = 2), O₂ = H₂O = 8% v/v in N₂, with variable NO₂/NOx ratio. A T = 200 °C; B T = 250 °C; C T = 300 °C. Legend: ♦ α = 1; ● α = 2

As observed in Fig. 6B, C, an increase in the inlet NH₃ feed concentration from 500 to 1000 ppm leads to an increment of the steady-state N₂O make at 250 and 300 °C, respectively: this is not surprising, being in fact NH₃ one of the key reactants for the ammonium nitrate formation, hence strongly impacting the N₂O formation rate [19, 51, 52]. Note that NO₂ and NH₃ react stoichiometrically according to (10), and the overall NH₄NO₃ formation should be controlled by the limiting reactant. With enhancing the NH₃ content at fixed NO₂, the latter is expected to be the limiting species. However, the experimental results collected at 250 and 300 °C clearly contradict this expectation, suggesting instead NH₃, probably the adsorbed one, to take on the role of limiting reactant. This is in line with the reduced ammonia coverage expected with increasing temperature [36].

The reverse effect is instead observed at 200 °C (Fig. 6A): indeed, less N₂O production is detected on increasing the NH₃ feed content. To our knowledge, this behavior is here reported for the first time. For example, Liu et al. [6] observed a promoting effect on steady N₂O formation on increasing the NH₃/NO_x ratio from 0.6 to 1.4 over a Cu-SSZ-13 catalyst in the same temperature range. However, at 200 °C, the observed trend was unclear, with similar N₂O outlet concentrations recorded when going from a NH₃/NO_x ratio of 1 to 1.4 at a GHSV of 60 kh⁻¹. It is possible that, with respect to ref. [6], in our case, the unexpected ammonia effect detected at 200 °C is enhanced by the lower gas space velocity selected (40 kh⁻¹). The inhibiting effect of NH₃ on N₂O production observed in Fig. 6A may also be explained by considering the NH₄NO₃ formation and decomposition reactions. At such a low temperature, the ammonium nitrate removal is in fact limited, while its formation (10) and accumulation into the nitrate pores is possibly promoted by the increase of the reactant concentrations, such as NH₃, hence hindering the N₂O formation.

However, it is worth noting that the negative NH₃ effect displayed in Fig. 6A is observed not only in the presence of NO₂ in the feed stream (NO₂/NO_x > 0), but also under Standard SCR conditions (NO₂/NO_x = 0), under which, according to recent publications [11, 17, 45, 46], NH₄NO₃ is not involved as an intermediate species for the N₂O formation. This evidence possibly suggests the existence of different low-temperature N₂O formation mechanisms when in the absence of NO₂ in the feed (see Section 3.3 for more details).

3.2 Dynamic SCR Measurements

Dynamic measurements are herein described to evaluate the effect of both gaseous and stored ammonia on the transient N₂O formation transient profiles. Given that its maximum steady-state production (considering a low-to-medium T-range) shifts with both temperature and NO₂ feed content (see Fig. 3), as a good compromise 250 °C is here selected as the reference operating temperature in order to perform Protocols A, B, and C described in Sect. 2.2. The experimental results are displayed in Fig. 7 for the case of no NO₂ in the feed stream (NO₂/NO_x = 0).

Fig. 7

Dynamic Standard SCR measurements on Cu-CHA at 250 °C—effect of stored and gaseous NH₃. GHSV = 40 kh⁻¹. Feed: NO = 500 ppm, NO₂ = 0 ppm, NH₃ = 500 ppm (if present), O₂ = H₂O = 8% v/v in N₂. A Without pre-adsorbed NH₃, B with NH₃ pre-adsorbed at 320 °C (ϑ_NH3 = 50%), and C without gas phase NH₃ (ammonia is preadsorbed only, ϑ_NH3 = 50%)

Figure 7A shows NH₃, NO_x, and N₂O outlet concentration profiles under standard SCR conditions starting from an empty Cu-CHA catalyst, hence in the absence of pre-stored NH₃ (Protocol A). As soon as NO + NH₃ are added to the O₂ + H₂O feed stream (at t = 0 s), N₂O is slowly released together with a non-negligible amount of unreacted NO and NO₂. NH₃ being not pre-stored on the catalyst, the initial adsorption of ammonia represents the limiting step for the reaction under consideration. Only after about 600 s, once enough NH₃ is present on the catalyst, NO and NO₂ approach complete conversion, and a steady-state N₂O production of ~ 9 ppm is achieved.

The experimental results obtained by replicating the same protocol but over a NH₃-preloaded catalyst (Protocol B) are reported in Fig. 7B. The same steady-state results previously discussed in Fig. 7A are here obtained (i.e., complete NO_x conversion and ~ 9 ppm of N₂O), but with significant differences in the initial dynamics. Notably, in the presence of preadsorbed ammonia, no slip of NO or NO₂ is detected once NO + NH₃ are fed to the reacting system (in O₂ + H₂O) at t = 0 s. In addition, no slow dynamics are observed in the N₂O release, which quickly approaches the steady state value within just a few seconds. These differences clearly point out the effect of pre-adsorbed ammonia: in Fig. 7B, being pre-stored on the catalyst, at t = 0 s, ammonia is immediately available and ready to react with NO_x. At the same time, NH₃ being one of the key parameters for nitrous oxide formation, N₂O is suddenly released with faster dynamics compared to Fig. 7A. This appears to be in line with findings by Kamasamudram et al. [19], who observed the N₂O selectivity to linearly increase on varying the NH₃ storage on a Cu-zeolite sample from 0 to 0.2 g/L.

Figure 7C shows the experimental protocol replicated by adding 500 ppm of NO only in a O₂ + H₂O feed stream starting from a NH₃-preloaded catalyst (Protocol C). As soon as NO is step-fed (at t = 0 s), N₂O coming from the reaction between NO_x and pre-adsorbed NH₃ is released. Once its maximum production is reached (~ 4 ppm only), the N₂O outlet concentration starts to decrease along with the amount of preadsorbed NH₃: in fact, since no gaseous ammonia is fed, the NH₃-coverage of Cu sites is not restored. In addition, after an almost zero-emission period of about 70 s, unreacted NO is completely released. A steady-state value of ~ 475 ppm is reached, balanced by the production of ~ 25 ppm of NO₂ coming from NO oxidation (being in the presence of O₂) [40]. No unreacted stored ammonia is left on the catalyst surface, as confirmed by the TPD following the experiment, shown in Figure SI-11. Notably, due to the absence of gaseous ammonia able to top up the NH₃ storage when at t ≥ 0 s, a remarkably lower N₂O make is noted when comparing Fig. 7C, B.

The same considerations are applicable when reproducing the same protocols in the presence of NO₂ in the feed stream (NO₂/NO_x = 0.25–0.5–0.75). A description is reported in the Supplementary Information (Section SI.10). In accordance with the literature [6, 19] and with what described at 250 °C in the previous sections, the recorded N₂O production is higher with enhancing the NO₂ feed content.

3.3 NH₃ Strategic Dynamic Injection for N₂O Minimization

On the basis of what previously discussed, a novel approach based on the strategic injection of NH₃ has been established to assess the possibility of minimizing the N₂O production maintaining high NO_x conversions during the NH₃-SCR process. Specifically, we combined Protocols B and C to derive a new experimental approach (Protocol D), as described in Sect. 2.3.

The experiment performed at 250 °C with NO₂/NO_x = 0 is reported in Fig. 8A. Starting from an NH₃-coverage (ϑ_NH3) equal to 50%, 500 ppm of NO only (α₁ = 0) are added to the O₂ + H₂O feed mixture (at t = 0 s) for a time interval (t₁) of 70 s. Notice that the selection of this interval is based on the results obtained in Fig. 7C, according to which i) almost complete NOx conversion and ii) a N₂O production peak of only 4 ppm are achieved in the same period of time. As expected, during the first step (t₁ = 70 s, α₁ = 0), the same outcomes are also observed in Fig. 8A. Note that, simultaneously, the ammonia storage is decreasing from 50 to about 30% due to its reaction with NOx (see Figure SI-10). To prevent NO slip and maintain complete NO_x conversion, a consecutive NH₃-pulse is performed by feeding 1000 ppm (α₂ = 2) in the NO + O₂ + H₂O mixture. Again, a time interval (t₂) of 70 s is selected, based on equation (9) (where $\eta$ = 1). Notice that, by feeding NH₃ in excess with respect to NO, the initial ammonia storage is restored to about 50% (Figure SI-13). At the same time, complete NO_x conversion is preserved, with N₂O formation increasing up to ~ 10 ppm.

Fig. 8

NH₃ pulsed strategy on Cu-CHA at 250 °C (NH₃ pre-adsorbed at 320 °C, ϑ_NH3 = 50%). GHSV = 40 kh⁻¹. Feed: NOx = 500 ppm, O₂ = H₂O = 8% v/v in N₂. NH₃ = 0 ppm (α₁ = 0) during t₁ = 70 s; NH₃ = 1000 ppm (α₂ = 2) during t₂ = 70 s. A Standard SCR (NO₂/NOx = 0). B Fast SCR (NO₂/NOx = 0.5)

The procedure is then replicated for a total amount of 10 NH₃-pulses, sequentially alternating phases with α₁ = 0 and α₂ = 2. Interestingly, more than 99% of NO_x conversion is achieved at the end of the experimental protocol, together with a peculiar N₂O outlet profile. In fact, the latter increases when NH₃ is fed to the system, while it decreases when the ammonia feed is cut-off, with an average variation of about 3.5 ppm, in line with what described in Sect. 3.1.3. Moreover, the selected interval times (t₁ = t₂) ensure no NH₃ slip. The N₂O profiles obtained from this novel approach (in pink) and from a typical steady-state experiment starting from the same initial conditions (in blue) are also compared in Fig. 8A. Notably, by integration of the highlighted light blue area, we can claim that about 11.4% of the overall N₂O production is saved according to the new pulsed approach (production of 0.218 g_N2O l_mon⁻¹) with respect to the corresponding steady-state one (production of 0.246 g_N2O l_mon⁻¹). It is noteworthy to mention that, although small, this outcome is indeed significant, as achieving the mitigation of N₂O production alongside with a substantial reduction of NO_x is quite difficult.

The same pulsed procedure is replicated with a NO₂/NO_x feed ratio equal to 0.5, to evaluate the effect of NO₂ in the feed stream. Experimental results are reported in Fig. 8B. Also, in this case, i) no NH₃ slip and complete NO_x conversion are ensured at the end of the 10 NH₃-pulses and ii) the same peculiar N₂O outlet profile is observed, increasing and decreasing when switching on and off, respectively, the NH₃ feed. The average variation (~ 9 ppm) in the N₂O profile, however, is higher with respect the one previously observed at NO₂/NO_x = 0 conditions: indeed, a maximum and a minimum peak of ~ 20 and ~ 11 ppm, respectively, is achieved for NO₂/NO_x = 0.5. By additional integration of the N₂O (pink) profile, a value of 0.423 g_N2O l_mon⁻¹ is obtained: in line with what was described in the previous sections, this definitely confirms the promoting effect given by NO₂ on the formation of N₂O, even in a NH₃ pulsed environment (0.218 g_N2O l_mon⁻¹ observed for NO₂/NO_x = 0). Figure 8B also shows the comparison between N₂O profiles obtained from the pulsed strategy (pink profile) and the steady-state experiment (blue profile) starting from the same conditions (NO₂/NO_x = 0.5). Notably, by integration of the highlighted light blue area we can state that about 11% of N₂O production is saved implementing the NH₃-pulsed approach with respect to a typical steady-state experiment (0.475 g_N2O l_mon⁻¹).

It is worth mentioning that the validity of the proposed strategy is dependent on the selected operating conditions. As described above, this approach relies on the alternated injection of an excess of NH₃ to the reactor. As done in this work, both the time intervals (t₁, t₂) must be carefully selected to avoid NH₃ slip, ensuring complete NH₃ conversion due to its storage on the catalyst as well as to its reaction with NOx via SCR. If not, when under realistic conditions, the excess would otherwise reach the downstream ammonia slip catalyst (ASC), likely leading to the undesired formation of NO from NH₃ oxidation reactions [19]. Also, we have to mention that this strategy is likely valid as long as sufficient NH₃ storage is available, ideally up to 300 °C: it is not expected to work in driving cycles when higher temperatures are experienced, where also a different N₂O formation mechanism prevails, as also proved in Fig. 2 [6].

Protocol D has also been replicated at 200 °C for NO₂/NO_x = 0. Results are illustrated in Figure SI-14 and discussed in the following. This time, no complete NO_x conversion is achieved throughout the experiment. Indeed, the outlet NO concentration rises up to ~ 40 ppm during t₁, in the absence of gaseous NH₃, and drops to ~ 20 ppm during t₂, when NH₃ is injected, with an average conversion of ~ 94%. However, if compared to the corresponding steady-state results (yellow profile), no loss in the NO conversion is observed when applying the pulsed approach, as ~ 94% of deNO_x activity is also achieved under steady-state conditions. Notably, the N₂O production rises up to ~ 12 ppm during t₁, in the absence of gaseous NH₃, and drops to ~ 9 ppm during t₂, with a total production of ~ 0.305 g_N2O l_mon⁻¹. As a consequence, unfortunately, no N₂O saving is achieved during the pulsed approach at 200 °C, its production being slightly higher than the one observed at steady state (~ 0.300 g_N2O l_mon⁻¹).

To better understand, the outlet N₂O profiles obtained at 250 °C (pink profile) and 200 °C (green profile), respectively, are comparatively discussed in the following. By looking at Fig. 9, and by integration of the experimental curves, we observe a higher N₂O production at 200 °C (0.305 g_N2O l_mon⁻¹) with respect to 250 °C (0.218 g_N2O l_mon⁻¹): this is in line with what also observed in the previous sections (also from steady state runs). Interestingly, the N₂O profile shows an opposite trend when the NH₃ pulsed strategy is applied at T = 200 °C. Indeed, its concentration decreases during t₂ (with NH_3in = 1000 ppm), while it increases during t₁ (with NH_3in = 0 ppm), in contrast to the results recorded at 250 °C. According to some literature findings [11, 17], we hypothesize that the observed behavior may be linked to two different controlling phenomena prevailing at the two investigated temperatures: i) at 200 °C, the redox mechanism in which Cu sites are involved (Cu^II ↔ Cu^I) may play a governing role and ii) at 250 °C, NH₃ coverage may become the controlling parameter. Additionally, it is worth mentioning that, on increasing the temperature, a decrease of the Lewis/Bronsted NH₃ ratio is expected [27]; on the other hand, the data in Fig. 9 point out a decrease in the N₂O production when increasing the reaction temperature from 200 to 250 °C (in line with results from steady-state experiments). This may suggest an active role of the Lewis NH₃ in the formation of N₂O, an aspect that will be further investigated in future work.

Fig. 9

N₂O profiles from NH₃ pulsed strategy on Cu-CHA—effect of temperature. GHSV = 40 kh⁻¹. Feed: NO = 500 ppm, NO₂ = 0 ppm, O₂ = H₂O = 8% v/v in N₂. NH₃ = 0 ppm (α₁ = 0) during t₁ = 70 s; NH₃ = 1000 ppm (α₂ = 2) during t₂ = 70 s. T = 200 °C (green) vs. T = 250 °C (pink)

It is commonly accepted that the formation and decomposition reactions of ammonium nitrate leads to the production of N₂O during low-temperature SCR processes [6, 21, 41]. NH₃ being one of the key reactants for reaction (10), this is well in line with the outcome reported in Fig. 9 at 250 °C: when feeding NH₃ during t₂ (NH_3in = 1000 ppm), more NH₄NO₃ should be formed, thus yielding a larger amount of N₂O. This is also in agreement with what was observed in Fig. 6B when evaluating the effect of α at steady state at 250 °C. On the other hand, as observed by Xi et al. [17] and Janssens et al. [11], at 200 °C, the redox state of Cu sites may control the formation of N₂O under Standard SCR conditions (NO₂/NO_x = 0) [34]. In this respect, Xi et al. [17] observed that, at 200 °C, more oxidized Cu-CHA catalysts (i.e., with a higher fraction of Cu sites as Cu^II) produce higher amounts of N₂O, thus suggesting the possibility of N₂O formation to originate, as a by-product, from the reduction of Cu^II sites (RHC: Cu^II → Cu^I) occurring during the standard SCR redox mechanism. To validate this hypothesis, starting from preoxidized conditions, we performed a NO + NH₃ titration experiment at 200 °C over the investigated Cu-CHA catalyst. As described in Section SI.1, the NO + NH₃ mixture is able to readily reduce all the Cu^II sites according to a Cu:NO = 1:1 stoichiometry, in line with refs. [25, 26, 53]. Actually, together with the consumption of NO, a small peak of N₂O is also detected, in line with recent literature findings [11, 17, 53]. That said, during the low-temperature pulsed protocol, the presence of excess ammonia (during t₂, with α = 2) in the reacting environment should keep Cu sites mainly in the reduced form (Cu^I), hence making less Cu sites available for further reduction and corresponding N₂O formation: this may explain why at 200 °C, the N₂O made decreased when feeding 1000 ppm of NH₃.

4 Conclusions

Steady-state NH₃-SCR experiments are performed over a commercial Cu-CHA catalyst in a 150–500 °C T-range. Despite the high deNO_x activity, N₂O formation is observed, showing a bimodal trend indicative of the existence of two different reaction mechanisms, depending on the reaction temperature [6, 7]. In this work, we focus on the low-to-medium T-range (up to 300 °C).

Stationary runs reveal N₂O to increase with the NO₂ feed content (NO₂/NO_x = 0–1). However, at low temperature (T ≤ 200 °C), too high NO₂ inlet concentrations lead to mass transfer limitations associated with clogging of CHA pores due to NH₄NO₃ formation and deposition. As a result, lower N₂O formation and NO_x conversion are observed. Additional runs performed at different NH₃/NOx feed ratios (1–2) suggest the N₂O production to increase with enhancing the NH₃ feed content at T > 200 °C. Unexpectedly, the opposite trend prevails at lower temperature. In addition to gaseous ammonia, the effect of prestored NH₃ is also evaluated on N₂O formation by means of dynamic runs. In particular, faster N₂O transients are observed in the presence of a NH₃-preloaded catalyst, thus suggesting a positive effect of the NH₃ storage on the N₂O make, in line with the literature [19].

Based on these findings, an innovative approach involving the strategic dynamic injection of NH₃ is applied with the aim of minimizing the undesired release of N₂O while ensuring high deNO_x activity. Specifically, starting from an ammonia preloaded catalyst, ten consecutive NH₃ pulses are performed in a constant NO_x + O₂ + H₂O environment, repeatedly varying α from 0 to 2. At 250 °C, complete NO_x conversion is achieved throughout the experiment, together with a significant N₂O saving with respect to the corresponding stationary reference conditions. Notably, the N₂O concentration increases when NH₃ is pulsed (α₂ = 2) and decreases when NH₃ is cut-off (α₁ = 0). In line with steady-state runs, the opposite trend is instead detected at 200 °C. Notably, the present results are observed not only in the presence of NO₂ in the feed, but also under Standard SCR conditions, where, according to the literature [11, 17], and as confirmed in this work, the N₂O formation can be in charge of Cu redox processes (Cu^II → Cu^I) rather than NH₄NO₃ decomposition. Therefore, regarding the N₂O formation, we suggest the existence of two different controlling phenomena depending on the T-range, namely, i) the Cu/redox reactions at T ≤ 200 °C and ii) the NH₃ coverage at higher temperatures, ideally up to 300 °C. As the ammonia coverage is known to decrease with increasing temperature, a different mechanism should be instead invoked to explain the growing N₂O make at temperatures exceeding 300 °C.

The present results offer valuable insights into the N₂O formation pathways over Cu-CHA during NH₃-SCR. They pave the way for future studies focused on elucidating the correlation between N₂O formation and Cu/SCR redox processes, which is indeed crucial in order to comprehend how to mitigate its undesired formation.