Anand Pariyara,
Siddharth Gopalakrishnanb,
Joseph Stansberyb,
Rajankumar L. Patelb,
Xinhua Liangb,
Nikolay Gerasimchukc and
Amitava Choudhury*a
aDepartment of Chemistry, Missouri University of Science and Technology, Rolla 400 W 11th Street, Rolla, MO 65409, USA. E-mail: choudhurya@mst.edu; Fax: +1-573-341-6033; Tel: +1-573-341-6332
bDepartment of Chemical and Biochemical Engineering, Missouri University of Science and Technology, Rolla, MO 65409, USA
cDepartment of Chemistry, Missouri State University, Springfield, MO 65897, USA
First published on 8th April 2016
Pyrolysis of a 1-D polymeric cobalt(II) coordination complex ([Co(BDC)(Mim)2]n, H2BDC = benzenedicarboxylic acid; Mim = N-methylimidazole) results in the formation of carbon embedded fcc cobalt nanoparticle composites, Co@C. The as-prepared Co@C shows an agglomerated secondary structure with a highly embedded carbon shell comprising of cobalt nanoparticles of 20–100 nm. These Co@C particles show excellent catalytic activity in the reduction of nitrophenol to aminophenol, studied as a model reaction, and evolves as a promising candidate for the gas phase reduction process.
Single-crystal X-ray diffraction (SCXRD) intensity data sets were collected on a Bruker Smart Apex diffractometer with monochromated Mo Kα radiation (0.7107 Å). Powder X-ray diffraction (PXRD) pattern was obtained from a PANalytical X'Pert Pro diffractometer equipped with a Cu Kα1,2 anode and a linear array PIXcel detector over a 2θ range of 5° to 90° with an average scanning rate of 0.0472° s−1. Thermogravimetric analysis (TGA) has been performed on the sample with a TA instrument Q50 TGA with a scan rate of 10 °C min−1 under N2 flow rate of 40 mL min−1. FT-IR spectrum was collected using Thermo Nicolet Nexus 470 FT-IR spectrometer over 400–4000 cm−1 on a sample embedded in KBr pellet. UV-Vis spectroscopy was performed in Varian CARY 100 Bio UV-Vis spectrophotometer. Temperature dependent UV-Vis spectra were recorded in 1 cm quartz cuvette (Starna) using the Agilent 8453 diode array spectrophotometer equipped with TC1 Peltier temperature controller by Quantum Northwest. Independent temperature control was provided by Cen-Tech 92242 digital thermometer with K-type probe. Scanning electron microscopy (SEM) images were taken using Helios Nanolab-600 equipped with Energy Dispersive Spectrometry (EDS) detector (Oxford Instrument) for elemental analysis. X-ray Photoelectron Spectrometry (XPS) was performed on a Kratos Axis 165 Photoelectron Spectrometer. The binding energies of all peaks were corrected as compared with reference peak of adventitious carbon (C1s, 284.8 eV).30 TEM images were obtained on FEI Tecnai F20 and Tecnai Osiris TEM operating at 200 kV. The as-pyrolized sample was dispersed in an acetone solution to prepare the TEM sample. A drop of the “as dispersed” solution was placed onto a carbon coated TEM grid and dried in air prior to TEM imaging and EDS. High resolution TEM was obtained with the Tecnai Osiris operated at 200 keV with a probe current of 2.5 nA. Raman measurement was performed in Horiba Jobin Yvon spectrometer using Nd:YAG laser with 30 s exposure time. Gas adsorption experiments were carried out in Quantachrome's Autosorb-1 surface area measurement instrument. Before the start of the gas-adsorption experiments, the as-prepared samples were outgassed at 120 °C under vacuum for 12 h. The specific surface areas of the samples were calculated using the Brunauer–Emmett–Teller (BET) method. The pore size distribution curves were derived from the adsorption and desorption isotherms using the Barrett–Joyner–Halenda (BJH) methods.
Synthesis of Co@C (2). The coordination polymer, 1 (0.5410 g) was taken in a heating boat and subjected to pyrolysis in a CVD furnace at 550 °C for 2 h under N2 gas with 50 SCCM flow rate. The sample was heated at a rate of 100 °C h−1 till 550 °C and held at that temperature for 2 h followed by cooling down to room temperature at the rate of 100 °C h−1. Such treatment yielded a black solid product (0.1550 g) with a weight loss of 76.8%. CHN elemental analysis C: 35.05%.
Fig. 1 Experimental (as synthesized) and simulated (single crystal diffraction data) PXRD pattern of 1. |
Empirical formula | C16H16N4O4Co |
Formula weight (g mol−1) | 387.26 |
Wavelength (Å) | 0.71073 |
Crystal system | Triclinic |
Space group | P |
a (Å) | 7.3256 (7) |
b (Å) | 9.0006 (9) |
c (Å) | 13.7874 (14) |
α (°) | 82.0240 (10) |
β (°) | 78.2670 (10) |
γ (°) | 72.3450 (10) |
Volume (Å3) | 845.25 (15) |
Z | 2 |
Calculated density (mg m−3) | 1.522 |
Goodness of fit on F2 | 1.056 |
Final R indices I > 2sigma | R1 = 0.0426 wR2 = 0.1059 |
R indices (all data) | R1 = 0.0503 wR2 = 0.1101 |
Largest diff. peak/hole | 0.460/−0.240 |
The Co(II) in 1 adopts a tetrahedral coordination with two oxygen atoms from two BDC units in bis-monodentate (syn–anti) fashion along with two N-donors from two Mim units (Fig. 2a). The polymeric 1-D chain is propagated via bridging BDC connecting Co atoms in a zig–zag fashion (Fig. 2b). The Co1–Co1′–Co1′′ angle is 100.56(1)° and the distance between two Co centers is 10.74(8) Å and 10.94(7) Å depending on the oxygen it is bonded to, O1 for the former and O3 for the latter (Table S1, ESI†). This is related to the fact that the Co1–O1 and the Co1–O3 bond lengths are 2.007(2) and 2.008(3) Å, respectively, which are typical for a tetrahedral Co(II) ion bonded to carboxylate ion.34 The Co(II) oxidation state is further supported by Co–N bond length of 2.029(2) Å (Co1–N1) and 2.056(2) Å (Co1–N3).34 Similar chain like structures are reported in other Cu, Zn and Ni complexes with bridged BDC where the M–M–M (M = metal) angles range between 65° to ideal 120° and are noted to be a function of steric contribution from the co-ligand.35–37 The 1-D polymeric complex 1 packs as infinite chains along a-axis via inter-chain hydrogen bonding to form 2-D array (Fig. 3a).
Two strong hydrogen bonding with donor (D)–acceptor (A) distances of 2.59 and 2.55 Å with D–H⋯A angles of 167 and 160°, respectively, for C10–H10A⋯O2 and C12–H12C⋯O1, were noted between the N-methyl hydrogens (Table S2, ESI†) and the metal coordinated carboxylate oxygen. Furthermore, the 2-D layers give rise to 3-D supramolecular network (Fig. 3b) via C–H⋯π interactions (Table S3, ESI†) as evident from short C–H⋯π distance of 2.83 to 2.93 Å. In FT-IR spectrum (Fig. S1, ESI†), coordination of carboxylic group to the metal center is evident from the change of ν(CO) band to lower wavelength, from 1720 cm−1 in uncoordinated BDC to 1588 cm−1 in 1.35–38 The presence of Mim unit is also evident through characteristic peaks at 2963, 1425 and 1360 cm−1, known for coordinated Mim.38 Thermogravimetric analysis (TGA) reveals incremental loss, first, at 200 °C and an overall mass loss of 42.4% at a temperature of 400 °C, that can be accounted for the loss of both Mim moieties (Fig. S2, ESI†) from the overall structure of 1. The decomposition of the Co-BDC backbone is therefore expected to decompose from temperature above 400 °C. Interestingly, PXRD of the calcined product of 1 after the TGA which was carried out up to 850 °C did not show any formation of oxides or carbides but metallic cobalt and the final weight (∼23%) found was slightly higher than required for just metallic cobalt (15.2%). This indicates retention of some chemical residues presumably carbon along with metallic cobalt. The polymeric complex 1 was then subjected to controlled pyrolysis under N2 in a CVD furnace. The synthesis was optimized at 550 °C and a dark carbonaceous solid Co@C (2) was formed. The PXRD of 2 revealed the formation of metallic fcc cobalt (Co0) phase (Fig. 4).39 It is to be noted that 2 is devoid of any conventional carbides of Co such as Co2C and Co3C, which may be due to lower decomposition temperature of Co2C (300 °C) and Co3C (315 °C).40 Elemental analysis of 2 prepared through five independent syntheses supports TGA results with carbon percentage between 35.07% and 35.36%, showing bulk carbonization of 2. This implies the presence of carbon along with the metallic cobalt phase. SEM image shows hierarchical near spherical carbonaceous aggregated metal clusters (Fig. 5).
Fig. 4 PXRD pattern of as synthesized 2 (top, black line) is shown against literature reported metallic fcc Co39 (bottom, red line). |
Furthermore, when a laboratory magnet was placed near a suspension of 2 in EtOH, almost quantitative collection of the content around the vicinity of the magnet was found within 2 min (Fig. S3, ESI†). This strong interaction with magnet shows that the Co-metal particles are highly magnetic and the carbon is highly embedded on Co-particles. TEM image of 2 taken from dispersion of ethanolic solution shows cobalt particles of 20–100 nm size highly embedded in carbon shell (Fig. 6a). Based on the mass contrast as shown in Fig. 6b on a single cobalt nanoparticle, the light contrast can be assigned to carbon rich materials and the dark contrast to the cobalt core. The dark domain is highly single-crystalline as evident from the SAED pattern which can be indexed to fcc cobalt while the shell around it is amorphous (Fig. S4, ESI†). The TEM micrograph shows the presence of domains of graphitic carbon (Fig. S5, ESI†) around the cobalt particles as evident from analysis of the lattice fringes with d-spacing of 0.34 nm (Fig. 6b). It is also noticeable that the thickness of the carbon shell is not uniform and varies from 5 to 8 nm. The presence of carbon coating is further supported by Raman spectroscopy of 2 which shows two signature bands at 1364 and 1587 cm−1 for D and G bands of disordered and graphitic carbon, respectively (Fig. 7).19,41 Furthermore, XPS analysis of 2 (Fig. S6, ESI†) done without Ar ion sputtering exhibited Co 2p3/2 peak at 778.3 eV characteristic of zero-valent metallic cobalt.42 However, peak at 781.3 eV along with shake-up satellite peaks for Co(II) species was also found. Since PXRD excludes the possibility of crystalline Co(II) species, the observed Co(II) indicates oxidation of reactive metallic cobalt surface to CoO, such slow oxidation in air has been observed previously for Co nanoparticles.17
Fig. 7 Raman spectra for Co@C (2) designating the D (1364 cm−1) and G (1587 cm−1) band (IG/ID = 0.37). |
The suitability of 2 as a catalyst was evaluated as cobalt particles are very well known for catalytic applications. As a model reaction, the hydrogenation of para-nitrophenol (PNP) to para-aminophenol (PAP) at room temperature was performed to test the catalytic activity of 2 (Scheme 1). PNP and other nitro phenol derivatives are common byproducts of pesticides, herbicides, and synthetic dye productions and the reaction is a strong indicator for catalytic potential.43
Preliminary reaction of PNP with 5 mol% of 2 using 20 eq. NaBH4 (2:PNP:NaBH4 = 1:20:400) in H2O at 20 °C gave quantitative conversion to PAP (Fig. S7, ESI†). For comparison, identical reaction without 2 (blank) or with 1 shows almost no or negligible PNP conversion which indicates the indispensable role of 2 for the conversion. Careful literature review shows ample use of carbon or silica supported metal particles for the catalytic conversion of PNP to PAP and almost all reports show that the support such as carbon, graphene oxide, silica or alumina augments the catalytic property of the material but themselves (support) are inactive (Table S4, ESI†).44–49 For example, the catalytic activity of carbon as in carbon nanotube with surface area as high as 176.7 m2 g−1 (Table S4, entry 1, ESI†) show extremely sluggish rate (Kapp = 0.001 min−1) for the reduction of nitrophenol to aminophenol and can be considered as almost inactive.44 Fig. 8 shows the reduction of PNP through the decrease in absorbance maxima at 400 nm due to the nitro functionality and the generation of PAP as indicated by the gradual emergence of peak intensity at 300 nm due to the amine functionality. Furthermore, kinetic measurements were performed by monitoring the decrease in absorbance at 400 nm with a molar ratio of 2:PNP:NaBH4 = 1:102:105 (catalyst = 0.375 μM).43–51 The ln(A/Ao) vs. time plotted in Fig. 8 (Inset) depicts negligible change in absorbance at 400 nm in the absence of catalyst (blank). On the other hand, upon addition of 2, a certain period of time was required for the reaction to start. Under limiting NaBH4 concentration, pseudo-first order kinetics with respect to PNP is applied to the kinetic data.43–51 The slope of near linear fit plot of the natural log of absorbance at 400 nm versus time (inset Fig. 8, red line) gives the apparent rate constant43–50 (Kapp = 0.31 ± 0.02 min−1). Under the given reaction condition, the induction period (t0) is 1.6 minutes for three independent kinetic measurements at 20 °C. It was noted that the induction period is independent of the concentration of NaBH4 and depends strongly on PNP concentration. This finding points to the fact that PNP restructuring over the active catalytic center dictates the induction period. Lowering of the induction time over higher temperature further supports substrate (PNP) induced surface restructuring necessary for the catalytic activity of Co@C composites.43 The activation energy measured between 20 and 50 °C for the catalytic process is calculated to be 24.97 kJ mol−1 (Fig. S8, ESI†) and is relatively lower than earlier reported Co/Ni nanoparticles (25.7–27.8 kJ mol−1)52,53 but relatively higher than Pt-cubes (12 kJ mol−1)54 (Table S5, ESI†).52–58 As shown in Fig. 9 the N2 adsorption–desorption isotherm of the composite Co@C (2) follow type IV isotherm with a type H3 hysteresis loop. A type IV adsorption–desorption isotherms typically indicate the presence of mesopores (pore diameter, 2–50 nm) and a type H3 hysteresis loop is often correlated with slit-shaped pores due to assemblage of non-rigid plate-like particles.59
Fig. 8 Catalytic reduction of PNP studied by UV-Vis with increment of time. Inset: plot of natural log of A/Ao vs. time showing the apparent rate constant36,37 marked by red line. |
The BET surface area of composite Co@C (2) is calculated to be 128 m2 g−1 with a pore volume of 0.4261 cm3 g−1 measured using N2 adsorption. The pore size distribution as obtained from the desorption isotherm of the composite yielded an average diameter of 4 nm (Fig. S9, ESI†) calculated using the BJH method.60 The catalytic activity of 2 can be explained by considering the reaction at the metal surface where the presence of porous yet conformal coating around cobalt nanoparticle allows easy and control access of substrates to the active metal center. This is also indirectly substantiated by the reaction of PNP in presence of non-porous parent Co(II)-polymer, 1 (SABET = 23.82 m2 g−1) as catalyst. The reaction produced negligible product (Fig. S7, ESI†) elaborating the role of the porous carbon shell around the metallic Co particles on the activity of the overall catalytic system. Although 3-D porous MOFs are well-known to play crucial role in producing uniform distribution of metal particles in a nano-porous carbonaceous matrix,7,8 in this article we showed that Co nanoparticles can be produced from 1-D non-porous coordination polymer with sufficient porosity and surface area making it a good candidate for catalyst.
Footnote |
† Electronic supplementary information (ESI) available: Tables for crystallographic details and interaction for 1; cif file for 1; IR, TGA, XPS and other details. CCDC 1436062. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c6ra04650a |
This journal is © The Royal Society of Chemistry 2016 |