Plasma Synthesis of Highly Dispersed Metal Clusters Confined in Nanosized Zeolite


  • Dr. Mickaël Rivallan,

    1. Laboratoire Catalyse et Spectrochimie, ENSICAEN, Université de Caen, CNRS, 6 Bd Maréchal Juin, F-14050 Caen (France), Fax: (+33) 2-3145-2822
    Search for more papers by this author
  • Ivan Yordanov,

    1. Laboratoire Catalyse et Spectrochimie, ENSICAEN, Université de Caen, CNRS, 6 Bd Maréchal Juin, F-14050 Caen (France), Fax: (+33) 2-3145-2822
    Search for more papers by this author
  • Dr. Sébastien Thomas,

    1. Laboratoire Catalyse et Spectrochimie, ENSICAEN, Université de Caen, CNRS, 6 Bd Maréchal Juin, F-14050 Caen (France), Fax: (+33) 2-3145-2822
    Search for more papers by this author
  • Dr. Christine Lancelot,

    1. Unité de Catalyse et Chimie du Solide, Ecole Nationale Supérieure de Chimie de Lille
    Search for more papers by this author
  • Dr. Svetlana Mintova,

    1. Laboratoire Catalyse et Spectrochimie, ENSICAEN, Université de Caen, CNRS, 6 Bd Maréchal Juin, F-14050 Caen (France), Fax: (+33) 2-3145-2822
    Search for more papers by this author
  • Dr. Frédéric Thibault-Starzyk

    1. Laboratoire Catalyse et Spectrochimie, ENSICAEN, Université de Caen, CNRS, 6 Bd Maréchal Juin, F-14050 Caen (France), Fax: (+33) 2-3145-2822
    Search for more papers by this author


Plasma preparation: Low-pressure plasma treatment of a templated Pt/BEA zeolite results in fast and total degradation of the organic template with simultaneous reduction of the Pt cations to clusters with a size of 1 nm. Operando FTIR study confirms the enhanced mechanism of reduction of Pt cations to clusters entrapped in the BEA zeolite pores.

original image

Metal-containing zeolites (and, more generally, porous materials) are very important for catalysis, electronics, and sensors, but their preparation is still a problem. Reproducibility and control of the properties of the metal cluster is difficult. Broad size distribution of the metal particles together with a low stability of the clusters is often observed. New methods have been reported, such as UV photolysis and γ irradiation of hydrated metal complex ions incorporated within the microporous molecular sieves.14 Organic templates such as tetraalkylammonium cations can stabilize separate metal clusters and prevent agglomeration and also limit the metal loading in the material.511 However, such solids still need to be heat treated (calcined), which may result in a slow agglomeration of the clusters and a possible migration to the outer surface of the matrix, with pore blocking and low energy efficiency.12 Postsynthetic treatment by nonthermal plasma was recently applied to metal-loaded templated mesoporous materials (MCM and SBA-15).1316 The plasma is believed to dissociate and ionize residual water molecules in the mesoporous matrix, creating hydrated electrons equation image and radicals H⋅ for reduction of the metal ions to clusters.17 The authors however still needed to use additional water and apply a final thermal treatment to remove residual impurities, resulting in partial sintering of the metal clusters.18

We report herein a one-step environmentally friendly approach for the preparation of active and highly dispersed metal clusters confined predominantly in the channels of nanosized BEA type zeolite (diameter of individual particles of 10 nm) by plasma treatment. This approach, which was discovered using operando FTIR in the plasma reactor, leads to a fast and simultaneous calcination of the nanosized molecular sieve (removal of the organic template) and reduction of the metal cations to clusters.

Prior to calcination, the nanosized pure BEA and Pt-containing BEA (Pt/BEA, 0.75 wt % Pt) zeolites were directly subjected to Ar or O2 cold plasma treatments at low pressure (350 Pa). The changes in the samples under plasma treatment followed by operando FTIR spectroscopy are shown in Figure 1.

Figure 1.

IR spectra of as-synthesized Pt/BEA 0.75 in (a) Ar, (b) O2 plasma and (c) pure BEA in O2 plasma as a function of time (0–25 s).

The initial spectra of the pure BEA and Pt/BEA samples exhibit the methyl stretching modes at 3100–2800 cm−1 originating from the tetraalkylammonium (TEA) organic template (Figure 1).19, 20 From the beginning of the plasma ignition, the intensity of the CH bands decreased due to the removal of the TEA template. In oxygen plasma, the IR feature of acidic Brønsted sites appeared at 3600 cm−1 (Figure 1 b, c). This phenomenon is better pronounced in the Pt-containing BEA samples (Figure 1 b). The degree of degradation of the organic template (D %) could be calculated by integrating the total absorbance, AHC, in the region of 3100–2800 cm−1 as a function of time according to the Equation (1):

equation image((1))

where equation image corresponds to AHC at t=0.

Additionally, the pseudo-reaction rate (r g−1 s−1) is expressed as a function of AHC according to the Equation (2a):

equation image((2a),)

where m is the weight in grams of the self-supported zeolite wafer.

On Pt/BEA samples, the presence of tetraammineplatinum(II) nitrate [Pt(NH3)4(NO3)2] (Pt source) is not clearly detected by IR spectroscopy, since the NH3 vibrations overlap with the OH bands. In addition to the region of the organic template, the vibration at 1630 cm−1 corresponding to water is also changed by the Ar or O2 plasma.21 The degradation of the template and desorption of the residual water from Pt/BEA in O2 or Ar plasma were calculated from the respective IR absorbances as a function of time (Figure 2 a).

Figure 2.

(a) Degradation of the template (square) and desorption of the residual water (circle) for Pt/BEA 0.75 in O2 or Ar plasma (P=350 Pa, half and full symbols respectively) as a function of time and (b) Evolution of ln (equation image) for the same sample in O2 plasma.

Water is often claimed to be the main reducing agent during conventional plasma treatment of metal-supported catalysts, without organic template, leading to the formation of metallic clusters (M0) according to reactions (1) and (2):17

equation image((1))
equation image((2))

In this study, the residual water in the nanosized BEA zeolite clearly enhanced the rate of template degradation during the first seconds (ca. 5 s) of Ar plasma treatment (Figure 2 a). Once the water was totally removed, the degradation rate seriously decreased. This indicated that the radical species incoming from water in reaction (1) are consumed and/or progressively removed under the dynamic conditions.

The O2 plasma treatment is carried out in excess oxygen. After water removal, the reaction rate given in Equation (2a) could be expressed as:((2b))

equation image((2b))

where k (given in g−1 s−1) is the pseudo-rate constant and α (dimensionless) is the rate order in the organic template.

Consequently, from the linear character of the plot of ln (equation image) versus time after water removal (t=5 s), α is determined to be close to 1 for the sample under O2 plasma treatment (Figure 2 b). In these conditions (P=350 Pa), a reaction half-life (t1/2) of approximately 7 s was calculated, whereas, under Ar plasma conditions, the half reaction was not reached. It may be concluded that the Ar plasma is an efficient process for water removal with moderate alteration of the template present in the channels of the zeolite nanoparticles.

To only evaluate the effect of O2 plasma on the degradation of the organic template and to negate the effect of the residual water, the Pt/BEA and pure BEA samples were subjected to Ar plasma pretreatment for 10 s.21 In the absence of water, the spectroscopic fingerprints of the TEA template are progressively consumed. In the meantime, the band at 3600 cm−1, due to the acidic Brønsted sites, became sharp and well pronounced. The O2 plasma process removed the organic template from the zeolite micropores, thus a validation of the previous results (Figure 1 b,c) was provided.

The rate of template degradation also appeared to be dependent on the O2 pressure. The higher the pressure, the faster the total degradation of the template is achieved. To calculate the degradation rate of the organic template as a function of oxygen pressure, we assumed the following Equation (3):

equation image((3))

where β (dimensionless) is the rate order in oxygen.

As the experiments were performed in excess oxygen, the oxygen partial pressure is assumed to be constant during the template removal and the integration of Equation (3) leads to Equation (4):

equation image((4))

The linear plots representing the template degradation at different oxygen pressures (Pmath image=250, 350, or 450 Pa) on sample Pt/BEA are shown in Figure 3 a. From the slopes of the three linear plots representing the template degradation as a function of the oxygen pressure, the rate order in oxygen (β) is calculated (the last several points are taken into consideration). It appears that this rate order is close to 3, which confirms the great influence of O2 pressure in the template degradation process.

Figure 3.

Neperian logarithm of the template degradation in O2 plasma as a function of time: a) on Pt/BEA 0.75 at various oxygen pressures; ▪:  Pmath image=250 Pa, t1/2=18 s; ▴:  Pmath image=350 Pa, t1/2=7 s; •:  Pmath image=450 Pa, t1/2=5 s; b) at a constant oxygen pressure of 350 Pa; ▴: on pure BEA, t1/2=42 s; •: on Pt/BEA 0.12, t1/2=16 s; ▪: on Pt/BEA 0.75, t1/2=7 s. The residual water is previously removed by Ar plasma treatment for 10 s.

Moreover, the O2 plasma treatment (350 Pa) is carried out on pure BEA and 0.75 wt %and 0.12 wt % Pt/BEA and pretreated with Ar plasma to remove the residual water. It is clear that the template degradation also occurs on pure BEA zeolite at room temperature (cold plasma), but the degradation rate is very slow (Figure 3 b). Therefore, for the total elimination of the organic template from pure BEA a longer treatment time is required in comparison to the Pt/BEA samples. To evaluate the influence of Pt metal on the rate of template degradation, ln (equation image) was plotted against time (Figure 3 b). No straight line was obtained, except that for the pure zeolite BEA. For the Pt/BEA samples, the Pt cations were progressively reduced in the O2 plasma and consequently started to be active during the process of degradation of the TEA organic template. The reduction of Pt cations to metallic Pt clusters during plasma treatment was subsequently confirmed by CO chemisorption and is discussed hereafter. From the slopes of these curves (the last several points are taken into consideration) representing almost the completely reduced Pt, the kinetic constant of the reaction (k) is calculated. The ratio between the constants found for the Pt/BEA 0.75 and pure BEA is about 30. The results indicate that, in the case of Pt/BEA samples, a clear signature for hybrid plasma-catalytic reaction is seen, where the Pt atoms participate in the elimination of the organic template present in the zeolite channels.

The crystalline structure of pure BEA and Pt/BEA samples before and after plasma treatment was investigated by XRD (Figure 4 b). All of the Bragg reflections expected for highly crystalline BEA zeolite were detected in the resulting XRD patterns, and a pronounced peak broadening due to the very small grain size of BEA zeolite is observed. Furthermore, two Bragg reflections at 2θ=39.76 and 46.25°, corresponding to (111) and (200) facets of Pt clusters, respectively, appear in the pattern for the Pt/BEA sample. These peaks are not present in the pattern of the ion-exchange sample prior to the plasma treatment, confirming the formation of platinum clusters (ca. 1 nm in size) in the Pt/BEA after plasma treatment. The size of the Pt clusters calculated using the Scherrer equation is in agreement with the TEM observations (Figure 4 a).

Figure 4.

a) TEM image of Pt/BEA nanocrystals after cold O2 plasma treatment at a pressure of 350 Pa; b) XRD patterns: 1) pure zeolite BEA; 2) ion- exchanged Pt/BEA 0.75; 3) Pt/BEA after O2-plasma treatment (the Pt Bragg peaks are shown on the bottom of Figure 4 (b); c) IR spectra of CO chemisorbed on Pt/BEA treated by O2 plasma at Pmath image=350 Pa.

The very high crystallinity of the nanosized BEA zeolite after plasma treatment can be seen in Figure 4 a. The BEA crystals with individual grains of 10 nm have very well-aligned crystalline fringes, which correspond to a BEA-type framework structure. The TEM images clearly indicate the presence of small and well-dispersed Pt clusters in the pores of the BEA nanocrystals (Figure 4 a, inset). This was also confirmed by CO chemisorption (monitored by FTIR spectroscopy, Figure 4 c) performed on O2 plasma-treated Pt/BEA in the plasma reactor. From the latter experiment, additional important information could be gleaned: a) the Pt atoms are accessible to the CO probe;22 b) the Pt cations are reduced to Pt0 (only one band is centered below 2100 cm−1) in the plasma treatment; c) a very high dispersion of Pt atoms above 60 % is achieved;23 d) the distribution of the Pt clusters is very homogeneous throughout the sample (a symmetrical CO band is measured).

The mechanism describing the decomposition and complete elimination of the organic template is still not clear. However, the species coming from the residual water in the BEA nanocrystals are clearly discarded. Many others species (radicals or other) could be involved during the O2 plasma treatment of metal-containing zeolites (O, O+, O⋅, O3). In addition, the molecules formed during the degradation of organic template, such as COx, NOx, and NHx, under oxidative conditions could also participate. Additional study has to be performed to clarify the complexity of the plasma process applied to metal- containing zeolites, which involves decomposition of organic template, water removal, and formation of metal clusters.

Plasma technology allows a one-step approach for the synthesis of highly dispersed Pt nanoparticles confined within BEA zeolite. Argon plasma treatment was found to be limited to residual water desorption, whereas the use of O2 as the gas-forming plasma strongly enhanced the elimination of the structure-directing agents from the pores of the BEA nanocrystals. The low amount of Pt ions in the templated molecular sieves was still enough to greatly enhance the mechanisms of template removal by a catalytic process.24 Such template removal occurred at nearly room temperature whereas the calcination process typically requires temperatures as high as 823 K. The benefits of such low-temperature plasma-assisted syntheses are multiple; firstly a very high and homogeneous dispersion (above 60 %) of metal clusters on the as-prepared catalysts is achieved, and secondly a fast and low-cost preparation procedure is applied.

Experimental Section

Nanosized BEA crystals were synthesized from a clear water precursor solution having the following molar composition: 15 TEAOH: 1 Al2O3: 25 SiO2: 375 H2O. Tetraethylammonium hydroxide (TEAOH, 35 wt %, Aldrich) was used in excess as an organic template; aluminum tri-sec-butoxide (98 wt %, Aldrich) and Cab-o-sil M5 (Riedel-DeHäen) were used as alumina and silica sources, respectively. The suspensions after aging for 3 days at ambient conditions were hydrothermally treated at 373 K for 3 days. The crystalline nanosized BEA particles were purified by two-step centrifugation (24 500 rpm, 2 h) and freeze dried. The ion-exchange of BEA was carried out in one step (333 K, 24 h) using [Pt(NH3)4(NO3)2] as platinum source; the ion exchange sample was washed with double-distilled water (3×200 mL) to remove the excess of Pt.

The crystalline nature of the material was determined by X-ray diffraction (XRD) on a STOE STADI-P diffractometer with CuKα1 source (λ=0.15406 nm, 40 kV, 30 mA) in Debye–Scherrer geometry, equipped with a linear position-sensitive detector and employing Ge monochromator. The chemical compositions of the BEA and Pt/BEA powders were determined by X-ray fluorescence (XRF) analysis with a MagiX PHILIPS PW2540. The nanosized BEA crystals had a Si/Al ratio of 14, and the BEA was loaded with 0.75 or 0.12 wt % Pt2+ (hereafter referred as Pt/BEA 0.75 and Pt/BEA 0.12 respectively). The operando FTIR spectroscopy studies were performed on self-supporting wafers (10 mg, 20 mg cm−2) placed in the plasma reactor, where a desired gas composition was sustained at low pressure (350 Pa). The dielectric barrier discharge was ignited with a 50 Hz sinusoidal power supply (2 kV) between two electrodes fused to the IR cell. During the experiments, the temperature increase due to cold plasma process was below ΔT=20 K. The IR spectra were collected in transmission mode on a Bruker Vertex 80v equipped with a cryogenic MCT detector and run at 4 cm−1 resolution (8 scans; acquisition time≈1.2 s).

CO chemisorption experiments were monitored by FTIR spectroscopy to evaluate the amount of Pt atoms accessible in the BEA matrix and to estimate the dispersion degree.12, 25 The changes of the equation imageCO IR bands corresponding to CO chemisorbed on Pt were evaluated; CO chemisorption on metal particles was assumed to take place with a Pt/CO ratio of 1:1 (no residual gas phase in the IR cell before sample saturation and nearly no bridged CO or carbonates were detected). The equation imageCO band was integrated in the range 2120–2000 cm−1 after each introduction of CO.


The authors would like to acknowledge S. Aiello for the help in the conception and development of the plasma FTIR reactor.