1 Introduction

Over the last few decades, stricter regulations on nitrogen oxide (NOx) emissions have been implemented, to mitigate its impact on human health and the environment [1,2,3,4]. There are two technologies that have emerged as solutions for NOx reduction, NOx storage and reduction (NSR) catalysis and selective catalytic reduction (SCR) catalysis, as well as combining the two [5,6,7]. For SCR, the ammonia reactant is derived from hydrolysis of aqueous urea [8, 9]. Thus an extra tank of aqueous urea is needed and solutions that avoid this extra feature are desired. NSR is challenged by cost and sulfur poisoning, but avoids the need for an external reductant. The focus of this study is the NSR catalyst.

NSR technology works via a cyclic process, alternating between an oxidizing phase, called the lean phase, where the NOx is stored as nitrites or nitrates, and a rich phase with a higher content of reducing compounds relative to oxygen, where the adsorbed NOx species are reduced with ideally the formation of water and nitrogen [10, 11]. Model catalysts consist of a support with high surface area, such as alumina, a noble metal for the redox aspects and a storage element, which usually is an alkaline or alkaline-earth, such as barium [12, 13]. In addition, there has been research focused on the modification of the model catalyst by the incorporation of other metals, e.g. potassium to improve the soot resistance, Sn to improve sulfur resistance or ceria to increase the NOx adsorption and modify ammonia production [14,15,16]; or even the using other less expensive materials, such as perovskites [17]. Most of them have compared the catalysts in terms of oxidation and adsorption capacities but not so many papers have been focused on the reduction step of the cycle. Nitrogen selectivity is required, however, in certain conditions or depending on the materials, NH3 and/or N2O are observed, both undesired products that should not be emitted [18,19,20,21].

There have been several studies focused on understanding the role of different reductants that are present in the exhaust [22,23,24]. Some authors have compared the use of H2 and C3H6 and have observed a HC-intermediate NOx reduction pathway in addition to the well-stablished classical H2-driven NSR pathway, but with a weaker contribution. The better low temperature performance with hydrogen is due to the Pt site poisoning by CO or hydrocarbons. However, at higher temperatures, regeneration performance becomes comparable between the three, and for hydrocarbons, steam reforming may contribute [25].

Ammonia can form via NOx reduction during the regeneration phase, and this is the basis for the coupling of the NSR–SCR technologies [26,27,28]. It is noteworthy that parallel undesired reactions can also occur. The formation of N2O has been analyzed under different operation conditions, shown to be dependent on cycle frequency and temperature, when H2 or CO are used as the reductant [29]. In general, N2O production during the regeneration phase is not as significant when the reductants are highly active, i.e. at higher temperature for example, leading to high N2 selectivity [30]. For example, H2 leads to more N2O at lower temperatures, 200 °C, but propylene at 300 °C, where its activation begins [31]. In addition, the use of C3H6 as a reducing agent involves a reaction network that includes NCO species, that leads to N2O during the lean phase, but this is reduced in the presence of water [32]. When CO or C3H6 are fed, N2O production is higher relative to H2 being used, due to the formation of isocyanates and CO adsorbed on Pt sites [33].

Although there has been recent research, with some highlights mentioned above, focused on N2O formation, the relationship between the ammonia formed and reduction of the nitrate or nitrite species, or reaction with exhaust NO, leading to N2O formation has not been resolved. Therefore, in this paper we report a study focused on N2O formation during the regeneration step over a model Pt–Ba/Al2O3 catalyst using by-product NH3 as the reductant. Moreover, the study quantifies ammonia consumption, taking into account the presence of NO, and the reacted ammonia selectivity toward N2 as well as toward undesired N2O, under different stoichiometric conditions.

2 Materials and Methods

2.1 Catalyst Preparation and Characterization

The Pt–Ba/γ-Al2O3 (1/20/100, w/w) catalyst was prepared by a two-step process. First, the Pt/γ-Al2O3 sample was made with a first impregnation of γ-Al2O3 (Puralox, Sasol, surface area 200 m3/g, pore volume 0.7 m3/g) using a solution of Pt(NH3)2(NO2)2 (diamminedinitroplatinum (II) (Pt(NH3)2(NO2)2), Pt content 3.4 wt%, density 1.05 g ml−1, Aldrich Chemical) followed by drying at 373 K and calcination at 773 K for 5 h. Ba was added to this Pt/Al2O3 via incipient wetness using an aqueous solution of Ba(CH3COO)2 (Merck, 99%). The final calcination was carried out at 773 K, in air for 5 h. More detailed information about the catalyst preparation and properties were given previously [20, 34].

2.2 Activity Study

The reaction studies were performed in a quartz fixed-bed reactor connected to a QMS 200 mass spectrometer (Pfeiffer Vacuum Prisma™) for outlet gas analysis and the sample temperature was measured by a thermocouple placed directly into the catalyst bed. 60 mg (100–120 µm) of the catalyst sample was used in each test with a constant total flow of 100 ml min−1. The mass spectrometer (MS) was well-calibrated, eliminating the cross sensitivity effect of the species presented in the gas mixture, to allow NO, N2O and NO2 measurements. The following mass-to-charge ratios were used to monitor the evolution of reactant/product concentration profiles during reaction: 2 (H2), 15 (NH3), 18 (H2O), 28 (N2), 30 (NOx = NO + NO2), 32 (O2), 40 (Ar), 44 (N2O), 46 (NO2). For additional information, supplementary mass-to-charge ratios were taken into account: 12 (secondary ionization of CO2), 14 (secondary ionization of N2), 16 (100% signal of NH3). The interferences of the mass-to-charge ratios were considered during data analysis and MS signal to concentration (ppm) calculations. The MS data were quantitatively analyzed using factors determined experimentally during the calibration process. For calibration, calibration mixtures were used and fragmentation patterns were recorded.

Prior to an experiment, a temperature programmed desorption (TPD) in He was done to provide a clean catalyst surface. The TPD conditions were 300–773 K, 10 K min−1. Temperature-programmed surface reactions (TPSR) were performed in feed streams containing NO (NO-TPSR), NH3 (NH3-TPSR) and NO + NH3 (NO + NH3-TPSR), with He as the balance gas.

For NO-TPSR, 500 ppm of NO was used for catalyst exposure to saturation and the 500 ppm NO concentration was maintained during the TPD, which was from room temperature to 773 K with a ramp of 10 K min−1. Similarly, for the NH3-TPSR experiments, the sample was exposed to 333 ppm of NH3 to saturation and then followed by the TPD with NH3 still present in the gas phase.

The TPSR experiments were carried out over a catalyst without preformed nitrites/nitrates (experiment type I) and for a catalyst with preformed nitrites/nitrates on the surface (experiment type II). For both types NO + NH3-TPSR type I and II, three ammonia concentrations were used, 166, 333 and 1000 ppm NH3, while keeping the NO at 500 ppm. For all experiments, first NO was admitted to the reactor at RT and after NO saturation, ammonia was added to the flow until ammonia saturation. The sample was then heated up to 723 K with a ramp of 10 K min−1. This admittedly leads to formation of surface NOx species, specifically nitrites as will be discussed below. However, the amounts are significantly less than those found in the type II experiments.

In the case of TPSR with preformed nitrites/nitrates, type II, prior to TPSR nitrites/nitrates were formed over the catalyst samples. The nitrites/nitrates storage were performed during lean-rich cycling applying transient response method (TRM) at 623 K (GHSV = 1.5 × 105 h−1, at 1 atm and 293 K). A rectangular pulse of NO (1000 ppm) + O2 (3%) in He flow and H2 (2000 ppm) in He flow were subsequently fed during the oxidation and reduction steps, respectively. There was no He purge applied between the lean and rich stepswas admitted to the reactor. After 4–5 adsorption–reduction cycles at given temperature, a stable catalyst behavior was obtained and nitrites–nitrates storage has been performed. The amounts of stored nitrites/nitrates were estimated taking into account that the amount of stored nitrites/nitrates is proportional to the area A1 in Fig. S1.

After the storage step, the catalyst was cooled in He to RT and the NO + NH3-TPSR was performed. To ensure the complete removal of the stored nitrites/nitrates, a regeneration phase included exposure to 1000 ppm of NH3 in He at 623 K at the end of TPSR type II, followed by the TPD up to 723K. The calculated selectivities to N2\(({S_{{N_2}}})\) and to N2O \(({S_{{N_2}O}})\) are based on molar amounts of N2, N2O, NO and NO2 produced. The calculated Nbalance closed with an error less than 57%.

Diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) studies were performed in a continuous flow Harrick Praying Mantis reaction chamber. A FTIR Nicolet Nexus 470 spectrometer was used to characterize the interactions between the gas species and the catalyst surface during the TPSR experiments. The reactions were carried out in a temperature range between room temperature and 723 K. Spectra were recorded in diffuse reflectance mode, and 45 scans were collected at a resolution of 1 cm−1 (Table 1).

Table 1 Assignments of peak positions in DRIFT spectra results
Full size table

3 Results and Discussion

3.1 NH3 versus H2 Reduction of Nitrites/Nitrates

Figure 1 shows the reduction of nitrites/nitrates stored over Pt–Ba/Al2O3 using H2 (Fig. 1a) and NH3 (Fig. 1b, c). Upon H2 admission to the reactor at t = 0 s, H2 was completely consumed for about 180 s and then, the H2 outlet concentration gradually increased with time, reaching the inlet value at ca. 1400s. N2 was formed and after 350 s, NH3 was observed, reaching ca. 100 ppm. H2O was also observed as a product, reaching 1300 ppm after 200 s. The amounts of N2 and H2O produced is close to the 1:5 ratio, indicating nitrites/nitrates reduction through reactions (1) and (2). In line with literature no N2O was observed under these conditions [18, 20, 49, 50].

Fig. 1
figure 1

TRM profile for the Pt–Ba/AlPox catalyst during rich phase a 2000 ppm of H2 in He, b 1500 ppm, c 2500 ppm of NH3 in He at 623 K, d H2 vs NH3 in He comparison. Total flow 100 ml min−1

Full size image
$$3Ba{(N{O_3})_2}+5{H_2} \to {N_2}+5{H_2}O+BaO$$
(1)
$$3Ba{(N{O_3})_2}+8{H_2} \to 2N{H_3}+5{H_2}O+BaO$$
(2)

The results obtained during the reduction of stored NOx with different amounts of NH3, 1500 versus 2500 ppm, are presented in Fig. 1b, c. In the experiment with 1500 ppm NH3, Fig. 1c, full reductant consumption and N2 production occurred for ca. 85 and 250 s, respectively. Increasing the NH3 concentration led to faster regeneration, Fig. 1b. The NH3 consumption and N2 production time decrease by ca. 50% in comparison to the same experiment with less NH3, although less than twice as much was added. With 2000 ppm NH3, the regeneration time decreased by ca. 70% compared to H2, although on a per H basis it should have been only 50% if linearly correlated to reductant amount. These results indicate that for the given reaction conditions, at 623 K NH3 is a more efficient reducing agent than H2 (Fig. 1d).

Based on the results obtained during these specific experiments, reaction (3) is sufficient to explain the reaction stoichiometry, with the error (Nadsorbed/Nproduced) found to be in the 3–5% range.

$$3Ba{(N{O_3})_2}+10N{H_3} \to 8{N_2}+15{H_2}O+3BaO$$
(3)

The obtained data reveal that under similar conditions, the reduction of stored NOx was much faster with NH3 compared to H2. All these results also show high reactivity between NO and H2. With increasing NH3 concentration, the reaction onset temperature was not significantly changed, and with H2 as a reductant NH3 formation was detected among the reaction products, what is in line with previous study [51]. Ultimately, N2 is formed, and possible pathways for its production have been proposed, involving NO adsorption on reduced Pt particles, followed by decomposition to N and O ad-species [19, 51, 52].

O(ads) species are removed by H2, leading to water formation. Along similar lines, the presence of NH3, observed for high H2 concentrations, would be due to reaction between N and H ad-species [51]. Ba also apparently plays a role in those reactions. It has been reported that Ba addition to supported Pt catalysts results in an increase in NOx reduction activity in the absence of O2 [19, 52, 53]. It was suggested that according to reactions (4)–(6), the presence of the alkaline or alkaline-earth metal ion weakens the N–O bond [54]. This will favor dissociation and subsequent reduction. The addition of Ba also increases the selectivity of the reaction to N2.

In the case of ammonia as a reductant, previous studies have shown that it behaves similar to H2, with the same NOx conversion efficiency [55, 56]. On the other hand, some literature data have revealed a difference in NH3 versus H2 reactivity towards stored NOx, for different catalyst compositions [50]. Specifically, a higher reactivity of ammonia for Pt–K/Al2O3 regeneration was noted relative to Pt–Ba/Al2O3, possibly due to the different nature of stored nitrites/nitrates (different ionic/bidentate ratio) and the Pt sites. In the results shown in Fig. 1, the reductant choice influenced the reduction rate. As the same catalyst was used for both experiments, it is not a difference in initial surface species type. In order to better understand the NH3 reactivity with NSR catalyst surface species, TPSR experiments were carried out over the nitrite/nitrate-free (or at least minimized) Pt–Ba/Al2O3 and for the same catalyst with nitrites/nitrates on the surface.

3.2 TPSR over a Catalyst Without Preformed Nitrites/Nitrates

TPSR experiments over the sample without preformed nitrates/nitrites can help isolate the reaction between gas phase NO and NH3. First, an NO-TPSR was examined (Fig. 2), and reveals N2O formation with a maximum at 657 K, likely through NO decomposition on Pt. The operando DRIFTS study (Fig. S2) showed mostly nitrite formation on the surface, with two main peaks located at 1300 and 1230 cm−1.

Fig. 2
figure 2

NO-TPSR profile for the Pt–Ba/AlPox catalyst 500 ppm of NO in He. Total flow 100 ml min−1

Full size image

The NO + NH3-TPSR results from the sample without preformed nitrates/nitrites are shown in Fig. 3, with corresponding in situ DRIFTS catalyst surface analysis shown in Fig. S3. The DRIFTS results show mainly nitrite formation on Ba sites, similar as in case with NO- and NH3-TPSR (peaks at 1300 and 1230 cm−1, Fig. S3 lines NN350 and NN He vs Fig. S2a and b). At the beginning of the (NO + NH3)-TPSR reaction, all experiments show an ammonia desorption peak as the temperature increased (Fig. 3). With 500 ppm NO + 166 ppm NH3 (Fig. 3), the maximum NO consumption occurs between 423 and 473 K and more product N2O, relative to N2, was detected. The results for TPSR experiments with higher NH3 concentration (333 and 1000 ppm, Fig. 3b, c) showed, that for these conditions, NO reduction by NH3 over Pt–Ba/Al2O3 produces only N2 (1.43 and 1.48 × 10−3 mol/gcat) above 423 K. In the 323–423 K temperature range, N2O is formed, however the overall amounts of N2O produced are below ca. 5 × 10−5 mol/gcat. The amounts of N2O and N2 detected most likely correspond to the following global reactions:

Fig. 3
figure 3

Reactivity of 500 ppm NO and a 1000, b 333 and c 166 ppm of NH3 over the Pt–Ba/AlPox catalyst without preformed nitrites/nitrates—TPSR type I. Flow 100 ml min−1, He as a carrier. T ramp 10 K min−1

Full size image
$$3NO+2N{H_3} \to {\raise0.7ex\hbox{$5$} \!\mathord{\left/ {\vphantom {5 2}}\right.\kern-0pt}\!\lower0.7ex\hbox{$2$}}{N_2}+3{H_2}O$$
(4)
$$4NO+N{H_3} \to {\raise0.7ex\hbox{$5$} \!\mathord{\left/ {\vphantom {5 2}}\right.\kern-0pt}\!\lower0.7ex\hbox{$2$}}{N_2}O+{\raise0.7ex\hbox{$3$} \!\mathord{\left/ {\vphantom {3 2}}\right.\kern-0pt}\!\lower0.7ex\hbox{$2$}}{H_2}O$$
(5)

As was shown in Fig. 3, N2O via reaction (5) dominates at low ammonia concentration (Fig. 3a) and conversion (up to 438 K Fig. 3a–c). The higher NO consumption at 423–473 K, compared to that at temperatures above 473 K, is probably due to the stoichiometry of reactions (4) and (5). However, the relative significance of the two reactions changes with temperature in such a way that at low temperatures reaction (5) dominates and at higher temperatures, its importance decreases. Also at higher temperatures, above ca. 573–623 K, N2O can decompose [57]. For reaction (4), the opposite trends occur. It participates in the TPSR process more significantly above 473 K in all cases.

The mechanism N2O formation describes the recombination of N-species on reduced Pt particles. Previous studies showed that the recombination of N ad-species gives N2 as a product and recombination of N and NO ad-species leads to N2O formation according to reactions (6)–(8) [54, 58, 59].

$$N{O_{(g)}} \to N{O_{(ads)}} \to {N_{(ads)}}+{O_{(ads)}}$$
(6)
$${N_{(ads)}}+{N_{(ads)}} \to {N_2}_{{(g)}}$$
(7)
$$N{O_{(ads)}}+{N_{(ads)}} \to {N_2}{O_{(g)}}$$
(8)

Moreover it was shown that the addition of Ba, especially its presence in close proximity to Pt, also increases the selectivity of the reaction to N2. This enhanced selectivity was associated with the increase in the NO dissociation rate, which decreases the formation of N2O according to reaction (8) [54].

In Fig. 4, the calculated amount of ammonia consumption towards N2 and N2O formation is shown for different NH3 concentrations (166, 333 and 1000 ppm). For low ammonia concentration, reaction (5) plays an important role up to 773 K (Fig. 4, 166 ppm), while with higher ammonia concentrations (333 and 1000 ppm, Fig. 4), at temperatures above 473 K the contribution of reaction (5) to the total consumption of ammonia is negligible. For all TPSR type I experiments, the maximum consumption of ammonia via reaction (5) occurs at ca. 423–448 K (Fig. 4) and at temperatures higher than 473 K an increase in the total amount of ammonia reacted toward N2 production was observed. At the later stage of the TPSR reaction, above 473 K, the estimated NH3 consumption values are roughly consistent with ammonia concentration inlets.

Fig. 4
figure 4

Temperature dependence of reaction participation towards N2 and N2O over Pt–Ba/AlPox in the presence of: a 500 ppm NO + 1000 ppm NH3, b 500 ppm NO + 333 ppm NH3 and c 500 ppm NO + 166 ppm NH3. Total flow 100 ml min−1, He as a carrier

Full size image

According to the results above, the amounts of N2O produced depend on ammonia concentration and the reaction temperature. In Fig. 5, N2O production during (NO + NH3)-TPSR using 166, 333 and 1000 ppm of NH3 is presented.

In general, N2O is a dominant product at low temperature, when the NH3 conversion is low. The largest amount of N2O was produced during the experiment with 166 ppm of ammonia (Fig. 5a), around 3.81 × 10−4 mol/gcat, and for this ammonia concentration most likely reaction (5) dominates till ca. 548 K. For higher ammonia concentrations (Fig. 5b, c), reaction (5) occurs in a narrow temperature window, ca. 323–373 K, and the values of N2O produced decreased by ~ 90% with respect to the experiment with 166 ppm NH3. Moreover, for higher ammonia concentrations, reaction (4) starts to dominate at much lower temperatures, at ca. 423 K. This may suggest that there is a minimum temperature, which limits reaction (4). This temperature would be tied to NH3 activation on Pt sites or product desorption from the catalyst surface. Characterizing the slopes of the N2O and N2 concentration profiles, up to 323 K, leads to the assumption that the reaction rates for all experiments are similar. Moreover, the amounts of N2O and N2 produced up to 323 K indicate that the rates of reaction most likely are not controlled by ammonia concentration.

Fig. 5
figure 5

N2O production during TPSR type I experiment upon admission of 500 ppm of NO and a 166 ppm, b 333 ppm and c 1000 ppm of NH3

Full size image

In general, during NO assisted TPSR experiment type I, with higher than stoichiometric ammonia concentrations, reaction (5) is practically negligible and N2 is a dominant product through all experiments. At lower ammonia concentrations, more N2O will form as shown in Fig. 5.

3.3 TPSR over Catalyst with Preformed Nitrites/Nitrates

The TPSR type II experiments, with nitrites/nitrates preformed on the catalyst surface, were carried out in order to clarify the role of the NO-NH3 reaction during the reduction of surface nitrates/nitrites (regeneration of the NSR catalyst) and also to compare the reactivity of NH3 with the gas-phase NO entering the reactor, still of course on the catalyst surface, or the nitrite/nitrate species on the catalyst surface, or if both reactions take place together (Figs. 6, 7, 8, with corresponding in situ DRIFTs spectra shown in Fig. 9).

Fig. 6
figure 6

NO-TPSR over Pt–Ba/AlPox catalyst with nitrites/nitrates preformed. Flow 100 ml min−1, gas composition 500 ppm NO in He, Tramp 10 K min min−1

Full size image
Fig. 7
figure 7

He-TPD over Pt–Ba/AlPox catalyst with nitrites/nitrates preformed. Flow He 100 ml min−1. Tramp 10 K min min−1

Full size image
Fig. 8
figure 8

NH3-TPSR over Pt–Ba/AlPox catalyst with nitrites/nitrates preformed. Flow 100 ml min−1, gas composition 333 ppm NH3 in He, Tramp 10 K min−1

Full size image
Fig. 9
figure 9

In situ DRIFT spectra for Pt–Ba/AlPox catalyst with nitrites/nitrates preformed. Gas composition: a 333 of NH3 in He, b 500 ppm NO in He, and c He. Flow 100 ml min−1. T ramp 10 K min−1

Full size image

NO-TPSR data are shown in Fig. 6, where species evolution are plotted as a function of time along with temperature. For this reaction, little influence of NO on the decomposition of the stored nitrites/nitrates was found (NOx-species stored shown Fig. 9a–c lines NN350, NNHe). Apart from the higher NO concentration, the evolution of NO at temperatures higher than 623 K resembled those obtained for just nitrate decomposition in He (Fig. 7). However it is worth mentioning that nitrite/nitrate decomposition in the presence of He starts at ca. 50 K lower and slightly more NO2 is produced when compared to nitrite/nitrate decomposition during NO-TPD. The maximum in nitrite/nitrate decomposition occurred at ca. 723 K. In both experiments (TPD in He and TPD in NO + He), after reaction ca. 6–9% of nitrite/nitrates remain on the catalyst surface as unreacted or stable; however they can be removed by a short H2 or ammonia pulse at 723 K. The amounts of NOx detected from nitrite/nitrate decomposition are consistent with the amounts of NOx stored during the lean phase of the experiment and the mass balance error was found to be less than 8%.

During the NH3-TPSR type II experiment with the Pt–Ba/Al2O3 catalyst, complete nitrite/nitrate reduction was achieved as shown in Fig. 8. The amounts of N2O, NO and N2 produced during the experiment correspond to the amounts of NOx stored during the adsorption step. Reduction occurs from ca. 373 K up to ca. 673 K, however below 423 K the reduction of stored nitrite/nitrates is only partial. Based on these data, at this temperature the reaction probably is under kinetic control and the N-adsorbed species decomposition and/or reactivity of ammonia will limit the reaction. Above 443–473 K, reduction is fast and the reaction rate and product distribution are controlled by ammonia concentration. Moreover, for the NH3TPSR experiment, decomposition of ammonia was detected at temperatures higher than 623 K and the amount of H2 and N2 (ratio 1:3), observed at the end of the NH3-TPSR experiment, correspond to the stoichiometry of reaction (9).

$$2N{H_3} \to {N_2}+3{H_2}$$
(9)

N2O production is small and the reaction is highly selective to N2. The amount of the N2O produced through all experiments was lower than that observed over the samples without preformed nitrites/nitrates with 500 ppm of NO and 333 ppm NH3 (Fig. 8), and was below 5 × 10−5 mol/gcat.

In situ DRIFTS spectra are shown in Fig. 9 for the experiment with nitrites/nitrates preformed; NH3–TPSR (Fig. 9a), NO–TPSR (Fig. 9b) and nitrite/nitrate decomposition in He (Fig. 9c). The spectra recorded during NH3–TPSR shows gradual consumption of nitrites (bands at 1300 and 1230 cm−1) and a small amount of linear and bidentate nitrates (bands at 1480 and 1510 cm−1) with increasing temperature. The formation of any other surface species was not evident during this experiment. At 773 K only a small amount of residual nitrite species are still present, 1300 cm−1. Similar results were obtained for NO–TPSR (Fig. 9b) and nitrite/nitrate decomposition in He (Fig. 9c). For these experiments, the nitrites/nitrates decomposed also without other surface species formation. The highest amount of remaining nitrite species at 773 K occurred with the NO–TPSR experiment (Fig. 9b).

The results of the TPSR experiment with preformed nitrates/nitrites are shown in Fig. 10. For low ammonia concentration (500 ppm NO + 166 ppm NH3), the N2O profile resembled that for (NO + NH3)-TPSR over the catalyst with no preformed nitrites/nitrates (Fig. 3). At the beginning of the (NO + NH3)-TPSR reaction, all experiments show an ammonia desorption peak as the temperature increased (Fig. 10 vs 3). With 500 ppm NO + 166 ppm NH3 (Fig. 10), the maximum NO consumption occurs between 423 and 473 K with N2O and N2 formation. Nitrite/nitrate decomposition is observed between 657 and 773 K, similarly as in case of NO-TPSR with nitrites/nitrates preformed (Fig. 6). In this temperature range, the nitrites/nitrates decomposition for a stoichiometric NO:NH3 ratio, i.e. 500 ppm NO + 333 ppm NH3 is also observed. The results for TPSR experiments with higher NH3 concentration (333 and 1000 ppm) showed, that for these conditions, NO reduction by NH3 reacts to form only N2 above 423 K. In the 323–423 K temperature range, N2O is formed. For the higher ammonia concentrations, NH3 decomposition was evident. It can be assumed that reaction (5) dominates from the beginning of the reaction test up to ca. 623 K. Above this temperature the NO peak reaches its maximum at 723 K. This peak can be attributed to nitrate/nitrite decomposition as they become unstable, similar as what was observed for NO-TPSR, Fig. 6.

Fig. 10
figure 10

Reactivity of 500 ppm NO and a 1000, b 333 and c 166 ppm of NH3 over Pt–Ba/AlPox catalyst with nitrites/nitrates preformed—TPSR type II. Flow 100 ml min−1, He as a carrier. Tramp 10 K min−1

Full size image

The in situ DRIFTS spectra obtained during (NO + NH3)-TPSR with nitrites/nitrates preformed are shown in Fig. 11. The spectra resemble those obtained during NH3–TPSR (Fig. 9a). Upon NH3 admission, there is gradual consumption of nitrites (bands at 1300 and 1230 cm−1), and linear and bidentate nitrates (bands at 1480 and 1510 cm−1) with increasing temperature, without any other species formation. At 773 K, only a small amount of residual nitrite species are evident, 1300 cm−1, again similar as for NH3–TPSR (Fig. 9a). However, with increasing ammonia concentration, 333 ppm (Fig. 11b) and 1000 ppm (Fig. 11c), the surface species reduction was faster and more nitrite/nitrate surface species were reduced between 423 and 523 K. The amount of residual species were similar for all TPSR experiments, regardless of ammonia concentration.

Fig. 11
figure 11

In situ DRIFT spectra for a Pt–Ba/AlPox catalyst with nitrites/nitrates preformed—TPSR type II. Gas composition: 500 ppm NO and a 1000, b 333 and c 166 ppm of NH3 in He. Flow 100 ml min−1. T ramp 10 K min−1

Full size image

To explain the reactivity of preformed nitrates/nitrites in NO + NH3, it is necessary to take into account reactions (4) and (5) as well as the reactions between nitrates/nitrites and ammonia according to equations (10) and (11):

$$3Ba{(N{O_3})_2}+10N{H_3} \to 8{N_2}+15{H_2}O+3BaO$$
(10)
$$Ba{(N{O_2})_2}+2N{H_3} \to 2{N_2}+3{H_2}O+BaO$$
(11)

During TPSR experiment type II with low ammonia concentration (Fig. 10), it was observed that NH3 was completely consumed, and decomposition of unreacted nitrites occurred at temperatures higher than 623 K. N2 and N2O production was observed, which most likely took place according to the stoichiometry of reactions (4) and (5). At low temperatures, below 438 K, reaction (10) occurs to a very small extent, if at all, taking into account the amount of NO and NO2 from nitrate decomposition, with the decomposition peak maximum at ca. 723 K. Also, with NO present, nitrites/nitrates decomposition is shifted at higher temperature by 20 K, similar to the data shown in Fig. 6 and discussed above. There was no nitrites nor nitrates (Fig. 11, DRIFTS) decomposition up to 438 K. Therefore, presumably up to this temperature only NO and NH3, which are co-fed to the reactor, participated in reaction (4) and (5) on the Pt sites, forming N2 and N2O. At temperatures higher than 473 K, observed nitrates to nitrites reduction, and nitrite decomposition (Fig. 11, DRIFTS), suggests that previously stored N-species also participate in N2 and N2O formation.

With a stoichiometric NO to NH3 concentration (500 ppm NO + 333 ppm NH3, Fig. 10), the reduction of some nitrites/nitrates occurs at the beginning of the TPSR experiment. Nitrite/itrate reduction takes place in the 448–623 K range; however only ca.15% of all nitrites/nitrates stored are reduced. Above the temperature of nitrites/nitrate stability (623 K), again decomposition of stored nitrites/nitrates can be observed.

Total NO consumption and nitrites/nitrate reduction was obtained when using 1000 ppm of NH3 (Fig. 10). When NH3 is not limiting, reaction (5) is almost negligible, reactions (4) and (10) are fast and take place together in the 448–623 K range. Above 623 K no nitrate decomposition peak was observed, indicating that the reduction of stored nitrites/nitrates is complete (the amount of residual nitrites/nitrates is below 0.5%, if any). According to the N-balance and calculations with different ammonia concentrations (Fig. 12), in the later stage of TPSR type II for these experiment conditions, up to 448 K most likely only the co-fed NO and NH3 react together via reaction (4), and above 448 K the reaction with preformed N-species (nitrites/nitrates) participate towards N2 and N2O formation via reactions (10) and (11).

Fig. 12
figure 12

Product mol g−1 distribution as a function of ammonia concentration during TPSR type II experiment: a NH3 consumed with reaction (I) and (II), b NH3 left after reaction (I) and (II), c NO3− consumed by NH3 left via reaction (III) and d N2O produced during all experiment

Full size image

Taking into account the results presented above, it can be assumed that when the reductant is limiting, then the reaction between gas-phase NO and ammonia over the cagtalyst is faster and/or competitive, than the reduction of surface nitrites/ nitrates.

In comparing TPSR type II experiment data (Fig. 10, 166, 333 and 1000 ppm NH3), with preformed nitrites/nitrates, and type I experiment (Fig. 3) some changes in the NO2 profile can be observed, and suggest that reaction (9) occurs under the investigated conditions. It can be speculated that during (NO + NH3)-TPSR reaction with preformed nitrates/nitrites, the reaction between gas-phase NO with NH3 over the catalyst takes place at the expense of nitrate/nitrite reduction. The decomposition of ammonia via reaction (9) was not observed.

The obtained data reveal that N2O formation starts at a temperature as low as 373 K and when the ammonia concentration is low, N2O is produced during all TPSR type II experiments, Fig. 13. Comparing the N2O profiles of TPSR with and without preformed nitrites/nitrates (Fig. 13 vs 5), an increase in N2O production is observed with 333 ppm of ammonia (Fig. 13a, b). According to the stoichiometry of reactions (4), (5) and (10), the data suggest that reaction (5) is carried out at the expense of reaction (4) due to the (N-species + NO)/NH3 ratio and involvement of NH3 in partial reduction of nitrites/nitrates via reaction (10). Figure 12 shows product mol g−1 distributions as a function of ammonia concentration during all TPSR type II experiment, according to reactions (4), (5) and (10). The amount of N2O produced when NH3 is in abundance (1000 ppm of NH3), is roughly constant for both experiments with and without nitrites/nitrates present and is ca. 4 × 10−5 mol/gcat. For all TPSR experiments, a maximum in N2O production occurs at 448 K.

Fig. 13
figure 13

N2O production during TPSR type II experiment upon admission of 500 ppm of NO and a 166 ppm, b 333 ppm and c 1000 ppm of NH3

Full size image

When taking into account the set of data when 333 ppm of NH3 was added, i.e. (NO + NH3)-TPSR without preformed nitrites/nitrates, Fig. 5, (NO + NH3)-TPSR with preformed nitrites/nitrates, Fig. 10, and NH3-TPSR with preformed nitrites/nitrates, Fig. 8, it can be concluded that first the consumption of the NO from the gas phase occurs and later reduction of surface nitrites/nitrates with ammonia. As was found previously, Pt must be involved in the reaction, because the reduction of stored nitrates/nitrites by NH3 is not effective over Pt-free samples [54, 60]. Moreover, this current study shows that when ammonia is a limiting agent, nitrates/nitrites remain on the catalyst surface unreacted, and at temperatures higher than 623 K decompose. Therefore, this suggests that the reaction between NO and NH3 on the catalyst is faster/preferential than the reaction between NH3 and the nitrites/nitrates. Also, most likely weakly adsorbed N-species will react more easily with ammonia than for example bulk nitrates. It has been suggested that nitrates adsorbed on Ba sites in the vicinity of Pt are more easily reduced than N-species adsorbed on Ba sites situated farther away from Pt [61,62,63].

The calculated conversions and selectivities show that when NH3 is not limiting, the NO conversion reaches its maximum at ca. 448 K. At the same temperature, the NH3 conversion reaches 100% and when NH3 is in abundance, the extra ammonia stays unreacted. In general, for all performed experiments, N2O was the dominant product at low temperature, when NO and NH3 conversions are low. At higher temperature, with increasing NO and NH3 conversions, an increase in reaction (5) prevails and selectivity to N2 increases. Moreover, the selectivities towards N2 and N2O depend on ammonia concentration, which was also proposed previously [61,62,63]. N2 is the main product for higher than stoichometric NH3-to-NO concentrations, while for low NH3 concentrations, N2O production is significant.

4 Conclusions

N2O formation was studied over Pt–Ba/γ–Al2O3. The results show:

  • N2O is the dominant product at low temperature, and when NH3 concentrations are substoichiometric. At higher temperature, with increasing NO and NH3 conversions, N2 selectivity increases.

  • In characterizing the slopes of the N2O and N2 production rates, up to 423 K, it is apparent that the reaction rates are similar. Moreover, the amount of N2O and N2 produced up to 423 K indicate that the rates of reaction most likely are not controlled by NH3 concentration.

  • In experiments with nitrites/nitrates on the surface, roughly similar amounts of N2O are formed, indicating it is gas-phase NOX leading to N2O byproduct selectivity.

  • In comparing the results from the samples with nitrites/nitrates on the surface to those without, it is apparent that NH3 first reacts with NO from the gas phase and then nitrites/nitrates.

  • The reduction of nitrites/nitrates is complete only for high NH3/NO ratios, and when NH3 is the limiting reactant relative to gas-phase NOx, nitrites/nitrates remain on the catalyst surface unreacted, and at temperatures higher than 623 K decompose giving NO2 and O2.