1 Introduction

The fact that bentonites, both natural and modified, are widespread, inexpensive; available and environment-friendly sorbents permits to consider them as “materials of the twenty-first century” [1]. This denomination is confirmed by numerous studies concerning the synthesis of novel modified materials, particularly, pillared clays (PILCs) prepared by the polyhydroxy cation intercallation into montmorillonite and investigated from seventies of the last century [2]. Intercallation processes are very complicated and dependent on very many factors. Therefore, the repeatability of results and their comparison are hardly possible [3]. The properties of intercallation products depend on the nature of polyhydroxy cations (Al, Ti, Fe, Cr, Zn or Cu), synthesis conditions (concentration of metal polycations, hydrolysis, reaction time, washing, drying, and calcination), and, mainly, on origins of bentonites used [3, 4].

Pillared clays are prevalent as acid catalysts in organic synthesis and sorbents for both heavy metals and toxic gaseous substances [5,6,7,8]. Incorporation of transition metal ions into pillared clays extends the range of their application. These modified pillared clays are used as catalysts for oxidation of numerous organic compounds contained both in air and in aqueous solutions [2, 8,9,10,11,12] as well as for reduction of NO [7, 13, 14] and nitrate ions contained in water [15]. For the latter process, a Pd0–Cu0/Al-PILC catalyst was used. M/Al-PILC catalytic compositions (M is Cr, Cu or Ag) proposed for carbon monoxide oxidation [16, 17] are active only at high temperatures: t > 200 °C for Cu and Ag and t > 400 °C for Cr.

Pillared clays were not studied as supports for metal complex compounds which could be active at ambient temperatures in the reactions of deactivation of some toxic gaseous substances, e.g. CO, SO2, and O3 [18,19,20]. As can be seen [18,19,20,21,22,23,24], anchored metal complex catalysts for CO oxidation necessarily contain palladium(II) and copper(II) compounds. Mainly, their catalytic activity can be improved by varying the support nature, the nature and concentrations of palladium(II) and copper(II) precursors, introduction of ligands, L, and by changing Pd(II):Cu(II):L ratio. As for halide ligands (L = Cl−, Br−, and I−), palladium–copper complexes with bromide ligands demonstrated the maximum catalytic activity in the reaction of CO oxidation with air oxygen [18,19,20]. Our previous investigations [19, 20, 24] showed that chemical and phase compositions, structural, structural-adsorption, and physicochemical properties of natural minerals and their modified forms substantially influence the catalytic activity of palladium–copper complexes in the above-named reaction. For changing these properties of supports, different modification methods were used.

The aim of the current work was to synthesize nanosized pillared clay from one of Ukrainian bentonites and to evaluate the effect of this PILC on the activity of palladium–copper complexes supported on PILC in the reaction of carbon monoxide oxidation with air oxygen.

2 Experimental

Al-pillared clay was synthesized from natural bentonite (N-Bent) from Dashukovskii deposit (Ukraine) with the following chemical composition (wt%): SiO2—50.60, Al2O3—15.58, Fe2O3—8.72, TiO2—0.50, MgO—2.64, CaO—2.07, Na2O—0.20, K2O—0.05 by the known procedure [15] under the following conditions: an OH/Al molar ratio of 2.0, a bentonite content in aqueous suspension of 1.0 wt%, an Al content in bentonite of 20 mmol/g, calcination at 500 °C for 2 h.

К2PdCl4–Cu(NO3)2–KBr/\({\bar{\text{S}}}\) compositions (\({\bar{\text{S}}}\) is N-Bent or Al-PILC) were prepared by incipient wetness impregnation of each support with aqueous-alcohol solution at certain ratios of the above-named salts. The wet mass obtained was kept in a closed Petri dish at 20–25 °C for 24 h and then dried in an oven at 110 °C till constant weight. A palladium(II) content in the samples thus obtained was varied from 1.36 × 10−5 to 4.08 × 10−5 mol/g while a copper(II) content was varied from 1.17 × 10−5 to 8.8 × 10−5 mol/g; a KBr content for each sample was 1.02 × 10−4 mol/g. The catalyst samples were denoted as Pd(II)–Cu(II)/\({\bar{\text{S}}}\) (\({\bar{\text{S}}}\) is N-Bent or Al-PILC).

The X-ray diffraction powder analysis was carried out on a Siemens D500 diffractometer (CuKα radiation, λ = 1.54178 Å) with a secondary beam graphite monochromator.

Analysis of FT-IR spectra in the range from 400 to 4000 cm−1 with resolution of 4 cm−1 was carried out using a Perkin Elmer FT-IR spectrometer. All spectra were recorded using pellets consisting of 1 mg of the material under study and 200 mg of KBr compressed under pressure of 7 tons/cm2 for 30 s.

The thermal behavior of the samples (0.25 g each) was investigated by DTG–DTA using a Paulik, Paulik and Erdey derivatograph in the temperature range from 25 to 1000 °C with a linear heating rate of 10 °C/min. The measurement error was ± 5%.

Water vapor adsorption/desorption was studied gravimetrically using a vacuum setup temperature-controlled at 21 ± 0.2 °C containing a conventional McBain–Bakr silica-spring balance [25].

Protolytic properties of the bentonite surface were estimated by measurement of a pH equilibrium value in aqueous suspensions of 0.2 g N-Bent or Al-PILC in 20 mL of distilled water. pH values were measured by a pH-340 instrument with an ESL 43-07 glass electrode and an EVL 1M3 chlorsilver electrode under continuous stirring of the suspension at 20 °C.

The samples (1 g) of К2PdCl4–Cu(NO3)2–KBr/\({\bar{\text{S}}}\) compositions were tested in a gas-flow setup with a fixed-bed reactor at an initial CO concentration in the gas-air mixture (GAM), \({\text{C}}_{\text{CO}}^{\text{in}}\), of 300 mg/m3 and also 50 mg/m3 (see Fig. 6 and Table 3) and the following GAM parameters: temperature of 20 °C, relative humidity, φGAM, of 67%, a volume flow rate, w, of 1 L/min, a linear velocity, U, of 4.2 cm/s. As a result, an effective residence time, τ′, was of 0.095 s. By monitoring final CO concentrations, \({\text{C}}_{\text{CO}}^{\text{f}}\), a carbon monoxide conversion in a steady state mode, ηst (%) was determined as the parameter characterizing the catalytic activity of the compositions.

3 Results and discussion

3.1 Phase composition

X-ray diffraction patterns for N-Bent, Al-PILC, and Pd(II)–Cu(II)/\({\bar{\text{S}}}\) at CPd(II) = 2.72 × 10−5 and CCu(II) = 2.9 × 10−5 mol/g are shown in Fig. 1. The concentrations of these metal ions were the same in all cases if not specified otherwise. After intercalation of the Al polyhydroxy cation, the montmorillonite, Mont (M in Fig. 1), and α-quarts. Q, phases remained in Al-PILC and Pd(II)–Cu(II)/Al-PILC samples whereas the calcite phase, C, disappeared. No additional reflection were detected in X-ray diffraction patterns 2 and 4 being evidence that initial crystalline Pd(II) and Cu(II) salts, oxide, PdO, CuO, or reduced, Pd0 and Cu0, forms were absent.

Fig. 1
figure 1

X-ray diffraction patterns for N-Bent (1), Pd(II)–Cu(II)/N-Bent (2), Al-PILC (3), and Pd(II)–Cu(II)/Al-PILC (4) (M is montmorillonite, Q is α-quartz, and C is calcite)

Full size image

Catalyst components were homogenized over N-Bent and Al-PILC surfaces as palladium–copper complexes. Table 1 demonstrates the influence of Al polyhydroxy cation intercalation and active component anchoring on XRD parameters of supports and corresponding Pd(II)–Cu(II)/\({\bar{\text{S}}}\) catalyst samples. Taking into account d001 values and a thickness of the aluminosilicate pack (9.6 Å), we determined ∆d = d001—9.6 Å characterizing an interlayer spacing. ∆d Value was the highest for Al-PILC and interlayer spacings decrease in the order Al-PILC > Pd(II)–Cu(II)/Al-PILC > N-Bent > Pd(II)–Cu(II)/N-Bent. Intercalation of the Al13 polycation always results in expansion of aluminosilicate layers. However, according to some reported data, ∆d values depend on the bentonite origin and intercalation conditions.

Table 1 X-ray diffraction parameters for N-Bent and Al-PILC as well as for Pd(II)–Cu(II) compositions based on them
Full size table

For instance, ∆d values for Al-PILCs synthesized from Spanish bentonites were 9.4 Å [26] and 8.8 Å [27], while, for Turkish bentonites, those were 7.99 Å [28] and 8.9 Å [29]. Thus, our result is similar to that reported in [28]. The Al13 intercalation has no influence on the position of (060) reflection (Table 1) testifying that the crystalline structure of the aluminosilicate packs is not changed. After supporting palladium(II) and copper(II) compounds on Al-PILC the inter-pack space collapses that is consistent with some results reported elsewhere [2, 29]. However, in the case of Cu/Al-PILC containing 0.8 wt% of Cu, an expansion of the interlayer distances occurs [10]. The crystallite size, D (nm), for N-Bent is lower than those for Al-PILC and Pd(II)–Cu(II)/\({\bar{\text{S}}}\).

3.2 FT-IR spectral investigations

Figure 2 shows FT-IR spectra for N-Bent, Al-PILC, and Pd(II)–Cu(II)/\({\bar{\text{S}}}\). The Al13 intercalation and metal salts anchoring for both supports result in some changes in their IR spectra. The number and positions of absorption bands in the IR spectrum for Pd(II)–Cu(II)/N-Bent do not change as compared with the IR spectrum for N-Bent whereas IR spectra for Al-PILC and Pd(II)–Cu(II)/Al-PILC substantially differ from the IR spectrum for N-Bent. As can be seen from Fig. 2, the absorption bands at 3692 cm−1 (assigned to stretching vibrations of M–OH structure groups (M is metal), at 876 cm−1, and 915 cm−1 (assigned to deformation vibrations in Al–Fe3+–OH and in Al–Al–OH structure groups, respectively), and at 1421 cm−1 (assigned to stretching vibrations in \({\text{CO}}_{3}^{2 - }\) ion) disappear; the absorption band at 3622 cm−1 (assigned to stretching vibrations in Si–(OH)–Al structure groups) is shifted to a high-frequency region by 15 cm−1 for Al-PILC and by 11 cm−1 for Pd(II)–Cu(II)/Al-PILC; the bands assigned to stretching and deformation vibrations of OH groups in water molecules are shifted to high-frequency region by 10 cm−1 and 13 cm−1, respectively; in the range of Si–O–Si stretching vibrations, instead the shoulder at 1096 cm−1, a clear band at 1047 cm−1 shifted to high-frequency range by 8 cm−1 as compared with N-Bent appears; the band of deformation vibrations in Si–O–Si fragments (468 cm−1) is shifted by 6 cm−1.

Fig. 2
figure 2

FT-IR spectra for N-Bent (1), Pd(II)–Cu(II)/N-Bent (2), Al-PILC (3), and Pd(II)–Cu(II)/Al-PILC (4)

Full size image

All these changes confirm the fact of successful completion of the intercalation process and the sharp profile of Si–O–Si band is a proof of the highly ordered pillar structure. The data obtained are in line with those reported for Al-PILC [27, 30]. For Pd(II)–Cu(II)/Al-PILC, the high-frequency shift (Δν = + 13 cm−1) of the absorption band assigned to stretching vibrations of OH group in the associated water molecules shows that the palladium–copper complexes are anchored just on these sites.

3.3 Water vapor adsorption

The available data concerning water vapor adsorption by pillared clays are very scant [31, 32]. Our water vapor isotherms (Fig. 3) show that the adsorption by Al-PILC and Pd(II)–Cu(II)/Al-PILC is lower than for N-Bent and Pd(II)–Cu(II)/N-Bent.

Fig. 3
figure 3

Isotherms of water vapor adsorption(open circle)/desorption(black circle) by N-Bent (1), Pd(II)–Cu(II)/N-Bent (2), Al-PILC (3), and Pd(II)–Cu(II)/Al-PILC (4) (the isotherms are shifted relative to each other by 0.4 P/Ps)

Full size image

Their hysteresis loops, especially for Al-PILC, are smaller in size and are closed at higher P/Ps values than the loop for N-Bent. These data are evidence of considerable structural changes in bentonite after Al13 intercalation.

Using the BET equation, we estimated the monolayer capacity. am, and the BET constant, C, characterizing the affinity between water molecules and the adsorbent surface. The heats of adsorption for the first layer, Q1, and the specific surface areas, Ssp, were also evaluated using the known formulas [33]. The analysis of these parameters summarized in Table 2 shows that the intercalation and anchoring of palladium(II) and copper(II) compounds onto N-Bent and Al-PILC result in the decrease in am and Ssp values.

Table 2 Structural-adsorption parameters and a thermodynamic activity of adsorbed water for N-Bent, Al-PILC and Pd(II)–Cu(II) compositions based on them
Full size table

The water vapor adsorption isotherms were used for evaluation of a thermodynamic activity of adsorbed water, \(a_{{{\text{H}}_{2} {\text{O}}}}\) = P/Ps, that is an universal quantitative index considering structural and physicochemical properties of porous sorbents, their geometric and energetic inhomogeneity [18]. For comparison of water activity values for different sorbents, they must be determined for the same adsorption value. \(a_{{{\text{H}}_{2} {\text{O}}}}\) Values presented in Table 2 were determined for a values of 2 mmol/g and 4 mmol/g. One can see that the minimal \(a_{{{\text{H}}_{2} {\text{O}}}}\) values were observed for N-Bent.

3.4 Thermochemical properties

As can be seen from thermochemical behavior of the two materials (Fig. 4), Al-PILC and Pd(II)–Cu(II)/Al-PILC, in contrast to N-Bent that demonstrates two endothermic effects, show a single endothermic effect which TM1 for Al-PILC is the same as that in the case of the first endothermic effect for N-Bent. The metal compounds anchored on Al-PILC decrease its TM by 50 °C as compared with pure Al-PILC. The absence of the second endothermic effect \(({\text{T}}_{{{\text{M}}_{2} }} = 580\,^\circ {\text{C}})\) for Al-PILC and Pd(II)–Cu(II)/Al-PILC shows that all surface OH groups of bentonite were involved in the Al13 intercalation process.

Fig. 4
figure 4

Derivatograms for N-Bent (1), Pd(II)–Cu(II)/N-Bent (2), Al-PILC (3), and Pd(II)–Cu(II)/Al-PILC (4)

Full size image

3.5 Protolytic properties

It is known that the Pd2+ and Cu2+ aqua ions in aqueous medium can be hydrolyzed depending on pH [34]. The hydrolysis increases on the surface of porous supports [18, 35]. It is important to know how protolysis of water molecules occurs on acid (T+) and basic (TO−) surface sites

$${\text{T}}^{ + } + {\text{HOH}} \leftrightarrow {\text{TOH}} + {\text{H}}^{ + } ,$$
$${\text{TO}}^{ - } + {\text{HOH}} \leftrightarrow {\text{TOH}} + {\text{OH}}^{ - } ,$$

and also about changes in pH level in aqueous suspensions of the samples under study. Time dependences of pH shown in Fig. 5 for N-Bent and Al-PILC suspensions show significant differences: pH equilibrium values for N-Bent and Al-PILC are 8.97 and 6.03, respectively.

Fig. 5
figure 5

Time dependences of pH values in aqueous suspensions of N-Bent (1) and Al-PILC (2)

Full size image

3.6 Pd(II)–Cu(II)/\({\bar{\text{S}}}\) nanocomposition testing in the reaction of CO oxidation with air oxygen

Figure 6 shows time dependencies of a final carbon monoxide concentration, \({\text{C}}_{\text{Co}}^{\text{f}}\), for the reaction of CO oxidation by air oxygen over К2PdCl4–Cu(NO3)2–KBr/\({\bar{\text{S}}}\) compositions (\({\bar{\text{S}}}\) is N-Bent or Al-PILC). As can be seen, the modification of initial N-Bent by Al13 intercalation permits to prepare the Pd(II)–Cu(II)/Al-PILC nanocatalysts providing a steady state mode of CO oxidation as opposed to К2PdCl4–Cu(NO3)2–KBr/N-Bent composition for which \({\text{C}}_{\text{Co}}^{\text{f}}\) increases with time and becomes equal to \({\text{C}}_{\text{Co}}^{\text{in}}\) in 50 min.

Fig. 6
figure 6

Time dependences of \({\text{C}}_{\text{CO}}^{\text{f}}\) for CO oxidation by air oxygen over К2PdCl4–Cu(NO3)2–KBr/\({\bar{\text{S}}}\) nanocompositions where \({\bar{\text{S}}}\): N-Bent (1) and Al-PILC (2–8). Experimental conditions including CPd(II) and CCu(II) are presented in Table 3

Full size image

Variation of Pd(II) concentration from 1.36 × 10−5 to 4.08 × 10−5 mol/g at CCu(II) = 2.9 × 10−5 mol/g and variation of Cu(II) concentration from 1.17 × 10−5 to 8.8 × 10−5 mol/g at CPd(II) = 2.72 × 10−5 mol/g showed that the maximum catalytic activity (ηst = 33%) and the highest values of Win (an initial reaction rate), Wst (a reaction rate in a steady state mode), kI (a reaction rate constant calculated in the case of a steady state mode at an effective residence time, τ′, of 0.095 s) were attained at CPd(II) = 2.72 × 10−5 mol/g and copper(II) concentrations of 2.9 × 10−5 and 5.9 × 10−5 mol/g (Table 3). Decreasing an initial CO concentration from 300 to 50 mg/m3 (Fig. 6), we obtained the similar values of kI and ηst, that was evidence of the first-order reaction with respect to CO.

Table 3 Effect of palladium(II) and copper(II) concentrations in К2PdCl4–Cu(NO3)2–KBr/Al-PILC nanocatalysts on kinetic and stoichiometric parameters of the reaction of CO oxidation (CKBr = 1.02 × 10−4 mol/g; U = 4.2 cm/s; mcat = 1 g)
Full size table

Based on the results of N-Bent, Al-PILC, and Pd(II)–Cu(II)/\({\bar{\text{S}}}\) characterization, a conclusion can be drawn that the Al13 intercalation into the bentonite structure undoubtedly influence the activity of the nanocatalysts obtained. The changes in the bentonite structure, particularly, a Δd increase, for Al-PILC and Pd(II)–Cu(II)/Al-PILC (Table 1) cause a decrease in their pore-diffusion resistance and increase an access of CO molecules to active sites of the nanocatalyst. Despite the fact that Ssp for Pd(II)–Cu(II)/Al-PILC decreases (Table 2), this composition, as opposed to Pd(II)–Cu(II)/N-Bent, is characterized by clearly defined catalytic properties: the process of CO oxidation over this composition proceeds in a steady state mode (Fig. 6). Thus, a conclusion can be drawn that a Ssp value is not a critical factor influencing the catalytic activity of palladium–copper complexes anchored not only on Al-PILC but also on natural sorbents that has been shown in some our reports [18, 20, 23, 25]. Thermodynamic parameters of adsorbed water, particularly, a water activity and an ability of water molecules to protolytic dissociation depend on structural and physicochemical properties of a support. These thermodynamic parameters, by-turn, can influence the mechanism of surface complex formation and composition of palladium–copper complexes [18, 36].

As compared with N-Bent, in the case of Al-PILC, the Al13 intercalation results in increased \(a_{{{\text{H}}_{2} {\text{O}}}}\) (Table 2) and decreased pH (down to 6.03) values measured in aqueous suspensions and the latter value indicates an increase in activity for hydronium ions (Fig. 5). Taking into consideration values of the first hydrolysis constants of palladium(II) and copper((II) i.e. pK(Pd2+) = 2.0 and pK(Cu2+) = 8.0 [34], one can come to a conclusion that Pd(II) and Cu(II) are hydrolyzed on the N-Bent surface whereas only Pd(II) is hydrolyzed on the Al-PILC surface. As can be seen from our investigations [18,19,20], in the latter case, the surface palladium–copper complexes formed are able to catalyze the reaction of CO oxidation with air oxygen. The maximum effect attained at certain contents of Pd(II) and Cu(II) (Fig. 6 and Table 3) is caused by synergy of their action [18].

4 Conclusions

Pillared clay used in the current study was prepared from Ukrainian bentonite by Al polyhydroxy cation (Al13) intercalation. Nanosized pillared clay (D = 18 mm) and Pd(II)–Cu(II)/Al-PILC nanocatalysts (D = 19 mm) have been obtained. Though no change in crystalline structure for montmorillonite phase is observed, the interlayer spacings (in other words, “free space” between the aluminosilicate packs) for Al-PILC and Pd(II)–Cu(II)/Al-PILC are larger than those for N-Bent and Pd(II)–Cu(II)/N-Bent allowing easier access for CO molecules to active sites of the nanocatalyst. Active components of Pd(II)–Cu(II)/\({\bar{\text{S}}}\) compositions (\({\bar{\text{S}}}\) is N-Bent or Al-PILC) are not transformed to oxides or reduced forms, on the contrary, they are anchored on the support surface as palladium–copper complexes. Their compositions depend on some thermodynamic parameters of a certain support i.e. a thermodynamic activity of adsorbed water, \(a_{{{\text{H}}_{2} {\text{O}}}}\), and hydronium ion activity, \(a_{{{\text{H}}_{3} {\text{O}}^{ + } }}\). Substantial differences in these parameters results in hydrolysis of both palladium(II) and copper(II) on the N-Bent surface (\(a_{{{\text{H}}_{2} {\text{O}}}}\) = 0.46 and pH 8.97) whereas only Pd(II) is hydrolyzed on the Al-PILC surface (\(a_{{{\text{H}}_{2} {\text{O}}}}\) = 0.86 and pH 6.03). These combinations of Pd(II) and Cu(II) forms of existence on the Al-PILC surface leads to appearance of their catalytic activity and synergistic effect at CPd(II) = 2.72 × 10−5, CCu(II) of 2.9 × 10−5 and 5.9 × 10−5, and CKBr = 1.02 × 10−4 mol/g.