The formation of colloidal copper nanoparticles

PAPER

www.rsc.org/pccp | Physical Chemistry Chemical Physics

The formation of colloidal copper nanoparticles stabilized by zinc stearate: one-pot single-step synthesis and characterization of the core–shell particlesw Andre´ Rittermeier,a Shaojun Miao,a Marie K. Schro¨ter,ab Xiaoning Zhang,b Maurits W. E. van den Berg,a Shankhamala Kundu,a Yuemin Wang,a Sabine Schimpf,a Elke Lo¨ffler,a Roland A. Fischerb and Martin Muhler*a Received 22nd April 2009, Accepted 3rd June 2009 First published as an Advance Article on the web 13th July 2009 DOI: 10.1039/b908034a A highly efficient one-step process to generate Cu–Zn colloids was developed, in which the colloidal particles were synthesized from Cu and Zn stearates by reduction with H2 in a continuously operated stirred tank reactor. The resulting spherical, well separated particles have a size of 5–10 nm, consisting of a crystalline Cu0 core (fcc) stabilized by a Zn stearate shell without long-range order. In situ attenuated total reflection FTIR spectroscopy was used to monitor the shift of the C–O stretching vibration of adsorbed CO as a function of temperature and pressure. The absence of the CO rotation–vibration bands of dissolved CO allowed us to obtain FTIR spectra at a CO pressure of 1.0 MPa at 473 K resulting in three shifted CO bands at 2030–2025, 1979–1978, and 1920 cm 1. These bands indicate the presence of reduced coadsorbed Zn species on the metallic Cu surface. Cyclic CO adsorption experiments demonstrated the dynamics of the interaction between the Cu core and the Zn stearate shell.

Introduction With a world production capacity of 48 million tons in 2008, methanol is one of the most important chemicals used in industry.1,2 It serves as a basic feedstock for the production of other chemicals. Furthermore, it is a potential energy source for fuel cell applications.3 Methanol is traditionally produced from synthesis gas (CO, CO2 and H2) over a solid ternary Cu/ZnO/Al2O3 catalyst. In the 1980’s a liquid-phase methanol synthesis process (LPMeOHTM) was introduced for methanol production from coal to improve the quality of methanol and prolong the catalytic lifetime of the Cu catalysts.4 Despite the enormous number of publications on methanol synthesis over copper-based catalysts, the nature of the active sites, the reaction mechanism, the role of Cu and ZnO in the solid catalyst, and further issues still remain controversial.5,6 However, it is commonly accepted that by combining Cu and ZnO a synergetic effect is achieved, which requires an intimate contact between Cu and ZnO. For example, under severe reducing conditions, frequency shifts of adsorbed CO were detected by infrared spectroscopy for Cu/ZnO catalysts, which were proposed to originate from the migration of Zn species a

Laboratory of Industrial Chemistry, Ruhr-University Bochum, Universita¨tsstr. 150, D-44780 Bochum, Germany. E-mail: [email protected]; Fax: +49 234 3214115; Tel: +49 234 3228754 b Chair of Inorganic Chemistry II, Ruhr-University Bochum, Universita¨tsstr. 150, D-44780 Bochum, Germany. E-mail: roland.fi[email protected]; Fax: +49 234 3214174; Tel: +49 234 3223629 w Electronic supplementary information (ESI) available: Further information about the particle size distribution and XRD, XPS, and ATR measurements. See DOI: 10.1039/b908034a

8358 | Phys. Chem. Chem. Phys., 2009, 11, 8358–8366

onto the Cu surfaces (so-called SMSI effect, strong metal– support interactions).7 A homogeneous catalyst with a defined molecular structure could help to address all these questions. There are several reports in the literature about homogeneously catalyzed methanol synthesis,8 typically applying transition metal based catalysts (e.g. Ru, Ni, Re).9 An alternative approach for bridging the gap between heterogeneous and homogeneous catalysts is the preparation of quasi-homogeneous systems consisting of colloidal nanoparticles, which can be applied in gas–liquid methanol synthesis. Colloids of definite size and shape should be less complex than industrial catalysts and could be used to access structure–activity correlations. Furthermore, in comparison to the micrometre-sized grains used in the industrial liquid methanol process,4 the colloidal nanoparticles expose a much higher Cu surface area. So far, methanol synthesis over nanosized trialkylaluminium-stabilized Cu colloids10 and Cu/ZnO1 x colloids11 has been reported. These nanoparticles are highly active for the production of methanol in the liquid phase, but originate from partially complex, expensive, and difficult to handle organometallic compounds. Here, we report on the further development of the simple and easy to scale-up stearate synthetic route developed by Kimura and co-workers12–14 to obtain bimetallic Cu–Ni colloids for the amination of alcohols. This route is extended to easily tunable Cu–Zn colloids for gas–liquid methanol synthesis. The colloids were synthesized from Cu stearate combined with stabilizing metal stearates (M = Zn, Al, Ca) by H2 reduction in the same continuously operated stirred tank reactor (CSTR) used for the catalytic reaction. The generation of stable Cu colloids via stabilization with salts of This journal is

c

the Owner Societies 2009

long chain acids is well known.15 The goal of our work was to generate a colloidal system that is stable under the reaction conditions of methanol synthesis such as 473 K and a CO pressure of 1.0 MPa. The present publication deals with the synthesis and characterization of Cu colloids with different compositions and stabilizers in the reduced state. The catalytic performance and the kinetics of methanol synthesis are part of an ongoing investigation. The Cu–Zn stearate colloid with a Cu : Zn ratio of 50 : 50 was investigated by means of transmission electron microscopy (TEM), selected area electron diffraction (SAED), X-ray diffraction (XRD), X-ray absorption spectroscopy (XAS), and X-ray photoelectron spectroscopy (XPS). The adsorption of CO, one of the feed gas components in methanol synthesis, was monitored under high pressure conditions via in situ attenuated total reflection infrared spectroscopy (ATR-FTIR) to investigate the chemical state of the exposed Cu surfaces.

Experimental Synthesis and catalytic tests of the Cu–Zn stearate colloids Cu-based colloids were prepared from Cu stearate and M stearate (M = Zn, Al, Ca). Cu stearate was synthesized by precipitation in ethanol using cupric acetate and stearic acid according to methods reported in the literature.13 The other metal stearates were obtained from Sigma-Aldrich and used without further purification. For the preparation of the Cu–Zn stearate (Cu : Zn = 50 : 50) colloids, Cu stearate (Cu(CH3(CH2)16COO)2, 2.005 g, 3.18 mmol) and Zn stearate (Zn(CH3(CH2)16COO)2, 2.011 g, 3.18 mmol) were suspended in squalane (C30H62, 200 ml). The colloid synthesis and the catalytic tests were performed in a CSTR operated at high pressures. The suspension was stirred for 20 h at 800 rpm under reducing conditions (5–20% H2 in N2) at 0.5 MPa. The temperature was ramped from 293 K to 493 K with 1 K min 1. Cu–Zn stearate colloids with different Cu : Zn ratios and Cu–M stearate (M = Al, Ca) colloids were synthesized according to this method. All ratios given in the following refer to the atomic metal to metal ratio obtained when combining the stearates. Subsequent to reduction, the reactor was purged with pure N2 and pressurized to 2.6 MPa. The catalytic tests were carried out by switching from N2 to syngas (72% H2, 10% CO, 4% CO2, and balance N2; total flow rate 50 ml min 1 referred to normal conditions). Due to its low boiling point, the formed product methanol was continuously evaporating out of the high-boiling solvent squalane and transported to the online GC and analyzed (two injections, Porapak and molecular sieve columns) every 20 min. The methanol concentrations were in the range up to 2%. The ternary reference powder catalyst was crushed, sieved (o63 mm), suspended in squalane and pre-reduced with diluted hydrogen at 493 K to yield the active Cu/ZnO/Al2O3 state. Transmission electron microscopy (TEM) Samples were prepared under an inert gas atmosphere (glove box) by placing a droplet of the colloidal solution in THF onto This journal is

c


a carbon-coated Au grid. TEM analyses were carried out at 200 kV using a Hitachi H-8100 transmission electron microscope (200 kV, LaB6 filament) and at the Hahn-Meitner Institute in Berlin with a Philips CM30 instrument using an accelerating voltage up to 300 kV. X-Ray absorption spectroscopy (XAS) The absorption edges of Cu and Zn at 8979.0 and 9659.0 eV, respectively, were measured at Hasylab E4 station (Hamburg, Germany). This beamline was equipped with a Si(111) double-crystal monochromator that was used to detune to 50% of the maximum intensity in order to exclude higher harmonics present in the X-ray beam. Sample preparation was performed as follows: after reduction, squalane was removed, and the nanoparticles were transferred to a gas-tight measurement cell excluding any contact with the ambient atmosphere. The spectra m(k) were measured at room temperature in transmission mode using ionization chambers. A metal foil (between the second and the third ionization chamber) was measured at the same time for energy calibration purposes. Data treatment was carried out using the software package VIPER35.16 For background subtraction a Victoreen polynomial was fitted to the pre-edge region. A smooth atomic background m0(k) was evaluated using smoothed cubic splines. The radial distribution function FT[k2w(k)] was obtained by Fourier transformation of the k2-weighted experimental function w(k) = (m(k) m0(k))/m0(k) multiplied by a Bessel window. Duplicate spectra were recorded to verify data reproducibility. X-Ray diffraction (XRD) All X-ray diffractograms were recorded under ambient conditions on a D8 Advanced Bruker AXS diffractometer (Cu-Ka radiation) equipped with a position-sensitive detector in y–2y geometry. X-Ray photoelectron spectroscopy (XPS) The Cu–Zn stearate colloids were analyzed after reduction using high resolution X-ray photoelectron spectroscopy (XPS). Squalane was removed, the nanoparticles were dispersed in THF and placed on a silica wafer to dry in an inert gas atmosphere using a glove box. Before the measurement the samples were shortly exposed to air during transportation. The XPS measurements were carried out in an ultra-high vacuum (UHV) set-up equipped with the Gammadata-Scienta SES 2002 analyzer. A monochromatic Al-Ka (1486.6 eV) X-ray source was used as incident radiation. The base pressure in the measurement chamber was 2 10 8 Pa. The analyzer slit was set to 0.3 mm, and a pass energy of 200 eV was chosen resulting in an overall energy resolution better than 0.5 eV. Charging effects were compensated by a flood gun. The binding energies were calibrated based on the aliphatic C 1s peak at 285 eV as a reference. The XPS peaks were analyzed using a Shirley-type background and a nonlinear least-squares fitting of the experimental data based on a mixed Gaussian– Lorentzian peak shape. Phys. Chem. Chem. Phys., 2009, 11, 8358–8366 | 8359

In situ ATR-FTIR spectroscopy ATR spectra were recorded using a Sensir DuraSamplIR II flow-through cell equipped with a single reflection diamond internal reflection element (IRE). The cell was mounted in a Nicolet Nexus Fourier transform infrared spectrometer (FTIR) equipped with a DTGS and a MCT detector. Spectra were acquired at 4 cm 1 resolution, accumulating 300–750 scans. The ATR cell was operated at temperatures up to 473 K and pressures up to 1.0 MPa. The colloidal dispersion (Cu–Zn stearate colloids in hexadecane) was saturated with gases in a home-made steel reactor equipped with a glass inliner and was continuously circulated through the cell with a gear pump (ISMATEC Reglo-ZS). All flow rates used in the experiments were 50 ml min 1. The gases used had the following purities: H2 (99.9999%), CO (99.997%), CO2 (99.9995%), Ar (99.9999%) and synthetic air (20% O2 in N2) (99.999%, hydrocarbon-free). CO was further purified by a trap filled with zeolite Y operated at 573 K to remove carbonyl impurities.

Results and discussion The reduction of Cu stearate in the presence of Zn stearate in squalane resulted in deep red solutions of Zn stearatestabilized Cu colloids (Cu–Zn stearates (25 : 75) and (50 : 50)). Transmission electron microscopy (TEM) After reduction, spherical Cu particles were observed for the Cu–Zn stearate (50 : 50) colloid (Fig. 1). According to the particle size distribution (Fig. S1, ESIw) over 80% of the particle diameters range from 5 to 10 nm, interspersed with some smaller and larger particles. An average particle size of around 7 nm can be estimated. Lattice spacings matching the Cu(111) plane of fcc Cu are found in the HRTEM image (Fig. 1b). The nanoparticles are well isolated from each other. The interparticle diameter of about 2 nm and the slight tendency towards assembly in hexagonally 2D ordered lattices indicate a stabilization of the particles by adsorbed Zn stearate molecules. EDX analysis confirmed the presence of Zn. To compare the influence of different stabilizing metal species, other metal stearates such as Al and Ca stearate were

Fig. 1 (a) TEM image of the Cu–Zn stearate (50 : 50) colloid after H2 reduction, scale bar = 50 nm; SAED, reflections from center: fcc-Cu(111), (200), (220), (311). (b) HRTEM image, scale bar = 5 nm. Reference: JCPDS pattern 4-0836.


Fig. 2 TEM images of (a) Cu–Al stearate (50 : 50) colloids, (b) Cu–Ca stearate (50 : 50) colloids, scale bars = 200 nm.

chosen as reactants. The TEM images of Al and Ca stearatestabilized Cu particles with a metal to metal ratio of 50 : 50 are shown in Fig. 2. The particles are roughly spherical, and their sizes range from 5 to 30 nm. Fig. 3 shows TEM images of a ternary Cu–(Zn + Al) stearate colloid with a composition of Cu : Zn : Al = 1 : 1 : 0.1. The morphology of the particles did not change upon the addition of Al stearate, and no indication for the presence of crystalline AlxOy species was observed. These TEM images also illustrate that the reduction of Cu stearate by H2 in the presence of stabilizing non-reducible stearates, such as Al and Ca stearate, is a versatile, simple and efficient synthetic route to obtain stable binary and ternary Cu colloids. SAED and XRD To determine the crystalline Cu phases and the oxidation stability of the Cu–Zn stearate (50 : 50) particles, diffraction patterns were recorded in an inert gas atmosphere (SAED) and under ambient conditions (XRD). The SAED pattern after reduction displays reflections corresponding to the (111), (200), (220) and (311) lattice planes of the cubic copper phase (Fig. 1b). Cu0 is exclusively observed when the sample was kept under an inert gas atmosphere. The XRD pattern (Fig. 4) recorded in an ambient atmosphere shows additional reflections, which can be assigned to Cu2O, indicating surface oxidation of the Cu particles. For lower 2y values a high background is obtained, indicating the presence of amorphous material. This could be due to Zn stearate or stearic acid, but could also arise from organic decomposition products. The

Fig. 3 TEM images of a Cu–(Zn + Al) stearate (1 : 1 : 0.1) colloid, (a) scale bar = 100 nm. (b) HRTEM, scale bar = 20 nm.

This journal is

c


Fig. 4 XRD pattern after reduction of the Cu–Zn stearate (50 : 50) particles. Reference: JCPDS patterns 4-0836 (fcc Cu) and 5-0667 (cubic Cu2O).

range below 2y = 301 is dominated by broad overlapping reflections between 2y = 101 and 201 and a sharp reflection at around 2y = 221, originating from stearic acid. Two shoulders at 2y = 19.61 and 22.71 can be attributed to Zn stearate (Fig. S2, ESIw). X-Ray absorption spectroscopy (XAS) The local structure around the constituent metals after reduction was examined by X-ray absorption spectroscopy (XAS) at the Zn and Cu K-edges. In Fig. 5, the spectra of the

samples with a Cu : Zn ratio of 50 : 50 and 25 : 75 are compared to those of a metallic Cu foil and ZnO. The Cu edge reveals that the metallic state is predominantly found in the nanoparticles (Fig. 5a and b). The XANES resemble the structure of the metal foil, and also in the EXAFS (Fig. 5b) the higher shells indicate a typical fcc structure. In the spectrum of the nanoparticles with a Cu : Zn ratio of 50 : 50 a small shoulder is seen on the left side of the main Cu–Cu peak. This shoulder is likely due to a small amount of a light element, possibly carbon or oxygen. In the XANES there was no clear evidence found for the oxidation of the bulk of the Cu particles. When the nanoparticles are deliberately oxidized, however, a strong white line is seen to develop in the XANES (Fig. 5a), which is evidence for the oxidation of the Cu particles. In this case, it was not possible to use Cu EXAFS due to a highly disordered structure as well as a poor signal-to-noise ratio. The Zn K-edge spectra (Fig. 5c and d) show that the XANES and EXAFS were dissimilar to ZnO (shaded traces), and a good agreement was found with the spectra of Zn stearate (dashed traces). Consistent with the XRD observations, the Zn phase lacks long-range order, which would be seen in the case of e.g. crystalline ZnO, as evidenced by the very rapid amplitude decay beyond 4 A˚ (uncorrected FT). Deliberate oxidation of the 50 : 50 sample did not lead to the conversion of the bulk of the Zn stearate to ZnO, as the XANES and EXAFS still resembled the structure of Zn stearate very closely. Thus, the XAS data of the two reduced Cu–Zn colloids are consistent with metallic Cu particles and a Zn stearate phase without long-range order. X-Ray photoelectron spectroscopy (XPS)

Fig. 5 XAS of the colloid samples after reduction. (a) Cu K-edge XANES and (b) Cu EXAFS; (c) Zn K-edge XANES and (d) Zn EXAFS. All samples were measured at room temperature. Shaded: Cu foil (a,b) and ZnO (c,d). Dashed traces: Cu stearate (a,b) and Zn stearate (c,d).

This journal is

c


XPS was applied to further investigate the chemical state of the freshly reduced Cu–Zn stearate colloids. Fig. 6 shows the Cu 2p, C 1s and O 1s spectra of the Cu–Zn stearate (50 : 50) colloid. The corresponding results of the quantitative analysis are summarized in Table 1. The surface concentrations were derived from the intensities of the photoelectron peaks and the corresponding sensitivity factors.17 The Cu oxidation states can be determined based on the Cu 2p and Cu L3M45M45 Auger spectra. In the Cu 2p spectrum (Fig. 6a) no shake-up satellite peaks are observed, confirming that no Cu2+ ions are present. Thus, Cu stearate was reduced in H2 at 493 K. The Cu L3M45M45 spectrum (Fig. S3, ESIw) shows a dominant peak at 916.1 eV (kinetic energy, KE) which is characteristic for Cu+ species. In addition, a small amount of metallic Cu was found, indicated by the weak Cu L3M45M45 peak at 919.4 eV (KE). Based on quantitative analysis, the surface concentration ratio of Cu+ to Cu0 amounts to 6.9 (Table 1). Unfortunately, it was not possible to transfer the samples fully under an inert gas atmosphere into the XPS set-up. Thus, the formation of Cu+ is due to the oxidation sensitivity of the colloids when exposed to air. An incomplete reduction of the Cu stearate is unlikely, because the XAFS and SAED analysis did reveal Cu0 as the only chemical state of Cu present under inert gas conditions. Phys. Chem. Chem. Phys., 2009, 11, 8358–8366 | 8361

Fig. 6 Table 1

XPS data of the Cu–Zn stearate (50 : 50) particles: (a) Cu 2p, (b) C 1s, (c) O 1s regions.

Quantitative analysis of the Cu–Zn stearate particles by XPS

Surface composition (%) Zn 2p

Cu 2p

O 1s

C 1s

Cu+/Cu0

Cu/Zn

1.03

0.65

11.32

87.00

6.9/1

0.63/1

In the Zn 2p spectrum (not shown) two peaks are detected at 1022 and 1045 eV, which are attributed to the 2p3/2 and 2p1/2 components, respectively. Whereas the Zn 2p core levels are not sensitive to the oxidation state of Zn due to the screening effects in the final core–hole state,18 Zn2+ and metallic Zn can be clearly distinguished by the L3M45M45 Auger spectra, which exhibit large changes in the line shape and an energy difference of about 4 eV.19 The Zn L3M45M45 Auger spectrum (Fig. S3, ESIw) is dominated by a peak at 987.2 eV (KE) which is assigned to Zn2+ species. The existence of Zn stearate is also demonstrated by the intense hydrocarbon C 1s peak (C–C and C–H) at 285 eV and the carboxylate (–COO) peak at 288.7 eV (Fig. 6). The latter species is further characterized by the O 1s peak at 531.7 eV being the only feature observed in the O 1s spectrum (Fig. 6). The absence of an O 1s peak around 530 eV indicates that the Zn2+ ions are bound to stearate, and that ZnO is not present. In summary, all structural techniques and the surface analysis by XPS indicate that the reduction in H2 of the Cu and Zn stearates dissolved in squalane leads to colloidal metallic Cu nanoparticles with spherical shape, which are stabilized by a Zn stearate shell. CO adsorption studied by ATR-FTIR spectroscopy CO adsorption studies were performed to determine the accessibility of the Cu core of the colloids and its chemical state. In the literature, three spectral ranges are generally associated with the stretching vibrations of the carbon–oxygen bond of linearly adsorbed CO bound to Cu in different electronic states.20,21 The region of 2220–2150 cm 1 is assigned to CO molecules bound to Cu2+. The range from 2160 to 2080 cm 1 is assigned to CO molecules bound to Cu+, and the bands lower than 2130 cm 1 are assigned to Cu0. 8362 | Phys. Chem. Chem. Phys., 2009, 11, 8358–8366

In the last decades many FTIR studies on Cu single crystals have been performed, observing CO adsorption bands in the region of 2102–2065 cm 1.22,23 Recent IR studies found large differences in the CO frequency during CO adsorption on Cu-based catalysts after different reducing conditions, ranging from 2115 to 2060 cm 1.7,24 Greeley et al.25 applied density functional theory (DFT) to derive CO vibrational frequencies on modified Cu(111) single crystal surfaces including coadsorbed atomic oxygen, Cu–Zn alloys, and coadsorbed Zn atoms. The results of these DFT calculations are summarized in Table 2. Strong shifts to lower frequencies were obtained for coadsorbed Zn atoms.25 For Cu/ZnO catalysts the differences in the CO vibrational frequencies during CO adsorption on Cu were interpreted as an indication of the migration of Zn species onto the Cu surface forming new surface structures.7 Correspondingly, ATR-FTIR spectroscopy is a powerful tool to validate the presence of Zn adspecies on the Cu colloids. Here, first measurements under high pressure conditions in the liquid phase are presented. Due to the absence of the CO rotation– vibration bands because of the limited rotation of the dissolved CO molecules, high pressure measurements were possible without complete absorption of the IR radiation in the region of the CO vibrations. Furthermore, an increased intensity of the FTIR bands was previously reported for CO adsorption on Pd and Pt in the liquid phase.26

ATR-FTIR measurements at room temperature. Fig. 7 shows the IR spectra of CO adsorption on the Cu–Zn stearate (50 : 50) colloid dispersed in hexadecane at room temperature after reduction in the CSTR monitored by in situ ATR-FTIR. The background spectrum used for the difference spectra was the spectrum of the colloid after reduction and transfer to the ATR set-up without any contact to ambient atmosphere. During this measurement the set-up was flushed with Ar. Spectrum a in Fig. 7 was obtained after 360 min in 100% CO atmosphere at 0.4 MPa. After this period of time CO adsorption reached its maximum, represented by a strong band at 2102 cm 1. Simultaneously, bands evolved at 2915, 2847, 1587, 1537, 1457, and 1398 cm 1, which continuously grew during the adsorption of CO. This journal is

c


Table 2 Adapted summary of calculated vibrational frequencies for CO on different Cu(111) surfaces as a function of fractional coverage, Y25 Vibrational frequencies/cm 2093 2073 2158 2101 2089 1924 2035 1952 2058

1

Shift in vibrational frequency w.r.t. reference state/cm 1

Y (ML)

Description

— — 65 8 4 169 38 141 15

1/4 1/9 1/4 1/4 1/4 1/4 1/9 1/4 1/9

Reference states, Cu(111) unstrained

Fig. 7 ATR-FTIR spectra of the CO adsorption on the colloidal Cu–Zn stearate (50 : 50) particles at room temperature: (a) after 360 min in CO (100%, 0.4 MPa); (b) after 220 min in synthetic air (20% O2/80% N2, 0.3 MPa); (c) after additional 180 min in CO (100%, 0.4 MPa); (d) ATR spectrum of hexadecane.

Spectrum b in Fig. 7 was obtained subsequently after 220 min in synthetic air (20% O2/80% N2) at 0.3 MPa to observe the effects due to oxidation. The formerly sharp band of the CO stretching vibration broadened and shifted to higher wavenumbers. The oxidizing conditions had no influence on the bands at 1587, 1537, 1457 and 1398 cm 1. Spectrum c in Fig. 7 was obtained after another 180 min in 100% CO atmosphere at 0.4 MPa subsequent to oxidation. The broad band observed during oxidizing conditions sharpened and shifted back to 2102 cm 1. The bands at 1587, 1537, 1457 and 1398 cm 1 again remained unchanged. Because the ds(CH2) vibrations of hexadecane at 1378 cm 1 were not visible in the measured spectra, any influence of the solvent can be excluded (spectrum d in Fig. 7). The sharp band at 2102 cm 1 can be assigned to linearly adsorbed CO on Cu0 or Cu+. Former investigations assigned bands in this region to CO adsorption on incompletely reduced copper.23 Here, XAS and SAED measurements exclude the presence of Cu+ in the colloids under an inert gas atmosphere. It has been shown that under certain conditions nano-brass is formed.27 In our case, this can be excluded, because the XRD, XAFS and XPS measurements did not provide evidence for the presence of Cu–Zn alloys. The shift of the band to higher wavenumbers is likely due to the This journal is

c


O adatom on Cu(111) Cu3Zn(111) bulk alloy Cu3Zn surface alloy on top of Cu(111) Zn adatoms (hcp sites) on Cu(111) Zn adatoms (fcc sites) on Cu(111)

surface modification originating from the presence of Zn stearate species on the Cu particles, changing the electronic state of the Cu surface. A shift of the band for linearly adsorbed CO on Cu0 to higher wavenumbers upon introduction of Zn species onto Cu colloids has been described before.28 The broadening and the shift to higher wavenumbers during the oxidizing treatment indicates a partial surface oxidation of the Cu particles, which leads to superimposed bands. The second CO treatment demonstrates that the surface of the Cu particles can be re-reduced. The bands at 2915 and 2847 cm 1 and 1457 and 1398 cm 1 are assigned to the nas and ns(C–Hx) vibrations and to the das and ds(C–Hx) vibrations of the stearic groups, respectively. The bands at 1587 and 1537 cm 1 belong to the carboxylate vibrations nas(COO) of Cu and Zn stearate. The increasing intensity of these bands can be explained by increasing adhesion of the Cu–Zn stearate particles on the ATR crystal during the measurements. ATR-FTIR measurements at 473 K and 1.0 MPa. To gain further insight into the properties of the colloidal particles adsorption studies under conditions near methanol synthesis (473 K, 1.0 MPa, in hexadecane) were performed. The reactants Cu and Zn stearate were reduced directly inside the ATR set-up with H2 for 16 h leading to the formation of the metallic Cu core, the Zn stearate shell, and presumably stearic acid and octadecanol: Cu(CH3(CH2)16COO)2 + H2 - Cu + 2CH3(CH2)16COOH CH3(CH2)16COOH + 2H2 - CH3(CH2)16CH2OH + H2O The resulting colloids were subsequently exposed to CO, CO2, and again to CO. First CO adsorption cycle. Fig. 8 shows the first CO adsorption cycle on the Cu–Zn stearate (50 : 50) colloids. The background spectrum used for the difference spectra was the spectrum of the colloids after reduction. Spectrum a in Fig. 8 was obtained after 5 min in 100% CO at 1.0 MPa and 473 K. After this short period of time four bands originating from CO adsorption were observed at 2125, 2025 (sh), 1978 and 1920 cm 1 (sh). Further bands were observed at 1465, 1423 cm 1 and a band with negative intensity at 1574 cm 1. After 10 min (spectrum b in Fig. 8) the intensity of the band at 1920 cm 1 increased and simultaneously the intensity of the band at 1978 cm 1 Phys. Chem. Chem. Phys., 2009, 11, 8358–8366 | 8363

Fig. 8 ATR-FTIR spectra of the CO adsorption on the colloidal Cu–Zn stearate (50 : 50) particles at 1.0 MPa, 473 K after (a) 5 min, (b) 10 min, (c) 30 min in CO (100%) and after (d) 5 min and (e) 40 min in Ar.

decreased. After 30 min (spectrum c in Fig. 8) four bands were observed at 2125, 2025 (sh), 1978 (sh) and 1920 cm 1. During the whole CO treatment the bands at 1574, 1465, 1457 and 1423 cm 1 increased only slightly. The further decrease of the band at 1978 cm 1 is accompanied by an increase of the band at 1920 cm 1. Spectrum d in Fig. 8 was obtained after 5 min in Ar at 1.0 MPa and 473 K. The intensity of the bands at 2125, 2025, and 1920 cm 1 decreased immediately, and the band at 1978 cm 1 appeared again. The intensity of the band at 1423 cm 1 decreased and a new band appeared at 1406 cm 1. Spectrum e in Fig. 8 was obtained after 40 min in Ar. No bands belonging to CO adsorption were observed any more. The band at 1423 cm 1 disappeared, and there was now a strong band at 1406 cm 1. Thus, the adsorption of CO on the Cu surface is reversible, which is also crucial for catalytic activity. In this measurement no adhesion of the Cu–Zn stearate particles at the ATR crystal was observed, because the background spectrum was obtained after the reduction in the ATR set-up. The band at 2125 cm 1 might be assigned to linearly adsorbed CO on Cu+. Since the adsorption took place under highly reducing conditions, this interpretation is not very likely, and the adsorption on Cu0 is favored. The unusually high position of the CO stretching frequency can be explained by the close contact of the positively charged surface with the Zn stearate shell, which results in a shift of the CO stretching frequency to higher wavenumbers (Table 2). Under reducing conditions and at high temperatures the Zn stearate shell is presumably partially reduced to metallic Zn, forming Zn adatoms on the Cu surface. In a comprehensive DFT study Greeley et al.25 calculated the CO stretching frequency on Cu(111) with Zn adatoms, which was up to 169 cm 1 lower than the CO stretching frequency on reference state Cu0 (Table 2) for high coverages of Zn. A similar situation for the present strongly reduced Cu–Zn stearate system is highly likely. Hence, we assume that the three shifted bands, which arise from CO adsorption (frequencies of 2025, 8364 | Phys. Chem. Chem. Phys., 2009, 11, 8358–8366

1978 and 1920 cm 1) at higher temperatures and pressures, can be assigned to CO adsorbed on Cu sites with Zn adatoms. These observations agree with the observed frequency shifts on strongly reduced Cu/ZnO catalysts,7 and provide further support for the presence of strong metal–support interactions under the reducing methanol synthesis conditions in Cu/ZnO catalysts.29,30 During CO desorption, the bands at 1978 and 1920 cm 1 decreased, and simultaneously the band at 1406 cm 1 was formed, which did not vanish in Ar. Therefore, the bands at 1423 and 1406 cm 1 are related to the formation of Zn adatoms. It can be assumed that the adsorption of CO under these conditions has a wide influence in terms of complex interactions and reactions on the organic stabilizers resulting in changes of the vibrations of the organic groups. The band at 1574 cm 1 is related to the carboxylate vibration nas(COO) of stearates, which decreased in intensity due to further reduction. The bands at 1465, 1457, 1423 and 1406 cm 1 can be assigned to the d(OH), das(C–Hx) and ns(COO) vibrations of the organic stabilizers (stearate, stearic acid, octadecanol).31 According to the observed shifts of the CO stretching frequencies, it is reasonable to assume that the Zn stearate shell is partially reduced to stearic acid in 100% CO at 1.0 MPa and 473 K during the ATR measurements: Zn(CH3(CH2)16COO)2 + CO + H2O - Zn + 2CH3(CH2)16COOH + CO2 A clear assignment of the bands in the region from 1600 to 1400 cm 1 and the occurring reactions of the organic stabilizers cannot be made without further investigations. CO2 adsorption cycle. The sharp intensive band at 2336 cm 1 observed after 85 min in 100% CO2 at 1.0 MPa and 473 K (Fig. S4, ESIw) corresponds to CO2 dissolved in hexadecane. No other bands were observed, which consolidates the assumption that no CO2 adsorption and therefore no formation of carbonate-like species took place on the colloids under these conditions.

Fig. 9 ATR-FTIR spectra of the CO adsorption (second cycle) on the colloidal Cu–Zn stearate (50 : 50) particles at 1.0 MPa, 473 K after (a) 5 min, (b) 30 min, (c) 40 min, (d) 75 min in CO (100%), (e) 10 min in Ar.

This journal is

c


Second CO adsorption cycle. Fig. 9 shows the second CO adsorption on the Cu–Zn stearate (50 : 50) colloid. Spectrum a in Fig. 9 was obtained after 5 min in 100% CO atmosphere at 1.0 MPa and 473 K. During this second adsorption the band at 1979 cm 1 also appeared. After 30 min (spectrum b in Fig. 9) the band at 1979 cm 1 was very strong, and the band at 2125 cm 1 also appeared. After 40 min (spectrum c in Fig. 9) the intensity of the band at 1979 cm 1 decreased, and at the same time the intensities of the bands at 2129, 2030, and 1406 cm 1 increased. After 75 min (spectrum d in Fig. 9) only the bands at 2129 and 1406 cm 1 were clearly visible. After 10 min in Ar (spectrum e in Fig. 9) no bands belonging to CO adsorption were observed anymore, whereas the intensity of the band at 1406 cm 1 did not decrease. To further validate that the presence of Zn adatoms is responsible for the three shifted bands at 2030–2025, 1979–1978, and 1920 cm 1, reference CO adsorption measurements with Cu–Al stearate colloids were performed at 1.0 MPa and 473 K. As expected, the spectrum of the CO adsorption on the Cu–Al stearate (50 : 50) colloid (Fig. S5, ESIw) exclusively displayed a small band at 2125 cm 1. Furthermore, this observation indicates that under reducing conditions a Cu–Al stearate colloid is formed, and that the surface of the Cu particles is partially oxidized through close contact with the Al stearate, similar to the effects occurring in the presence of Zn stearate. In summary, the CO adsorption experiments show that Zn stearate stabilizes the Cu nanoparticles, but does not inhibit the adsorption of dissolved species, which is a basic requirement for catalytic applications. The band position of adsorbed CO indicates that during the synthesis of the colloids, Cu stearate is reduced forming metallic Cu0 cores. The particle growth is obviously stopped by Zn stearate, which stabilizes the spherical colloidal Cu nanoparticles by covering their surfaces. Also, after adsorption of different gases, as well as oxidation and re-reduction, the Cu–Zn stearate colloids remain stable. The band positions and their assignments are summarized in Table 3. At room temperature the adsorption of CO on Cu0 was observed, and no changes of the electronic state of Cu took place. At higher temperatures (473 K) the Cu–Zn stearate colloids became a very dynamic system, as indicated by the changes of the CO stretching frequencies during the first (Fig. 8) and the second CO adsorption cycle (Fig. 9). The stabilizing Zn stearate shell was found to interact with the Cu surface in two ways. On the one hand, the Cu surface atoms are partially oxidized due to the presence of the Zn stearate shell (bands at 2102–2125 cm 1). On the other hand, under strongly reducing conditions strong evidence for the presence of Zn adatoms on the Cu surface (three bands at 2030–2025,

Table 3

Table 4 Weight-based catalytic activities of the stearate colloids compared to a Cu/ZnO/Al2O3 powder catalyst (slurry) after 60 h at 2.6 MPa in a feed gas mixture of 72% H2, 10% CO, 4% CO2, and balance N2. The total flow rate was 50 ml min 1 Sample

Rate at 493 K/mmolMeOH gCu

Cu–Zn stearate (50 : 50) Cu–Zn stearate (25 : 75) Cu–Al stearate (50 : 50) Cu/ZnO/Al2O3 powder

6408 5230 2280 6222

1

h

1

1979–1978, and 1920 cm 1) was obtained, originating from a partial reduction of Zn stearate on the Cu surface by CO. These findings contribute to the discussion on the relevance of Zn adspecies for the formation of catalytically active sites under the reducing methanol synthesis conditions.32 Single crystal investigations also indicated that Zn deposited on Cu(111) creates special Cu–Zn sites stabilizing reactive formate species, which were considered important for methanol synthesis.33 Finally, to support this conjecture, the catalytic activity of the Cu–Zn stearate colloids was tested at 493 K in the same CSTR used for the synthesis. The results are summarized in Table 4. It is remarkable that the weight-based rate of methanol formation obtained with the Cu–Zn stearate (50 : 50) colloid at 493 K and 2.6 MPa is roughly the same as that of a conventional ternary Cu/ZnO/Al2O3 catalyst applied as a fine powder in the CSTR.34 Ongoing studies focus on the further morphological and structural changes of the Cu–Zn stearate colloids under methanol synthesis conditions, structure–activity correlations, and the steady-state power-law kinetics.

Conclusions Zn stearate-stabilized colloidal Cu nanoparticles were synthesized in one step by H2 reduction of Cu stearate in the presence of Zn, Al or Ca stearates as stabilizers. The synthesis was performed directly inside the CSTR, simply by mixing the stable and easy to handle metal stearates allowing easy scale-up and tailoring of the colloids for different applications. After reduction of the Cu and Zn stearates, the colloidal solution contained spherical, well separated Zn stearatestabilized Cu nanoparticles (5–10 nm). The core of the nanoparticles consisted of crystalline metallic Cu0, whereas the stabilizing Zn stearate shell was not in a reduced state. CO was used as a probe molecule to assess the accessibility of the Cu surface. The ATR-IR spectra of adsorbed CO at 1.0 MPa CO and 473 K provided evidence for the presence of

Band positions and assignments for the adsorption of CO on the colloidal Cu–Zn stearate (50 : 50) particles

Vibrational frequencies/cm

1

2102 2129–2125 2030–2025, 1979–1978, 1920

This journal is

c

Conditions

Calculated vibrational frequency (DFT)25

After reduction, at room temperature After reduction, at 473 K and a CO pressure of 1.0 MPa After prolonged exposure to 1.0 MPa CO at 473 K

2158, 2093, 2073 — 2058, 2035, 1952, 1924


Assignment CO on Cu0 CO on Cu0 CO on Cu with Zn adatoms

Phys. Chem. Chem. Phys., 2009, 11, 8358–8366 | 8365

reduced Zn adspecies on the Cu surface and for the dynamic behavior of the system.

Acknowledgements Financial support by the Deutsche Forschungsgemeinschaft within the Collaborative Research Center (SFB 558) ‘‘Metal-Substrate Interactions in Heterogeneous Catalysis’’ is gratefully acknowledged.

Notes and references 1 G. Ertl, H. Kno¨tzinger, F. Schu¨th and J. Weitkamp, Handbook of Heterogeneous Catalysis, Wiley VCH, Weinheim, 2008. 2 J. N. OckerbloomThe Methanol & Derivatives Global Outlook, 2000–2012, PCI-Ockerbloom & Co., Inc., Maine, 2008; G. A. Olah, A. Goeppert and G. K. Surya Prakash, Beyond oil and gas: the methanol economy, Wiley-VCH, Weinheim, 2006. 3 B. D. McNicol, D. A. J. Rand and K. R. Williams, J. Power Sources, 1999, 83, 15. 4 M. Sherwin and D. Blum, Liquid Phase Methanol, in EPRI Report EPRI AF1291, Chem Systems Inc., California, 1977; E. C. Heydorn, V. E. Stein, P. J. A. Tjim, B. T. Street and R. M. Konosky, Liquid Phase Methanol (LPMeOHt) Project Operational Experience, Gasification Technology Council Meeting, San Francisco, 1998; From Coal To Chemicals, News release, US Department of Energy, National Energy Technology Laboratory, 2003, http://www.netl.doe.gov. 5 M. M. Gu¨nter, T. Ressler, B. Bems, C. Bu¨scher, T. Genger, O. Hinrichsen, M. Muhler and R. Schlo¨gl, Catal. Lett., 2001, 71, 37. 6 M. Schur, B. Bems, A. Dassenoy, I. Kassatkine, J. Urban, H. Wilmer, O. Hinrichsen, M. Muhler and R. Schlo¨gl, Angew. Chem., Int. Ed., 2003, 42, 3815. 7 N.-Y. Topsøe and H. Topsøe, Top. Catal., 1999, 8, 267; N.-Y. Topsøe and H. Topsøe, J. Mol. Catal. A: Chem., 1999, 141, 95. 8 G. Henrici-Olive´ and S. Olive´, The Chemistry of Catalyzed Hydrogenation of Carbon Monoxide, Springer, Berlin, 1983. 9 P. G. Jessop, T. Ikariya and R. Noyori, Chem. Rev., 1995, 95, 259; K. Tominaga, Y. Sasaki, M. Kawai and T. Watanabe, J. Chem. Soc., Chem. Commun., 1993, 629; M. Marchionna, L. Basini, A. Argano, M. Lami and F. Ancillotti, J. Mol. Catal. A: Chem., 1992, 75, 147; D. Mahajan, Top. Catal., 2005, 32, 209. 10 S. Vukojevic, O. Trapp, J.-D. Grunwald, K. Kiener and F. Schu¨th, Angew. Chem., Int. Ed., 2005, 44, 7978. 11 M. K. Schro¨ter, L. Khodeir, M. W. E. van den Berg, T. Hikov, M. Cokoja, S. Miao, W. Gru¨nert, M. Muhler and R. A. Fischer, Chem. Commun., 2006, 2498. 12 H. Kimura and H. Taniguchi, Catal. Lett., 1996, 40, 123.


13 H. Kimura, K. Matsutani, S.-I. Tsutsumi, S. Nomura, K. Ishikawa, Y. Hattori, M. Itahashi and H. Hoshino, Catal. Lett., 2005, 99, 109. 14 H. Kimura, K. Ishikawa, K. Nishino and S. Nomura, Appl. Catal., A, 2005, 286, 120. 15 C. Salzemann, I. Lisecki, A. Brioude, J. Urban and M.-P. Pileni, J. Phys. Chem. B, 2004, 108, 13242; C. Salzemann, I. Lisecki, J. Urban and M.-P. Pileni, Langmuir, 2004, 20, 11772. 16 K. V. Klementiev, (VIPER for Windows (Visual Processing in EXAFS Researches), freeware). Available from www.desy.de/ Bklmn/viper.html. 17 D. Briggs and M. P. Seah, Practical Surface Analysis, John Wiley & Sons, England, 2nd edn, 1994, p. 635. 18 A. F. Carley and M. W. Roberts, Top. Catal., 1996, 3, 91. 19 C. A. Leighton, A. J. Swift and J. C. Vickerman, Surf. Sci., 1991, 253, 220; J. A. Rodriguez, J. Phys. Chem., 1993, 97, 6509; E. Antonides and G. A. Sawatzky, J. Phys. C: Solid State Phys., 1976, 9, L547. 20 M. K. Schro¨ter, L. Khodeir, J. Hambrock, E. Lo¨ffler, M. Muhler and R. A. Fischer, Langmuir, 2004, 20, 9453. 21 K. I. Hadjiivanov and G. N. Vayssilov, Adv. Catal., 2002, 47, 307. 22 P. Hollins and J. Pritchard, Surf. Sci., 1979, 89, 486; B. E. Hayden, K. Kretzchmar and A. M. Bradshaw, Surf. Sci., 1985, 155, 553; R. Raval, S. F. Parker, M. E. Pemble, P. Hollins, J. Pritchard and M. A. Chesters, Surf. Sci., 1988, 203, 353; F. Boccuzzi and A. Chiorino, J. Phys. Chem., 1996, 100, 3617; J. K. Eve and E. M. McCash, Chem. Phys. Lett., 1999, 313, 575. 23 J. Pritchard, T. Catterick and R. K. Gupta, Surf. Sci., 1975, 53, 1. 24 S. Bailey, G. F. Froment, J. W. Snoeck and K. C. Waugh, Catal. Lett., 1995, 30, 99; N.-Y. Topsøe and H. Topsøe, J. Mol. Catal. A: Chem., 1999, 141, 95. 25 J. Greeley, A. A. Gokhale, J. Kreuser, J. A. Dumesic, H. Topsøe, N.-Y. Topsøe and M. Mavrikakis, J. Catal., 2003, 213, 63. 26 E. Zippel, M. W. Breiter and R. Kellner, J. Chem. Soc., Faraday Trans., 1991, 87, 637; S. D. Ebbesen, B. L. Mojet and L. Lefferts, Langmuir, 2006, 22, 1079. 27 M. Cokoja, H. Parala, M. K. Schro¨ter, A. Birkner, M. W. E. van den Berg, K. V. Klementiev, W. Gru¨nert and R. A. Fischer, J. Mater. Chem., 2006, 16, 2420. 28 M. K. Schro¨ter, PhD Thesis, Ruhr-Universita¨t, Bochum, 2007. 29 J.-D. Grunwaldt, A. M. Molenbroek, N.-Y. Topsøe, H. Topsøe and B. S. Clausen, J. Catal., 2000, 194, 452. 30 R. Naumann d’Alnoncourt, X. Xia, E. Lo¨ffler, O. Hinrichsen and M. Muhler, Phys. Chem. Chem. Phys., 2006, 8, 1525. 31 L. J. Bellamy, The Infrared Spectra of Complex Molecules, Halsted Press, New York, 3rd edn, 1975. 32 M. Spencer, Top. Catal., 1999, 8, 259. 33 I. Nakamura, T. Fujitani, T. Uchijima and J. Nakamura, J. Vac. Sci. Technol., A, 1996, 14, 1464. 34 R. Fischer, R. Fischer, M. K. Schro¨ter, M. Muhler and S. Miao, WO Pat., WO 2009/000 526 (A1), 2009.

This journal is

c
