Introduction

NOx removal from exhaust gases for a long time has been one of the major challenges in the area of the environmental catalysis [1–3]. The efforts aimed in reducing NOx concentration are mainly based on NOx reduction with various reducing agents (NH3, unburned or partially burned hydrocarbons and CO) in the exhaust or on direct decomposition. Nitrous oxide (N2O) is an important component of the earth’s atmosphere that has gained significant attention recently due to its environmental effects [4]. It is a potent greenhouse gas and is involved in the depletion of stratospheric ozone [5–7]. On the other hand, nitrous oxide is a non-selective product of NO reduction, mostly formed at low temperatures. The decomposition of N2O is thermodynamically favored, but it does not occur spontaneously for kinetic reasons. The alternative is to carry out a catalytic process in which the temperature is considerably lowered. In this sense, the heterogeneous catalytic decomposition seems to be the promising option for the N2O removal [5].

Various catalysts have been extensively studied for NOx elimination containing noble metals, ion-exchanged zeolites, and transition metal oxides [8–13]. Copper-containing catalysts are among the most active catalysts for these reactions [14–16]. However, the disadvantages of the metal oxides and mixed metal oxides as catalysts are due above all to their low surface areas that require their loading on a proper support. Recently, composite catalysts prepared by one or more active components supported on a support are of interest because of the possible interaction between the catalytic components and the support materials that could ensure much improved properties (catalytic activity, selectivity, stability, prolonged lifetime). Nanomaterials attracted interest in the last years due to their new properties and to the ability to their controllable design [17, 18]. Carbon and silica are well known as catalyst supports for reactions of NOx elimination [19, 20]. In this regard, the application of silica–carbon nanocomposites as support has drawn more attention because of the combination of hydrophilic and hydrophobic components that could contribute in different paths in the catalytic reaction. However, these composites depending on their pore architecture could be with ordered and disordered structure. Disordered mesoporous materials have wide pore distribution (sometimes with higher contribution of micropores) and these with ordered structure are with high surface areas and tunable pore morphology that facilitates the diffusion of the reagents and the reaction products inside the pores.

The purpose of the work is to investigate how the physico-chemical properties of copper catalysts supported on two types silica–carbon composites influence the activity in NO reduction with CO and in N2O elimination. We studied the effect of the support structure on the surface and the activity of copper catalysts deposited on ordered and on disordered mesoporous silica–carbon composites.

Experimental

Materials

Two series of silica–carbon materials were synthesized—with disordered and ordered porous structure with two Si/C compositions. The first series of disordered silica–carbon composites were prepared by sol–gel method with acidic catalyst. As silica precursors, TEOS and Na2Si2O3 were used. Finely ground carbon materials were mixed with Na2Si2O3 and the mixture was added to TEOS and water. The carbon component for the composites came from active carbon with S = 666 m2/g and Vt = 0.41 cm3. The second series of ordered mesoporous silica–carbon composites were prepared by tri-constituent co-assembly [21] of previously prepared resol (phenol–formaldehyde clay), silica oligomers from acid-catalyzed hydrolysis of TEOS, and triblock copolymer Pluronic F127 via the EISA method in ethanol solution [22]. The copper active phase was loaded on the two types of supports with different Si/C ratio by the incipient wetness technique, fourfold with a definite (concerning the total pore volume of the sample) volume of the solution of CuCO3Cu(OH)2 in ammonia, subsequent vacuum evaporation and treatment at 300 °C. The catalysts were denoted as Cu/DSC1 (copper on disordered silica–carbon with 85 wt% SiO2 and 15 wt% C), Cu/DSC2 (copper on disordered silica–carbon with 30 wt% SiO2 and 70 wt% C), Cu/OSC1 (copper on ordered silica–carbon with 85 wt% SiO2 and 15 wt% C), Cu/OSC2 (copper on ordered silica–carbon with 30 wt% SiO2 and 70 wt% C).

Characterization

The texture characteristics were determined by low-temperature (77.4 K) nitrogen adsorption in a Quantachrome Instruments NOVA 1200e (USA) apparatus. The nitrogen adsorption–desorption isotherms were analyzed to evaluate the following parameters: the specific surface areas (SBET) were determined on the basis of the BET equation, the total pore volume (Vt) was estimated in accordance with the Gurvich rule at a relative pressure close to 0.99. The volume of the micropores (VMI) and specific surface area connected to micropores (Smi), as well as the external specific surface area (SEXT) were evaluated according to the V-t-method [23]. The volume of the mesopores (VMES) was determined as a difference between Vt and VMI. Additionally, the pore size distribution (PSD) were calculated using nonlocal density functional theory (NLDFT) method using equilibrium models with slit-shaped/cylindrical pores in carbons (for silica–carbon) [24].

Powder X-ray diffraction patterns were collected within the range from 5.3 to 80° 2θ with a constant step 0.02° 2θ on Bruker D8 Advance diffractometer with Cu Kα radiation and LynxEye detector. Phase identification was performed with the Diffracplus EVA using ICDD-PDF2 Database. Mean crystallite size was determined with the Topas-4.2 software package using the fundamental parameters peak shape description including appropriate corrections for the instrumental broadening and diffractometer geometry.

The thermal analysis was carried out for determination of SiO2 and carbon content using the computerized combined apparatus for thermal analysis LABSYSEvo, SETARAM (France) in the temperature range 25–1000 °C. The experiments were carried out in dynamic mode, with heating rates of 10 °C/min and oxidizing gas flow rates of 20 ml/min.

A Thermo SOLAAR M5 flame atomic absorption spectrometer was used for the determination of the copper content. All measurements were performed in air-acetylene flame under standard conditions.

The TEM investigations were performed on a JEOL JEM-2100 instrument at an accelerating voltage of 200 kV. The specimens were prepared by grinding and dispersing them in ethanol by ultrasonic treatment. The suspensions were dripped on standard holey carbon/copper grids.

The X-ray photoelectron spectroscopy (XPS) measurements of catalysts supported on disordered composites were done in the UHV chamber of ESCALAB-Mk II (VG Scientific) electron spectrometer with a Al Kα1,2 radiation (hν = 1486.6 eV). The composition and electronic structure of catalysts supported on ordered composites were measured on AXIS Supra electron-spectrometer (Kratos Analitycal Ltd.) using monochromatic AlKα radiation (hν = 1486.6 eV). The binding energies (BE) were determined with an accuracy of ±0.1 eV. The chemical composition is determined monitoring the areas and binding energies of C 1 s, O 1 s, Si 2p, Cu 2p photoelectron peaks. Modified Auger parameter άCu was calculated as άCu = KE (Cu L3M4,5M4,5) + BE (Cu 2p3/2), where KE (Cu L3M4,5M4,5) is the kinetic energy of the Cu L3M4,5M4,5 and BE (Cu 2p3/2) is the binding energy of the Cu 2p3/2 peak.

Catalytic tests

The catalytic experiments for NO reduction with CO were carried out in catalytic flow apparatus in an isothermal flow reactor (quartz tube, i.d. 6 mm) in the temperature range 25–300 °C. The catalytic tests were performed with a gas mixture: NO + CO + Ar, containing 1200 ppm NO and 1200 ppm CO, gas hourly space velocity (GHSV) was 26,000 h−1. The outlet concentrations of NO, CO and CO2 were controlled by infrared gas analyzers. N2O was measured spectrophotometrically and a thermal converter was used for the total analysis of NOx (NO + NO2). The concentrations of NO, NOx, N2O, CO, and CO2 were continuously measured. Based on the experiment, we can calculate the nitrogen balance, taking into account the moles necessary for N2, N2O and NO2 formation.

N2O decomposition was performed using an integrated quartz micro-reactor and mass-spectrometer system (CATLAB, Hiden Analytical, UK) in the temperature range 40–300 °C. The composites were pinned between plugs of quartz wool in a quartz tubular vertical flow reactor (i.d. 6 mm). Before the measurements, the composites are treated under Ar flow at 100 °C for 1 h and then the reactor is cooled down to ambient temperature. Then N2O decomposition tests were performed with a concentration of 500 ppm N2O in argon and gas mixture with GHSV 24,000 h−1. During the tests, the following arbitrary mass units (amu) (m/z) were simultaneously collected by the Quadrupole Mass Spectrometer: 44 (N2O), 40 (Ar), 28 (N2), 32(O2), 12 (CO2).

Results and discussion

Materials characterization

X-ray diffraction

XRD data are presented in Fig. 1. The diffraction pattern of silica is only seen at around at 22.5° with a broad peak attributed to amorphous silica. XRD patterns for carbon represent broad diffraction features due to the turbostratic structure of disordered carbon. X-ray diffractograms show no significant reflections of copper containing phases in the wide angle region for the samples Cu/DSC2 and Cu/OSC1 (Fig. 1), indicating either the structure of copper species is amorphous or their size is smaller than the XRD detection limit. At the same time, well defined reflections, typical of CuO (JCPDS 05-0661) are registered for the sample Cu/DSC1 and of Cu2O (JCPDS 05-0667) for Cu/OSC2, respectively (Fig. 1). The average diameter of the crystallites in these cases, calculated on the base of the Scherrer equation, is about 10 nm for Cu/DSC1 and 8 nm for Cu/OSC2.

Fig. 1
figure 1

X-ray diffraction patterns of investigated catalysts: a Cu/DSC1; b Cu/DSC2; c Cu/OSC1 and d Cu/OSC2

Full size image

Nitrogen adsorption

Fig. 2 presents the adsorption–desorption isotherms of the silica–carbon composites (a) and copper catalysts (b) and Fig. 3 shows the pore-size distribution (PSD) curves of the copper catalysts. The isotherms of both type of composites and their copper catalysts are of IV type according to IUPAC classification with H1 hysteresis loops testifying for their globular structure with a reasonably uniform size and packing. The isotherms of composites with ordered structure and their copper modifications are with well-defined step due to capillary condensation in ~0.6–0.8 p/p0 region, typical of mesoporous materials with cylindrical mesopores and narrow pore size distribution. For the isotherms of the copper catalysts, small changes are observed—some narrowing of the hysteresis loops could be seen, associated with increase in the density of the samples and amending the size of the pores. The preservation of the isotherms and PSD for the cooper modified ordered composites indicates absence of significant structural changes with the host matrix after the copper deposition. The observed decrease in the BET surface area and total pore volume after the modification reveals pore blocking due to deposition of copper phase in the openings of the pores and on the external surface (Fig. 2; Table 1). The catalysts deposited on disordered composites show dual behavior. On one hand the copper phase fills the openings of the admission pores and blocks part of the porous space of the samples. As a result specific surface area, total pore volume and mesopore volume decrease (Table 1), but ill resolved peaks appear in the PSD curves (Fig. 3). On the other hand, the penetration in the inter- and intra- aggregate pores leads to the formation of well-defined narrow peaks in the direction to the fine mesopores.

Fig. 2
figure 2

Adsorption-desorption isotherms at −196 °C of (a) silica–carbon composites and (b) copper catalysts. Solid lines present adsorption, dash lines- desorption branches

Full size image
Fig. 3
figure 3

Pore-size distribution curves of copper catalysts, made by NLDFT using equilibrium models with slit-shaped/cylindrical pores in carbons: a Cu/DSC1; b Cu/DSC2; c Cu/OSC1 and d Cu/OSC2

Full size image
Table 1 Composition and texture parameters of the investigated silica–carbon composites and copper catalysts
Full size table

TEM

Bright-field micrographs of the investigated copper catalysts are presented in Fig. 4. TEM data show that all catalysts are heterogeneous in their morphology. The sponge type structure can be observed for the catalysts supported on disordered silica–carbon composites Cu/DSC1 and Cu/DSC2. The arrays of hexagonal two dimensional ordered layers of pores are clearly identified on the images of Cu/OSC1 and Cu/OSC2, supporting the data from the nitrogen adsorption for the preservation of the mesoporous ordered pore structure during the copper loading. Copper nanoparticles with almost spherical shape are distinguished for all materials. The observed copper particles are within 10–20 nm, but probably they are aggregates of particles with lower sizes, as for two of the catalysts XRD do not show diffraction patterns. According to the nitrogen physisorption data, copper phase is located on the external surface and blocks the pore openings (see above), although for Cu/OSC1 we suppose the copper phase is rather in the bulk than on the external surface. Presence of more finely dispersed particles is not excluded as well, but they seem to be in lower amount and we cannot quantify them precisely.

Fig. 4
figure 4

Bright field TEM images of Cu/DSC1, Cu/DSC2, Cu/OSC1 and Cu/OSC2

Full size image

XPS

In order to obtain more information about the surface oxidation state of the loaded copper particles XPS analyses are performed (Fig. 5; Table 2). The Cu2p3/2 line shape (broad peak, existence of satellite) and the binding energy of about 933.7 eV (Fig. 5a) for Cu/DSC1 and Cu/DSC2 suggests that copper is mostly in the Cu2+ state, although the satellite is not as strong as in CuO [25]. Generally, a binding energy greater than 934.4 eV is characteristic for Cu2+/CuO, while a binding energy of ~932.7 eV is attributed to Cu+/Cu2O [26]. In the case of Cu/DSC2, a partial reduction of copper could be seen from the spectra- the satellite decreases in intensity and the Cu2p3/2 main line shifts to lower binding energies. One can see the copper on the surface of Cu/OSC2 is predominantly in Cu+ state (Fig. 5b) that is confirmed also by XRD, upper presented (Fig. 1). These considerations are supported by the modified Auger parameters of the samples that are in the range 1848.5-1849.9 eV, evidencing the mixed Cu2+-Cu+ state of copper. Table 2 presents the surface composition of the catalysts. It could be seen that irrespective of the carbon and silica content (Table 1) in all cases the carbon content on the surface is higher than in the bulk. The surface carbon content in Cu/OSC- samples is higher than that in Cu/DSC-samples. Hence, carbon in the ordered composites is situated mostly on the external surface along the pores. This enhanced carbon content on the surface leads to partial reduction of copper species (Cu+/Cu2+ ratios for Cu/OSC samples is higher). The quantified deconvolution [27] of the curves (Fig. 5 and Table 2) show different ratios Cu+/Cu2+ for the investigated samples. On other hand, on the surface of Cu/OSC1 copper is in lowest content, probably due to the open pore architecture and the respective composition of the sample that ensures low micropore contribution.

Fig. 5
figure 5

Cu2p3/2 spectra for: left part—a Cu/DSC1, b Cu/DSC2 and right part—c Cu/OSC1, d Cu/OSC2. Solid lines—experimental data, dash lines—deconvolution curves

Full size image
Table 2 Surface composition of the analyzed catalysts
Full size table

Catalytic tests

NO reduction with CO

Specific activity values for NO and CO conversion (converted mol NO (CO) per surface unit of catalyst for hour) over the investigated catalysts are calculated by the formula:

$${\text{SA}} = \frac{{{\text{V}} . {\text{C}} . {\text{M}} .\eta }}{22.4.P.S}$$

where V is the space velocity (cm3/h), C is the inlet concentration of NO (vol%), M is the molecular mass of NO (CO), η is the respective conversion degree (%), P (g) is the weight of the sample, S specific surface area (m2/g). Fig. 6 illustrates the dependence of the specific activity of NO and CO on temperature for the two series catalysts. In the present experiments, N2O along with the target product N2 is detected in the reaction system up to 100 °C over the catalysts Cu/DSC1, Cu/DSC2 and Cu/OSC2 and up to 150 °C over the catalyst Cu/OSC1. When the temperature is higher than these temperatures, no N2O is measured and the reaction proceeds to N2 only. All catalysts show activity in the investigated reaction, the catalysts Cu/DSC1, Cu/DSC2 and Cu/OSC2 being more active than Cu/OSC1. The reaction proceeds almost stoichiometrically regarding the activity towards NO and CO for all catalysts up to 150 °C (Fig. 6a, b). At temperatures above 200 °C more active are the catalysts Cu/DSC2 and Cu/OSC2, the latter reaches 100 % conversion at 300 °C. However, less CO consumption than needed for the stoichiometric reaction with NO is seen above 200 °C for Cu/DSC2 and Cu/OSC2. At these temperatures, probably the carbon from the support could assist the reduction reaction, according to the mechanism proposed by Teng and Suuberg [28, 29]. This mechanism presents two routes for the catalytic reaction: a reversible hemi-sorption, accompanied by formation of nitric-oxide surface complexes via the reaction: C + NO → C(NO), or adsorption with formation of oxygen complexes C(O) and C(O2) and nitrogen release. The formation of C(O) and C(O2) is associated with intermediate formation of NO-dimers in the finest pores of the carbon material [30]. Hence, at elevated temperatures the role of CO in our case, we suppose that it is limited to the regeneration of active centers on carbon fragments, without taking direct part in the reaction.

Fig. 6
figure 6

Temperature dependence of NO (a) and CO (b) specific activity (square-Cu/DSC1; circle-Cu/DSC2, up-triangle- Cu/OSC1 and down-triangle- Cu/OSC2). Gas mixture: NO + CO + Ar; 1200 ppm NO and 1200 ppm CO; GHSV = 26,000 h−1

Full size image

TPD experiments show that all samples have desorption peaks for NO and no desorption of CO. Fig. 7 presents the TPD spectra of NO. The most intense is NO desorption for the catalysts with the highest activity—Cu/DSC2 and Cu/OSC2 (Fig. 6), the least intense desorption is for Cu/OSC1. NO desorption curves for catalysts Cu/DSC2 and Cu/OSC2 consist of two desorption peaks that represent two forms of NO adsorption probably. One could see that with the sample Cu/OSC2, the temperatures of the peaks are 100 and 130 °C, with Cu/DSC2, the temperatures of the peaks are 60 and 100 °C. These temperatures correspond with the activity of the catalysts (Fig. 6) where Cu/DSC2 is more active then Cu/OSC2 at low temperatures, the latter more active at higher temperatures. Cu/DSC1 has broad NO desorption in wide temperature range. It is important to notice that the desorbed NO quantities correlate on one hand with the order of activity towards NO of the samples and on the other hand with the surface copper content of the samples (Figs. 6 and 7; Table 2). However, no desorption peaks are registered above 150 °C (catalysts show enhanced activity) and we suppose that the activity of the catalysts above this temperature is due to the reaction of NO from the gaseous phase on the previously formed active sites on the surface of the catalysts.

Fig. 7
figure 7

TPD spectra of NO at 25–300 °C and a heating rate 13 °C min−1 in Ar: a Cu/OSC1; b Cu/DSC1; c Cu/OSC2 and d Cu/DSC2

Full size image

Due to the silanol and carbon species in the two types supported copper catalysts, Cu2+ and Cu+ are available on the surface of the composites. The quantities of copper in the two oxidation states depend on the type and the composition of the support (Table 2). Tsoncheva et al. found that during the redox reaction for the carbon supported copper catalysts, the activity depends on the catalytic active complex which is composed of at least two copper ions in various oxidation states (Cu2+, Cu+ and even Cu0) [31]. Hadjiivanov and Knözinger [32] have investigated adsorption and co-adsorption of NO and CO by FTIR on silica-supported copper catalysts. They have found that NO and CO have been selectively adsorbed on the Cu2+ and Cu+ sites, respectively. In our case, we suppose the activity of the catalysts to be determined by the surface copper content of the samples and by the ratio of Cu+/Cu2+ species on them.

N2O decomposition

The temperature dependence of the specific activity in N2O conversion for the investigated catalysts in terms of converted grams N2O from the unit surface of the catalyst for hour is shown in Fig. 8. It is evident that all catalysts are active in the investigated temperature interval. The more active are catalysts, supported on ordered mesoporous silica–carbon composites, irrespective to their silica–carbon composition, however big differences in the specific activity values are not observed. Fig. 9 shows the correlation between the activity of the catalysts on carbon, oxygen and on Cu+/Cu2+ ratio (from Table 2). It is evident that the activity increases with increasing the carbon content on the surface of the catalysts (decreases with increasing the oxygen content, respectively) and with increasing the Cu+/Cu2+ ratio. The most active catalyst in this reaction is Cu/OSC2, which carbon content on the surface is about 70 at. % and Cu+/Cu2+ ratio is with largest value.

Fig. 8
figure 8

Temperature dependence of N2O specific activity for the investigated catalysts (square-Cu/DSC1; circle-Cu/DSC2, up-triangle- Cu/OSC1 and down-triangle- Cu/OSC2). Gas mixture: 500 ppm N2O in Ar; GHSV = 24,000 h−1

Full size image
Fig. 9
figure 9

Correlations between N2O specific activity and surface carbon (square), oxygen (up triangle) and Cu+/Cu2+ (empty circle)

Full size image

Several studies have examined the mechanism of the N2O decomposition [33–35]. Generally, the authors agree that N2O decomposes on the active metallic sites to N2 and O-adsorbed species. N2O is efficiently removed over the active sites on the supported carbon materials [35]. It was confirmed that the carbonaceous materials play a double role in the tested reaction, both as an adsorbent and reactant, due to the formation of the surface oxygen complexes C(O). The latter is consequently oxidized by N2O and released as CO2. Indeed, in our case traces of CO2 are registered in the effluent gas (but not oxygen) along with N2, evidencing the participation of the carbon from the silica–carbon composites in the reaction. Hence, during the N2O decomposition on these catalysts, carbon becomes not only a support but also a reactant. Correct material balance cannot be done, as the intensity of the measured CO2 mass (12) is low.

According to [35], a reduced catalyst surface is required for N2O conversion and the role of the catalyst is to facilitate the oxygen transfer from N2O to carbon. On the basis of our results, the possible mechanism is suggested as follows:

$${\text{N}}_{ 2} {\text{O}} + {{\text{ S}}^{*}} \to {{\text{S}}^{*}}\left( {\text{O}} \right) \, + {\text{ N}}_{ 2}$$
$${{\text{S}}^{*}}\left( {\text{O}} \right) \, + {\text{ C}} \to {\text{ C}}\left( {\text{O}} \right) \, + {{\text{S}}^{*}}$$
$${\text{C}}\left( {\text{O}} \right) + {{\text{S}}^{*}}\left( {\text{O}} \right) \to {\text{CO}}_{ 2} + {{\text{S}}^{*}}$$

Here S* denotes active catalyst complex (Cu+-Cu2+). N2O seems to promote the formation of surface C(O) complexes, more or less stable. The decomposition of these complexes originates CO2 and free sites [36].

Conclusion

It is found that the catalytic behavior of the silica–carbon supported copper catalysts is determined by the nature of the support and the surface copper species. The specific activity of the catalysts in NO reduction with CO is influenced by the surface copper content. In the decomposition of N2O, differences between the two series copper catalysts supported on ordered mesoporous and disordered silica–carbon composites are found. The specific activity does not depend at all on the nominal copper content of the samples but on the ratio Cu+/Cu2+ and on the surface carbon, which is more with the catalysts supported on ordered mesoporous composites, situated mostly on the external surface along the pores. In this case carbon plays a role not only of a support but also of a reactant.