A new way of assessing the interaction of a metallic phase precursor with a modified oxide support substrate as a source of information for predicting metal dispersion

1 Introduction

The search for new functional materials is one of fundamental tasks of science. One of the catalysts of widest use is platinum, and because of its high cost the ways of its optimum applications continue to be of interest (Bond, 1991; Stepanov et al., 2014; Aminul Islam et al., 2014; Ahmadi et al., 1996; Chiu et al., 2011; Samorjai and Borodko, 2001). Different methods for the synthesis of platinum nanoparticles (PtNs) give particles of different sizes, shapes and distribution of PtNs clusters on the support surface, which determines the reactivity and catalytic selectivity of the catalytic systems obtained (Samorjai and Li, 2010; Ertl et al., 2001). Supported PtNs are most often obtained by impregnation (Spieker and Regalbuto, 2001; Job et al., 2006), co-precipitation (Butler et al., 2001), ion-exchange (MacNevin and Crummett, 1954; Kononova et al., 2010), decomposition of metal clusters (Gates et al., 1995) or PtNs encapsulation into a large surface area porous support (Rioux et al., 2005). The aim of this paper is to present the effects of application of an easy and inexpensive method for investigating metal-ligand interactions in a heterogeneous system. The application of this method provides new information concerning the phenomena taking place at the molecular level depending on the type of support, type of agent coupling with the ligand, type of metal and the liquid-solid environment. This new knowledge can be useful for designing and synthesizing metal-based catalysts. At the first stage a two-component method for the synthesis of SiO₂ support by the sol-gel technique was proposed. This approach allows fabrication of amorphous silica of high purity and highly developed surface area. The silica support was modified by deposition of a ligand that would couple Pt²⁺ ions. At the second stage potentiometric titration in the heterogeneous system (liquid-solid) was applied as the method for controlling the type of metal-support interactions. The platinum complexes synthesized with immobilized amine groups were subjected to reduction and in this way metallic clusters of platinum on the support were obtained. The catalysts obtained were characterized by transmission electron microscopy (TEM), X-ray powder diffraction (XRPD) and volumetric hydrogen chemisorption (VHC). After characterization by several techniques, the catalytic activity of each sample was tested for the N₂O decomposition to N₂ and O₂ at temperatures from the range of 200–600°C.

2 Experimental methods

2.1

2.1 Catalysts preparation

2.1.1

2.1.1 Synthesis of amorphous silica by anhydrous sol-gel method (Nowicki et al., 2014)

Silica gel was obtained by adding 100 cm³ of TEOS (0.44 mol) to an Erlenmeyer flask containing 300 cm³ of anhydrous acetic acid (5.25 mol) upon vigorous stirring. The homogeneous mixture obtained was closed in a glass reactor under autogenic pressure at 353 K for 48 h. The gel was dried at 393 K by slow evaporation for 72 h, into a solid monolith. Then the silica was placed in a muffle furnace and calcined in air atmosphere at 773 K for 4 h.

2.1.2

2.1.2 Immobilization of N-(2-aminoethyl-3-aminopropyl)trimethoxysilane on the silica surface in anhydrous conditions (Chang et al., 2009)

The silica activated with steam (SG) (5.1 g) was dispersed in 200 cm³ of dried toluene and then 1.8 cm³ of N-(2-aminoethyl-3-aminopropyl)trimethoxysilane (Ligand) was added. The mixture was vigorously stirred at 384 K (toluene boiling point) for 10 h. Then the mixture was cooled and filtered in order to obtain a solid product which was a few times washed with dried toluene. The final product was dried at 343 K for 24 h. The synthesis of SG-L system is schematically shown in Scheme 1. Quantitative composition of the organic components on the support surface was determined by elemental analysis (carbon, hydrogen, nitrogen) using an Elemental Analyzer Vario EL III.

Close

Scheme 1 Surface modification of amorphous silica (SG).

Export to PPT

2.1.3

2.1.3 Potentiometric measurements

All pH measurements were performed manually by potentiometric titration in a 75 cm³ glass double-walled vessel, using a titrator Dosimat 725 Metrohm equipped with a 713 pH Meter with a glass electrode 6.0233.100. The electrode was calibrated in terms of hydrogen ions concentration (Miller et al., 2002; Irving et al., 1967), using two standard buffers of pH = 4.002 and pH = 9.255. CO₂ was eliminated from the titration cell by flowing a purging steam of pure argon. Potentiometric titration was carried out at the ionic strength μ = 0.1 M (KCl) at 20.0 ± 0.1 °C, using CO₂-free NaOH solution (0.244 M) as a titrant. The ratio of Pt²⁺ : ligand was fixed as 1:2 (ligand concentration M = 0.1 × 10⁻³ mol dm⁻³, Pt²⁺ concentration M = 0.05 × 10⁻³ mol dm⁻³). The source of platinum ions was a solution of K₂[PtCl₄] in solvent water. The volume of titrated solution after introduction of the support and ligand increased by 1%, but it had no effect on the results of refinement. The calculations were performed using 100–150 results of each titration process. In all systems investigated, no additional precipitations were observed in the whole pH range applied. The complex composition, ligand protonation constant and complex stability constant were estimated by the software package Hyperquad2008 (Gans et al., 1996), while the distribution of species was calculated using the Haltafall program (Ingri et al., 1967). In computer procedure of model selection the criteria of verification were applied and the results of analysis of the programs performance were taken into account (Lomozik et al., 1991, 1995). The presence of platinum complexes (synthesized by the above method) on the silica surface was confirmed by XPS measurements (Nowicki et al., 2016).

2.1.4

2.1.4 Pt nanoparticle synthesis

Particular samples of PtNs supported on SG silica were obtained by cessation of the titration process at a selected pH value, then the suspension was filtered off and dried at 393 K for 24 h. The samples obtained at particular pH, labelled as SG/Pt-pH3, SG/Pt-pH7 and SG/PtpH11, were subjected to reduction in pure hydrogen in a tube furnace Nabertherm R40, using the following thermal regime.

1. Ar – 10 [cm³ min⁻¹] keeping for 10 min at room temperature (rt),

2. H₂ - 10 [cm³ min⁻¹] keeping for 10 min at rt,

3. H₂ - 10 [cm³ min⁻¹] heating for 1 h from rt to 373 K,

4. H₂ - 10 [cm³ min⁻¹] keeping for 15 min at 373 K,

5. H₂ - 10 [cm³ min⁻¹] heating for 1 h to 532 K,

6. H₂ - 10 [cm³ min⁻¹] keeping for 15 min at 532 K,

7. H₂ - 10 [cm³ min⁻¹] cooling from 532 K to 373 K,

8. Ar - 10 [cm³ min⁻¹] cooling from 373 K to rt.

For the sake of comparison a reference sample of Pt/SiO₂ catalyst was prepared by the classical method of silica impregnation with a solution of hexachloroplatinic acid (SG/Pt). The reference sample was subjected to the same reduction with hydrogen. The mass percentage contribution of Pt was obtained from measurements by inductively coupled plasma emission (ICP-OES) on a Vista–MPX Varian emission spectrometer to be 1.1%, 1.1%, 1.0%, 0.9% for SG/PtpH3, SG/PtpH7, SG/PtpH11 and SG/Pt, respectively.

2.1.5

2.1.5 Analytical procedures

TEM observations were carried out using a Philips CM200 TEM microscope operating at 200 kV. The dried powders were sonicated in methanol and then deposited on carbon-film-coated copper grids and dried in air atmosphere. The obtained TEM images were analyzed by ImageJ (Schneider et al., 2012). The average size of platinum particles was calculated on the basis of measurements of at least 100 particles. XRPD measurements were performed on an AXS D8 Advance diffractometer made by Bruker. XRPD diffractograms were recorded using the radiation Cu Kα = 1.5406 Å, in the angular range 10°–90° 2-theta. The size of platinum crystallites was calculated from the Scherrer equation with K = 0.94 for spherical particles (Langford et al., 1978). Chemisorption measurements were made in an ASAP 2010C sorptometer. The samples were evacuated at rt for 15 min and at 532 K for 1 h, then the samples were subjected to additional reduction in hydrogen (40 ml/min.) at 532 K for 1 h and evacuated again at 532 K for 1 h. Hydrogen chemisorption was carried out at 308 K, the isotherms were recorded for 5 pressure values from the range 12–40 kPa. After the first set of pressures, the catalyst was evacuated for the first time (isotherm H_T) and then evacuated again at 308 K for 0.5 h in order to remove the reversibly adsorbed hydrogen (isotherm H_rev), then the whole procedure was repeated. The volume of adsorbed hydrogen was read off by extrapolation of the linear section of the isotherm to zero pressure. The difference between the volume of adsorbed hydrogen found from two isotherms is the amount of hydrogen chemisorbed on the sample (H_irrev = H_T – H_rev). Platinum dispersion was evaluated from the irreversibly chemisorbed hydrogen using the following formula Eq. (1):

(1)

$$\mathrm{D}=\frac{S\bullet M}{a_m\bullet N_A}$$ where S is the specific surface area of the metal, M is the atomic mass of platinum, N_A is the Avogadro number and a_m is the surface per one atom of platinum.

The specific surface area of platinum S was calculated from Eq. (2):

(2)

$$\mathrm{S}=\frac{V_m\bullet N_A\bullet n\bullet a_m}{22414\bullet m\bullet wt}$$ where V_m is the volume of chemisorbed hydrogen corresponding to the monolayer coverage of the platinum area (cm³), n is the stoichiometry of chemisorption, m is the sample mass and wt is the content of the metal in weight percentage %. The stoichiometry of hydrogen adsorption on platinum was assumed as 1:1 (Sinfelt and Yates, 1968).

The catalytic tests were performed in the flow reactor with nitrogen as a carrier gas. The concentration of N₂O in nitrogen was 5000 ppm. A PID controller was used to regulate and control temperature. The gas flow was controlled by mass flow controllers made by Brooks. The analyses were made in the gas analyzer ThermoStar made by Pfeiffer, model GDS301T2 with the analyzer QMA200M. The Pfeiffer Vacuum Quadstar program was used for analysis of the mass spectra and for controlling the process. In all experiments the sample was used in the amount of 100 mg, the systems had been earlier conditioned at 873 K for 2 h in argon atmosphere. Measurements were made in the range 473–873 K, while the rate of gas flow was 50 cm³/min. The m/e signals characteristics of possible products (O₂ – 32, N₂ – 28, NO – 30, NO₂ – 46) and the main compound of N₂O (m/e = 44) were analyzed. The degree of conversion N₂O was determined according to Eq. (3):

(3)

$$\%\mathrm{o}\mathrm{f}{\mathrm{N}}_2{\mathrm{O}}_{\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{v}}=\frac{N_2{O}_{\mathit{in}}-N_2{O}_{\mathit{out}}}{N_2{O}_{\mathit{in}}}\times 100$$

3 Results and discussion

The standard sol-gel process requires the following components: gel precursor, solvent, catalyst and water. The greatest problem of the sol-gel process is the control of the rate of hydrolysis-condensation reaction, which generally is too high and leads to the loss of control of the microstructure of the final product (Brinker and Scherer, 1990). The alternative approach, opening new possibilities, is to initiate the reaction of the sol-gel process without water in the system. The source of silica is TEOS. The use of anhydrous acetic acid as a reagent, a catalyst and solvent simultaneously, makes the system much easier to control. At the beginning, the reaction of acidolysis leads to the formation of alcohol. This alcohol can react with the acetic acid, thus releasing water which would lead to the generation of silanols by hydrolysis. Condensation of silanols generates the three-dimensional gel lattice. The synthesis of gels in such a system permits relatively effective control of the hydrolysis by gradual introduction of water as a product of esterification. In acidic environment it permits getting much more branched spatial gel lattice than that obtained by the classical method. The support obtained by this route does not contain inorganic impurities that by migrating from inside the gel onto its surface could hinder the work of platinum catalyst. The silica obtained in such a binary system shows a monomodal pore distribution and well-developed surface areas (600–700 m² g⁻¹), comparable with those of the silica obtained with the use of template (Nowicki et al., 2014). The obtained silica (SG) was subjected to activation and then amine ligand was deposited on its surface by impregnation. According to the theory of hard bases and acids, Pt²⁺ ions show very strong affinity to nitrogen atoms in amines. We decided to use N-(2-aminoethyl-3-aminopropyl)trimethoxysilane that permitted immobilization of the ethylenediamine multifunctional ligand on the surface of amorphous SiO₂ (SG-L). On the basis of potentiometric titration in the heterogeneous system the number and type of platinum complexes forming on the support surface were found to depend on pH (Nowicki et al., 2016). Fig. 1 presents the types of platinum complexes formed on the support surface. Distribution of particular platinum species as a function of pH is shown in Fig. 2. In determination of the types of platinum species on the support surface by titration in a heterogeneous system it is very important to reduce the effect of the support surface charge by the proper choice of the agent coupling the support surface with the ligand. The proper choice of the coupling agent ensures that the ligand is in the diffusing layer whose thickness depends on the ionic strength of the solution in which metal ions are bonded to the ligands. The types of platinum complexes formed depend on the particular metal ions and type of the ligand, their molar ratio in the system and effectiveness of the ligand deprotonation depending on the type of ligand and pH. On the basis of the curves illustrating distribution of complexes representing different ways of platinum ions bonding, the choice of samples to be reduced in hydrogen was made (Table 1). Sample SG/PtpH3 before reduction contained practically only [Pt(Cl₃)H₂O]⁻ ions involved in electrostatic interactions with fully protonated immobilized amine ligands. It is known from literature that K₂[PtCl₄] in water solution forms [Pt(Cl₃)H₂O]⁻ in significant prevalence. The actual number of these ions depends on the ionic strength of the solution and varies from 70% to 95% (Sanders and Martin, 1961; Grantham et al., 1955). This way of obtaining the platinum catalyst could be treated as a modified ion-exchange method (Brunelle et al., 1978, 1979). In sample SG/PtpH7, 85 wt% of platinum are bonded through coordination bonds with the ligand immobilized on the support surface. The method for the synthesis of this sample corresponds to that of deposition of platinum complexes on the support surface (Muratsugu et al., 2013). In sample SG/PtpH11, over 95% of platinum species are [PtCl₃OH]²⁻, and the method for the synthesis of this sample corresponds to the method of impregnation in which the platinum precursor is adsorbed from a solution onto the support surface (Haber et al., 1995). A reference sample was made of platinum deposited on SiO₂ by the classical impregnation method (SG/Pt) in the same reaction conditions as those applied for the syntheses of the other samples. All the samples obtained were characterized by XRD, TEM and hydrogen chemisorption studies. Fig. 3 presents the XRD diffractograms allowing phase identification of the samples. The broad and relatively high intensity reflection in the range 15–30° 2-theta corresponds to short-range ordering of the neighboring SiO₄ tetrahedrons, proving the presence of amorphous silica. According to ICDD Powder Diffraction Data PDF4+, the same metallic phase of platinum crystalizing in a regular system of space group Fm-3m (s.g. 225) and catalogue number PDF 04-016-6758, (JCPDS PDF4+, 2014) was identified in all samples obtained. The Rietveld method was employed to determine the a parameter of the elementary cell of the samples (Table 2). The TEM image of sample SG/PtpH3 (Fig. 4) shows metallic platinum well-dispersed on the silica support. The particle size distribution obtained for the image of higher resolution (Fig. 6, SG/PtpH3) gives values from 1.5 nm to 6 nm, while the statistical average size, obtained by refinement of Voigt curve gives 3.5 ± 0.17 nm. Potentiometric titration results imply that at low pH the immobilized ligand is protonated, which permits design of a large number of systems that can be subjected to reduction. The TEM image of sample SG/PtpH7 (Fig. 4) shows also monodisperse particle size distribution of platinum species on the support. The particle size distribution (PSD) on this sample (Fig. 6, SG/PtpH7) shows a small shift of the particle size towards higher values, so the particle diameters varied from 2 nm to 7.5 nm. The average size found from the Voigt curve refinement is greater than for sample SG/PtpH3 and equals 4.7 ± 0.11 nm. The system studied Pt²⁺/SG-L in solution corresponds to Pt²⁺ ions and ethylenediamine (en). Because of a very slow reaction of ligand exchange, it is very difficult to establish the stability constants of Pt: en complexes (Paoletti et al., 1984). The situation changes when the ethylenediamine ligands are deposited on the silica support through the silane coupling agent (Nowicki et al., 2016). By changing the amounts of the ligand and metal, it is possible to control changes in the composition and content of certain platinum ions coordinated with the ligands immobilized on the support. The degree of platinum species dispersion and particle size distribution for samples SG/PtpH3 and SG/PtpH7 are comparable to the results reported by Benesi et al. (1968) for PtNs obtained by reduction of [Pt(NH₃)₄]²⁺ions adsorbed on SiO₂. However, Goguet et al. (2002, 2003) who have studied the preparation of PtNs on silica by adsorption of Pt(NH₃)₄(OH)₂ complex onto silica and further reduction under H₂, obtained very small (1–3 nm) and highly dispersed (65%) platinum nanoparticles. In our work we used a mixture of different types of platinum complexes immobilized on silica surface. For samples SG/PtpH11 and SG/Pt (Fig. 5) the nonuniform Pt particle size distributions were observed. According to the PSD calculated for sample SG/PtpH11 (Fig. 6), the size of platinum particles significantly increased and varied from 3 nm to 14 nm, while the average particle diameter was 6.8 ± 0.36 nm, which was also much greater than in the other samples. No PSD was established for SG/Pt as no reliable result could be obtained. In the SG/Pt sample, the particles are not readily recognizable due to the presence of the artifacts, and only a small number ofparticles could be counted. The TEM images of samples SG/PtpH11 and SG/Pt confirmed the main drawback of the impregnation method, which was the active phase of nonuniform particle size distribution, which significantly reduces the possibility of controlling metal particles distribution (Maatman and Prater, 1957). Thus, in order to obtain metallic catalysts by the method based on potentiometric titration in a heterogeneous system, in the range of high pH values the frequent formation of poorly-soluble hydroxo-complexes of transition metals of different charges and having different numbers of OH⁻ groups should be taken into account.

Close

Fig. 1 Structures of Pt(II)/SG-L complexes on the surface of SiO₂ obtained on the basis of potentiometric titration in the heterogeneous system, charge has been omitted for simplicity.

Export to PPT

Close

Fig. 2 Distribution diagrams of Pt complexes in the system Pt(II)/SG-L; percentage contribution of the species : PtHSG-L, Pt(HSG-L)₂, PtSG-L, and Pt(OH) with respect to the total amount of Pt in the system, (C_Pt = 0.05 · 10⁻³ mol dm³, C_SG-L = 0.1 · 10⁻³ mol dm³) (Nowicki et al., 2016).

Export to PPT

Close

Fig. 3 Identification of SG/PtpH3, SG/PtpH7, SG/PtpH11, and SG/Pt, samples made on the basis of XRD powder diffraction data.

Export to PPT

Close

Fig. 4 TEM images of Pt nanoparticles for SG/PtpH3 and SG/PtpH7.

Export to PPT

Close

Fig. 5 TEM images of Pt nanoparticles for SG/PtpH11 and SG/Pt.

Export to PPT

Close

Fig. 6 Particle size distribution (PSD) of Pt nanoparticles for SG/PtpH3, SG/PtpH7, and SG/PtpH11. Statistics based on counting Voigt fit histogram.

Export to PPT

The results of hydrogen chemisorption for all samples are collected in Table 2. The values for chemisorbed hydrogen monolayers were obtained by extrapolation of isotherms to zero pressure. The degree of the active phase dispersion on the support (D_irrev) was determined on the basis of irreversibly adsorbed hydrogen (H_irrev). The platinum species dispersion for samples SG/PtpH3 and SG/PtpH7 showing uniform particle distribution is much higher, reaching 25% and 37%, respectively, than for samples SG/PtpH11 and SG/Pt showing nonuniform particle size distribution, of 9% and 2%, respectively. For the samples of low degree of dispersion (SG/PtpH11, SG/Pt) the contribution of hydrogen weakly reversibly bound to the support (H_rev) reaches 80% and 95%, respectively, of the total hydrogen adsorbed on the sample. As suggested by Vannice et al. (1970), when the dispersion of Pt is low, more reliable results are obtained if Pt surface area is evaluated on the basis of the total amount of adsorbed hydrogen (H_T), than on the basis of chemisorbed hydrogen (H_irrev). Particle size distribution obtained by XRD measurements was not much different from that obtained from TEM images, which means that the samples sonication, carried out prior to TEM measurements, had no significant effect on the size of particles. However, as shown in Table 2, significant differences were observed between the size of particles determined by chemisorption (D_chem) for samples SG/PtpH11 and SG/Pt of much smaller Pt dispersion degree. In comparing the number of free-standing TEM particles and that characterized by selective chemisorption it should be taken into account that some portion of the supported Pt particles is engaged in boding with the silica surface and is unable to chemisorb gas (Spenadel and Boudart, 1960).

Nitrous oxide (N₂O) is a greenhouse gas. Nitrous oxide has a global warming potential about 300 times higher than CO₂. The production of nitric acid is the largest source of N₂O emission in the industrial activities. The nitrous oxide is isolated from technological process at raised temperatures as a nitric acid plant tail gas (Pérez-Ramı́rez et al., 2003). Note that the thermal decomposition of nitrous oxide to nitrogen and oxygen occurs to a measurable extent at about 630 °C, but is speeds up when a catalyst is used (Kapteijn et al., 1996, Kalback and Sliepcevich, 1978). Therefore, we still search for the catalyst for high temperature (400–600 °C) N₂O decomposition. According to literature, Pt-oxides supports have low activity in nitrous oxide decomposition (Parres-Esclapez et al., 2010; Pachatouridou et al., 2015). Ángeles-Pascual et al. (2015) have studied Pt/SiO₂ as a catalyst for the N₂O conversion. At a temperature of 600 °C they obtained approximately 30% N₂O decomposition. Under the same experimental conditions, Arenas-Alatorre et al. (2005) have achieved higher efficiency of N₂O conversion (60%). Catalytic test reaction was conducted for evaluation of samples which were synthesized by potentiometric titration in liquid/solid system. Fig. 7 compares the N₂O decomposition activities of PtNs clusters of SG/PtpH3, SG/PtpH7, SG/PtpH11, and SG/Pt catalyst in the temperature range 400–600 °C. The N₂O conversions at a temperature of 600 °C for SG/PtpH11 and SG/Pt are very similar and equal to 67% and 65% respectively. In contrast, for samples SG/PtpH3 and SG/PtpH7, the degree of N₂O conversion in the same temperature conditions was much higher, reaching 85% and 80%. The data indicate that the activities of SG/PtpH3 and SG/PtpH7 are clearly higher than those of the rest (SG/PtpH11, and SG/Pt) of the catalyst studied. The differences in the degree of conversion of N₂O followed from the difference in metal dispersion between the samples. These results gradually changed the negative opinion about platinum supported on silica as a catalyst for high temperature N₂O decomposition.

Close

Fig. 7 Catalytic activity profiles for catalyst of SG/PtpH3, SG/PtpH7, SG/PtpH11 and SG/Pt during the decomposition of N₂O gas.

Export to PPT

4 Conclusions

The use of potentiometric titration in a heterogeneous system permitted obtaining platinum clusters of different morphology depending on the conditions of synthesis. The platinum complex ions showing strong interaction with the immobilized ligand (SG/PtpH3 – ionic interaction, SG/PtpH7 - coordination) allow obtaining uniform systems of a good degree of dispersion of platinum clusters on the support. The use of titration was found effective for the synthesis of PtNs by reduction of stable complexes of platinum ions and ethylenediamine ligand immobilized on the support surface, while in solution platinum ions do not form stable species with ethylenediamine. The complex ions showing weak interaction with the immobilized ligand (SG/PtpH11 – adsorption of [PtCl₃OH]²⁻) give nonuniform systems of low degree of metallic platinum dispersion on amorphous silica. Analysis of TEM images of samples SG/PtpH11 and SG/Pt confirmed the main drawback of the impregnation method, that is poor control of distribution of platinum nanoparticles on the support as the Van der Waals interactions are too weak to force the active phase of uniform particle size distribution.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Conflict of interest

None.