Enhanced oxygen reduction activity of Pt shells on PdCu truncated octahedra with different compositions

Pd@Pt core–shell nanocrystals with ultrathin Pt layers have received great attention as active and low Pt loading catalysts for oxygen reduction reaction (ORR). However, the reduction of Pd loading without compromising the catalytic performance is also highly desired since Pd is an expensive and scarce noble-metal. Here we report the epitaxial growth of ultrathin Pt shells on Pd_xCu truncated octahedra by a seed-mediated approach. The Pd/Cu atomic ratio (x) of the truncated octahedral seeds was tuned from 2, 1 to 0.5 by varying the feeding molar ratio of Pd to Cu precursors. When used as catalysts for ORR, these three Pd_xCu@Pt core–shell truncated octahedra exhibited substantially enhanced catalytic activities compared to commercial Pt/C. Specifically, Pd₂Cu@Pt catalysts achieved the highest area-specific activity (0.46 mA cm⁻²) and mass activity (0.59 mA μg_Pt⁻¹) at 0.9 V, which were 2.7 and 4.5 times higher than those of the commercial Pt/C. In addition, these Pd_xCu@Pt core–shell catalysts showed a similar durability with the commercial Pt/C after 10 000 cycles due to the dissolution of active Cu and Pd in the cores.

Pd_xCu@Pt core–shell truncated octahedra were synthesized and exhibited substantially enhanced catalytic properties for oxygen reduction reaction relative to Pt/C.     

Highly dispersed Pt nanoclusters supported on zeolite-templated carbon for the oxygen reduction reaction

The formation of highly dispersed Pt nanoclusters supported on zeolite-templated carbon (PtNC/ZTC) by a facile electrochemical method as an electrocatalyst for the oxygen reduction reaction (ORR) is reported. The uniform micropores of ZTC serve as nanocages to stabilize the PtNCs with a sharp size distribution of 0.8–1.5 nm. The resultant PtNC/ZTC exhibits excellent catalytic activity for the ORR due to the small size of the Pt clusters and high accessibility of the active sites through the abundant micropores in ZTC.

Electrochemically synthesized highly dispersed Pt nanoclusters (PtNCs) stabilized by the nanocages of zeolite-templated carbon (ZTC) exhibit excellent electrocatalytic performance toward the oxygen reduction reaction.

Platinum (Pt) is currently considered one of the best electrocatalysts for the oxygen reduction reaction (ORR), which occurs at the cathode of a fuel cell and is the key process determining the overall performance.^1–5 However, the high cost and scarcity of Pt limit its wide commercialization in this field. According to the US Department of Energy, the total Pt loading is required to be below 0.125 mg cm⁻², in contrast to a presently used Pt loading of 0.4 mg cm⁻² or more for fuel cell application.⁴ Therefore, reducing the Pt loading without loss or with an improvement of the cathode performance has received significant interest in electrocatalytic research for fuel cell systems.^6–10 In this regard, reducing the size of Pt particles to a nanocluster scale (size < 2 nm) and maximizing the Pt dispersion may offer an efficient way to achieve maximum utilization of the Pt electrocatalyst with appropriate consumption.^4,11–15The size of nanomaterials generally plays a critical role in controlling the physical and chemical properties for catalytic applications.^16–20 With a decrease in the particle size to the nanoscale, quantum size effects are induced, which alter the surface energy of the material due to unsaturated coordination and change in the energy level of the d orbital of metal atoms, leading to spatial localization of the electrons.^17–20 This size-induced effect on the electronic structures at the active sites modifies the capability of binding the reactant molecules in catalytic reactions, thereby altering the activity of the nanocatalyst.²⁰ When the particle contains a few to several dozens of atoms with sizes, ranging from sub-nanometer to 2 nm often termed as nanocluster that bridges nanoparticle and a single atom.²¹ However, the Pt single atom is not an appropriate electrocatalyst for the ORR in a fuel cell system as the fast four-electron (4e⁻) pathway for the reduction of O₂ to H₂O requires at least two neighboring Pt atoms.^22,23 Anderson''s group demonstrated that the ratio between the production of H₂O (product of 4e⁻ process) and H₂O₂ (2e⁻) in the ORR strongly depends on the number of atoms in the Pt cluster. Typically, it requires more than 14 atoms in a Pt cluster to produce H₂O efficiently through the 4e⁻ pathway of the ORR.²⁴ Therefore, Pt nanoclusters having more than a dozen atoms have proven to be highly efficient ORR electrocatalysts for fuel cell systems.^13–15 Upon decreasing the size of the nanoparticles to a nanocluster, the electronic state and structure are known to be changed, leading to an increase of the catalytic activity in the ORR. Therefore, it is highly desirable to synthesize a Pt nanocluster-based material as an ORR electrocatalyst with high catalytic performance. To date, several synthesis strategies, such as wet-chemical, atomic-layer deposition, and photochemical methods, have been applied for the preparation of well-dispersed Pt nanoclusters on different types of support, such as dendrimer, metal oxide, and carbon materials.^{13–15,25–31}An alternative approach to synthesize Pt nanocluster (PtNC) is the encapsulation of the cluster within nanosized pores, for example, by utilizing microporous (diameters less than 2 nm) carbon materials.³² Among the microporous carbons, zeolite-templated carbon (ZTC) has been attractive for supporting Pt clusters due to its ordered microporous structure.^33–37 ZTC is a potentially promising material as catalyst support as it offers the advantages of extremely large surface area and high electrical conductivity of graphene-like carbon frameworks constituting a three-dimensional (3D) interconnected pore structure.³⁶ Moreover, the micropores of ZTC can serve as nanocages for stabilization of the Pt nanoclusters. Coker et al. used Pt²⁺ ion-exchanged zeolite as a carbon template to synthesize Pt nanoparticles in ZTC with size in a range of 1.3 to 2.0 nm.³³ Recently, atomically dispersed Pt ionic species was synthesized via a simple wet-impregnation method on ZTC containing a large amount of sulfur (17 wt%).²³ Itoi et al. synthesized PtNC consisting of 4–5 atoms and a single Pt atom in ZTC using the organoplatinum complex.³⁷ Although these methods produced Pt nanoclusters with narrow size distribution and atomic dispersion, they required multi-step processes and/or high-temperature treatment (>300 °C). High-temperature treatment often induces the sintering of nanoclusters to aggregated clusters. Therefore, it is highly desirable to develop a simple and low-cost method for the preparation of PtNC supported on ZTC (PtNC/ZTC) for use as an efficient ORR electrocatalyst. The electrochemical reduction approach offers an alternate and efficient route for the synthesis of PtNC in the micropores of ZTC. The electrochemical method is one of the popular ways to prepare electrocatalysts because it is a simple single-step procedure and ensures electrical contact between the nanoparticles and the support.^38,39Herein, we report a facile electrochemical method for the formation of PtNC with a narrow size range of 0.8–1.5 nm supported on ZTC. The resultant PtNC/ZTC shows higher electrocatalytic activities towards ORR compared to that of commercial Pt/C. Here, ZTC plays two important roles: (i) it provides nanocages to stabilize the PtNC and (ii) it accelerates the ORR activity by enhancing the accessibility of active sites through its abundant micropores. Fig. 1a shows a schematic representation of the typical electrochemical synthesis of PtNC/ZTC. In the first step, ZTC was impregnated with a Pt-precursor dissolved in a water–ethanol mixture. As ZTC possess ordered micropores (Fig. S1a^†) with high Brunauer–Emmett–Teller (BET) surface area of 3400 m² g⁻¹ (vide infra), the uniform adsorption and anchorage of PtCl₆²⁻ ions into the micropores of ZTC was favored. After impregnating and drying, the resultant ZTC–PtCl₆²⁻ was mixed with water–ethanol and Nafion to make the ink for the preparation of the electrode. Using the prepared electrode, a potential of 0.77 V vs. reversible hydrogen electrode (RHE) (Fig. 1b) was applied followed by potential cycling between 1.12 to −0.02 V vs. RHE until the cyclic voltammogram was stabilized. The Pt content of PtNC/ZTC was determined to be ∼10 wt% (Fig. S2^†) by thermogravimetric analysis (TGA). The obtained PtNC/ZTC was electrochemically characterized by cyclic voltammetry and electrochemical impedance spectroscopy. The cyclic voltammogram (Fig. 1c) after potential cycling in fresh KOH electrolyte shows the characteristic Pt peaks corresponding to hydrogen adsorption and desorption. The Nyquist plots (Fig. 1d) demonstrate that PtNC/ZTC has lower electrolyte resistance (42 Ω) than that of ZTC (70 Ω), implying an improvement in the conductivity of ZTC by the presence of PtNC. Due to the increase in the conductivity, PtNC/ZTC could facilitate the electron transfer more effectively than ZTC, enhancing its electrocatalytic activity.Open in a separate windowFig. 1(a) Illustration for the formation of PtNC/ZTC:Pt-precursor was impregnated into ZTC micropores, and then a potential (0.77 V vs. RHE) was exerted on the ZTC–PtCl₆²⁻ composite in a 0.1 M KOH solution to form PtNC/ZTC (b) Chronoamperometric response of ZTC–PtCl₆²⁻ at a constant potential of 0.77 V (vs. RHE) in 0.1 M KOH electrolyte. (c) Cyclic voltammogram of PtNC/ZTC in a fresh 0.1 M KOH at a scan rate of 20 mV s⁻¹. (d) Nyquist plots of ZTC and PtNC/ZTC in 0.1 M KOH. Fig. 2a and b show images from aberration-corrected scanning transmission electron microscope (STEM) with high-angle annular dark-field (HAADF). The HAADF-STEM images exhibit the typical morphology of the final product (PtNC/ZTC) after electrochemical reduction. As shown in Fig. 2a, it is very clear that isolated PtNCs are uniformly dispersed in ZTC. These PtNCs have a homogeneous distribution with a narrow size range (0.8–1.5 nm, Fig. 2b). On further magnification, the STEM image shows a cluster-like structure of Pt (Fig. 2c). The STEM image of selected PtNC (Fig. 2d) reveals that it consists of ∼20 atoms. The number of atom content in PtNC was further determined by matrix-assisted laser-desorption-ionization time-of-flight (MALDI-TOF) mass spectrometry using trans-2-[3-(4-test-butylphenyl)-2-methyl-2-propenylidene]malononitrile as the matrix.^40,41 As shown in Fig. S3,^† MALDI-TOF measurement produces a mass spectra with a predominant peak centered at ∼3700 Da corresponding to the Pt₁₉ cluster. The TEM image (Fig. S4 a and b^†) validates the formation of PtNC with an average size of 0.9 nm. In addition, the energy dispersive X-ray spectrometer (EDS) mapping images clearly shows the uniform dispersion of Pt nanocluster in ZTC (Fig. S4c^†). The X-ray powder diffraction (XRD) pattern (Fig. 2e) of PtNC/ZTC showed three broad peaks associated with small size metallic Pt corresponds to (111), (200), and (311) planes (Fig. 2e, inset), along with peaks of ZTC at 2θ = 7.8° and 14.9° corresponding to the ordered microporous structure. Along with the structural analysis, the porous texture of PtNC/ZTC was examined by Ar adsorption (Fig. 2f). PtNC/ZTC had a high BET surface area of 2360 m² g_ZTC⁻¹, which is 1.4 times lower than that of pristine ZTC (3400 m² g_ZTC⁻¹). The decrease in Ar adsorption capacity after the formation of PtNC in ZTC is interpreted as a result of the filling of ZTC micropores by PtNC. This micropore filling was confirmed in the pore size distributions of the pristine ZTC and the metal-loaded carbon (inset of Fig. 2f). The X-ray photoelectron spectroscopy (XPS) results reveal the signature of Pt in ZTC (Fig. S5^†). The elemental survey (Fig. S5a^†) shows the signature of C 1s, O 1s, F 1s (Nafion), and Pt 4f. The chemical nature of Pt in PtNC/ZTC was inspected by a detailed Pt 4f XPS analysis. The deconvoluted Pt 4f XPS spectra (Fig. S5b^†) reveals the presence of both metallic and ionic Pt species. The peaks observed at 71.0 (4f_7/2) and 74.2 (4f_5/2) eV correspond to metallic Pt whereas the other peaks positioned at 72.6 (4f_7/2) and 76.0 (4f_5/2) are attributed to Pt²⁺ and the peaks at 74.9 (4f_7/2) and 77.8 (4f_5/2) eV are attributed to Pt⁴⁺ originating from the surface oxidation of metallic Pt.⁴²Open in a separate windowFig. 2(a–d) Representative spherical aberration-corrected HAADF-STEM images of PtNC/ZTC at various magnifications. (e) XRD pattern of PtNC/ZTC and (f) Ar adsorption–desorption isotherms of ZTC and PtNC/ZTC. Inset in (e) shows a 30 times magnified high-angle region of XRD of PtNC/ZTC. Inset in (f) shows the pore size distributions of the ZTC and PtNC/ZTC.The formation of narrow sized PtNC by the electrochemical method can be ascribed to the stabilization of PtNC in the ZTC micropores, which serve as cages to impose a spatial limitation on the size of the Pt clusters. For comparison, Pt supported on ZTC was also prepared by the conventional incipient wetness impregnation and subsequent H₂-reduction at high temperature (300 °C). The Pt obtained by this incipient wetness impregnation method shows the formation of Pt nanoparticles on the exterior surface of ZTC (PtNP/ZTC) (Fig. S6^†). The formation of larger Pt nanoparticles is due to the sintering at high temperature, showing that even ZTC micropores could not prevent the aggregation of PtNCs at high temperatures. Fig. 3 shows the electrochemical ORR activity of PtNC/ZTC using linear sweep voltammetry (LSV) technique on a rotating disc electrode (RDE) in a 0.1 M KOH solution saturated with O₂ at a scan rate of 5 mV s⁻¹. The ORR activity of ZTC (without PtNC) was measured for comparison as well. As shown in Fig. 3a, PtNC/ZTC exhibited higher diffusion limiting current density and higher positive onset and half-wave potential compared to ZTC alone, indicating that PtNC is the active center for the ORR. To investigate the effect of the Pt loading amount on the ORR activity, PtNC/ZTC with various Pt loadings, 2–20 wt%, was used for the measurement of LSV at 1600 rpm. With an increase in Pt content, both the onset and half-wave potential shifted towards more positive potential up to 10 wt% loading of Pt (Fig. 3a and S7^†). Upon further increase of loading of Pt on ZTC to 20 wt%, both the onset and half-wave potential of PtNC/ZTC shifted towards less positive potential along with a slight decrease in the diffusion limiting current density (Fig. 3a). The decrease in the ORR activity of PtNC/ZTC at high loading of Pt (20 wt%) was attributed to the decrease in the electrochemically active surface area (Fig. S8^†) and decrease in the specific surface area (Fig. S9^†). The STEM image clearly shows that the aggregated Pt clusters were formed on the exterior surface of ZTC at 20 wt% loading of Pt (Fig. S10c^†), blocking the accessibility of active sites. Therefore, PtNC/ZTC with the optimum loading of 10 wt% of Pt leads to superior ORR activity with a high positive onset potential of 0.99 V, which is similar to commercial Tanaka Pt/C (Pt/C-TKK) (Fig. 3b), and a half-wave potential of 0.87 V, which is ∼10 mV more positive than that of commercial Pt/C-TKK (0.86 V) (Fig. 3b). Compared to the case of PtNC/ZTC, both the onset and half-wave potential of PtNP/ZTC prepared by the conventional incipient wetness impregnation and subsequent H₂-reduction with the same loading of Pt exhibited a less positive value (Fig. S11^†). The poorer activity of PtNP/ZTC is due to the blockage of active sites by larger PtNPs formed on the exterior surface of ZTC (Fig. S6^†).Open in a separate windowFig. 3(a) RDE ORR polarization curves of PtNC/ZTC with different mass loading of Pt. (b) Comparison of PtNC/ZTC (PtNC_10%/ZTC) with commercial Pt/C-TKK at the same loading of 40 μg_Pt cm⁻². (c) RDE ORR polarization curves of PtNC/ZTC at different rotation speeds. Inset in (c) shows the corresponding K–L plots at different potentials. (d) Represents the kinetic current density values of Pt/C-TKK and PtNC/ZTC at the potential of 0.8 V vs. RHE.To investigate the kinetics of the ORR activity of PtNC/ZTC, LSV measurements were performed with RDE at different rotating rates (Fig. 3c), and the kinetics was analyzed using a Koutecký–Levich (K–L) plot (Fig. 3c, inset). From Fig. 3c, it was observed that the current density increases with the increasing speed of rotation of the electrode, which is characteristic of a diffusion-controlled reaction. The corresponding linear K–L plots (Fig. 3c, inset) with a similar slope at different potentials reveal that the number of transferred electrons was ∼4, indicating that O₂ is directly reduced to OH⁻ and the ORR is dominated by the H₂O₂-free 4e⁻ pathway. To estimate the amount of produced peroxide ion, rotating ring-disc electrode (RRDE) measurement was performed and the produce peroxide ion calculated from RRDE curve was < 4% (Fig. S12^†). The kinetic current density (J_k) obtained from K–L plot at the potential of 0.8 V (Fig. 3d) for PtNC/ZTC (J_k = 50 mA cm⁻²) is 2.2 times higher than that of commercial Pt/C-TKK (J_k = 22 mA cm⁻²).As Pt-based electrocatalysts are known to be highly active in an acidic medium, the ORR activity of PtNC/ZTC in O₂-saturated 0.1 M HClO₄ was also evaluated by comparing it with that of commercial Pt/C-TKK with the same loading of Pt on the electrode surface using RDE at a scan rate of 5 mV s⁻¹. The PtNC/ZTC-based electrode exhibited ORR activity with an onset potential of 0.96 V (Fig. 4), which is close to that of Pt/C-TKK (0.98 V), and half-wave potentials of 0.84 V, which is 20 mV more positive than that of Pt/C-TKK (0.82 V). PtNC/ZTC showed a slightly higher diffusion-limiting current density of ∼5.9 mA cm⁻² (0.4–0.7 V) compared with that of the Pt/C-TKK catalyst (∼5.6 mA cm⁻²). The kinetics of the ORR in an acidic medium was further analyzed using RDE at different rotation rates (Fig. S13^†) and it was observed that the current density increases with the increasing speed of rotation of the electrode, as in the case of the alkaline medium. The number of electron involved and the amount of produced H₂O₂ estimated by RRDE measurement were ∼4 and < 5%, respectively (Fig. S14^†). The mass activity of PtNC/ZTC obtained using the mass transport corrected kinetic current at 0.8 V is 0.15 A mg⁻¹, which is 3.2 times higher than that of Pt/C-TKK (0.046 A mg⁻¹).Open in a separate windowFig. 4(a) RDE ORR polarization curves at 1600 rpm and (b) mass activity at 0.8 V of PtNC/ZTC and Pt/C-TKK in 0.1 M HClO₄.Furthermore, the methanol tolerance of PtNC/ZTC was assessed by intentionally adding methanol to the oxygen saturated electrolyte solution (both in alkaline and acidic media). The commercial Pt/C-TKK was used for comparison as well. The peak current densities for methanol oxidation with PtNC/ZTC were ∼2.8 and ∼3 times lower than that of Pt/C-TKK in alkaline (Fig. 5a) and acidic (Fig. 5b) media, respectively. These results indicate that PtNC/ZTC has much higher tolerance towards methanol than Pt/C-TKK does. This higher methanol tolerance of PtNC/ZTC can be attributed to the small size of the Pt cluster, which may not be sufficient to catalyze the oxidation of methanol efficiently, as the oxidation of methanol requires Pt ensemble sites.⁴³Open in a separate windowFig. 5ORR polarization curves of PtNC/ZTC and Pt/C-TKK in the absence (solid line) and presence (dotted line) of 0.1 M of CH₃OH at a rotation rate of 1600 rpm in (a) alkaline and (b) acid media.The durability of PtNC/ZTC was also investigated by the amperometric technique. The test was performed at a constant voltage of the half-wave potential in an O₂-saturated alkaline medium and at 0.7 V in an O₂-saturated acidic medium at a rotation rate of 1600 rpm (Fig. S15a and b^†). The durability of the PtNC/ZTC catalyst in the alkaline medium was higher than that of Pt/C-TKK, exhibiting a 30% decrease compared to a 40% decrease of Pt/C-TKK in 5.5 h of ORR operation (Fig. S15a^†). The higher durability of PtNC/ZTC compared to Pt/C-TKK in the alkaline medium may be due to the stabilization of PtNC by pore entrapment. In the acidic medium, however, PtNC/ZTC exhibited a 54% decrease in the initial current after 5.5 h of operation while a 33% decrease was observed in the case of Pt/C-TKK (Fig. S15b^†). The decrease in ORR activity in the acidic medium may be due to the leaching out of tiny Pt nanoclusters in acid electrolyte from the ZTC micropores. To understand the decrease in the ORR durability with time, STEM measurements of PtNC/ZTC after 5.5 h of ORR operation were performed. In the alkaline medium, the STEM image of post-ORR PtNC/ZTC shows a slight change in the size of PtNC (Fig. S15c^†) while the STEM image of PtNC/ZTC after ORR in the acidic medium exhibited sintering of PtNC into large particles with an average size of 30 nm (Fig. S15d^†), resulting in a decrease of the ORR activity. In the alkaline medium, the decrease in ORR activity with time may be due to the oxidation of the ZTC support in KOH.⁴⁴We attributed the excellent ORR activity of PtNC/ZTC to the interplay between the following: (1) the structure of the Pt cluster possessing a high ratio of surface atoms that benefits the surface reactions,^45–47 (2) the microporous 3D graphene-like structure of the ZTC support that enables easy access of O₂ and electrolyte molecules to the active sites,⁴⁸ and (3) the high conductivity and large accessible surface area of ZTC that facilitates the electron transfer.^49–51     

Size effect studies in catalysis: a simple surfactant-free synthesis of sub 3 nm Pd nanocatalysts supported on carbon

Supported Pd nanoparticles are prepared under ambient conditions via a surfactant-free synthesis. Pd(NO₃)₂ is reduced in the presence of a carbon support in alkaline methanol to obtain sub 3 nm nanoparticles. The preparation method is relevant to the study of size effects in catalytic reactions like ethanol electro-oxidation.

A simple surfactant-free synthesis of sub 3 nm carbon-supported Pd nanocatalysts is introduced to study size effects in catalysis.

A key achievement in the design of catalytic materials is to optimise the use of resources. This can be done by designing nanomaterials with high surface area due to their nanometre scale. A second achievement is to control and improve catalytic activity, stability and selectivity. These properties are also strongly influenced by size.^1–3 To investigate ‘size effects’ it is then important to develop synthesis routes that ensure well-defined particle size distribution, especially towards smaller sizes (1–10 nm).Metal nanoparticles are widely studied catalysts. In several wet chemical syntheses, NP size can be controlled using surfactants. These additives are, however, undesirable for many applications^4,5 since they can block active sites and impair the catalytic activity. They need to be removed in ‘activation’ steps which can negatively alter the physical and catalytic properties of the as-produced NPs. Surfactant-free syntheses are well suited to design catalysts with optimal catalytic activity⁶ but their widespread use is limited by a challenging size control.³Palladium (Pd) NPs are important catalysts for a range of chemical transformations like selective hydrogenation reactions and energy applications.^7–9 It is however challenging to obtain sub 3 nm Pd NPs, in particular without using surfactants.² Surfactant-free syntheses are nevertheless attracting a growing interest due to the need for catalysts with higher performances.^10–14Promising surfactant-free syntheses of Pd NPs were recently reported.^8,15 The NPs obtained in these approaches are in the size range of 1–2 nm and show enhanced activity for acetylene hydrogenation⁸ and dehydrogenation of formic acid.¹⁵ Enhanced properties are attributed to the absence of capping agents leading to readily active Pd NPs. The reported syntheses consist in mixing palladium acetate, Pd(OAc)₂, in methanol and the reduction of the metal complex to NPs occurs at room temperature. The synthesis is better controlled in anhydrous conditions to achieve a fast reaction in ca. 1 hour. Another drawback is that the synthesis must be stopped to avoid overgrowth of the particles. Therefore, a support material needs to be added after the synthesis has been initiated and no simple control over the NP size is achieved.^8,15In this communication a more straightforward surfactant-free synthesis leading to sub 3 nm carbon-supported Pd NPs in alkaline methanol at ambient conditions is presented. A solution of Pd(OAc)₂ in methanol undergoes a colour change from orange to dark, indicative of a reduction to metallic Pd, after ca. a day. However, only ca. 1 hour is needed with Pd(NO₃)₂, Fig. 1 and UV-vis data in Fig. S1.^† The fast reduction of the Pd(NO₃)₂ complex in non-anhydrous conditions is a first benefit of the synthesis presented as compared to previous approaches.Open in a separate windowFig. 1Pictures of 4 mM Pd metal complexes in methanol without or with a base (as indicated).For particle suspensions prepared with Pd(OAc)₂ or Pd(NO₃)₂ the NPs agglomerate and quickly sediment leading to large ‘flake-like’ materials. When the reduction of Pd(NO₃)₂ in methanol is performed in presence of a carbon support and after reduction the solution is centrifuged and washed in methanol, a clear supernatant is observed indicating that no significant amount of NPs are left in methanol. Transmission electron microscope (TEM) analysis confirms that NPs are formed and well-dispersed on the carbon support surface and no unsupported NPs are observed, Fig. 2a. Likely, the reduction of the NPs proceeds directly on the carbon support. However, the size of the NPs is in the range 5–25 nm, which is still a relatively large particle size and broad size distribution.Open in a separate windowFig. 2TEM micrographs of Pd NPs obtained by stirring 4 mM Pd(NO₃)₂ in methanol and a carbon support for 3 hours, (a) without NaOH and (b) with 20 mM NaOH. Size distribution histograms are reported in Fig. S4.^† The same samples after electrochemical treatments are characterised in (c) and (d) respectively. Size distribution histograms are reported in Fig. S7.^†Assuming a ‘nucleation and growth’ mechanism, the NPs should become larger over time.¹⁶ But the reaction is so fast that by stopping the reaction before completion, size control is not achieved and unreacted precious metal is observed, Fig. S2.^† To achieve a finer size control and more efficient use of the Pd resources, a base was added to the reaction mixture, e.g. NaOH.³ In alkaline media, the formation of Pd NPs is slower; it takes ca. 60 minutes to observe a dark colour for a 5 mM Pd(NO₃)₂ solution with a base/Pd molar ratio of 10 in absence of a support, Fig. 1.Also in alkaline methanol, the NPs agglomerate over time in absence of a support material. However, if the alkaline solution of Pd(NO₃)₂ is left to stir in presence of a carbon support the desired result is achieved, i.e. Pd NPs with a significantly smaller size and size distribution of ca. 2.5 ± 1.0 nm, Fig. 2b. The NPs homogeneously cover the carbon support and no unsupported NPs are observed by TEM suggesting that the NPs nucleate directly on the carbon surface. Furthermore, the supernatant after centrifugation is clear, indicating an efficient conversion of the Pd(NO₃)₂ complex to NPs, Fig. S3.^† Furthermore, there is no need for an extra reducing agent as in other approaches, for instance in alkaline aqueous solutions.⁹The benefits of surfactant-free syntheses of Pd NPs for achieving improved catalytic activity have been demonstrated for heterogeneous catalysis.^8,15 Surfactant-free syntheses are also well suited for electrochemical applications where fully accessible surfaces are required for fast and efficient electron transfer. Several reactions for energy conversion benefit from Pd NPs. An example is the electro-oxidation of alcohols,⁷ in particular ethanol¹⁷ (see also Table S1^†).Previous studies optimised the activity of Pd electrocatalysts by alloying,^18–20 by using different supports^17,21–23 or crystal structures.^24,25 Investigating NPs with a diameter less than 3 nm was challenging.^2,26,27 The surfactant-free synthesis method presented here allows to further study the size effect on Pd NPs supported on carbon (Pd/C) for electrocatalytic reactions.In Fig. 3, results for ethanol electro-oxidation in 1 M ethanol solution mixed with 1 M KOH aqueous electrolyte are reported based on cyclic voltammetry (CV) and chronoamperometry (CA) with Pd/C catalysts exhibiting 2 significantly different size distributions. The electrode preparation, the measurement procedure and the sequence of electrochemical treatments are detailed in the ESI.^† In order to highlight size effects, we compare geometric and Pd mass normalized currents (Fig. 3a and c) as well as the oxidation currents normalized to the Pd surface area (Fig. 3b).Open in a separate windowFig. 3Electrochemical characterisation of carbon supported Pd NPs with 5–25 nm (grey) and 2.5 nm (dark) size in 1 M KOH + 1 M ethanol aqueous electrolyte. (a) 2^nd CV before chronoamperometry (CA), (b) current normalised by the electrochemically active surface area of Pd, (c) CA recorded at 0.71 V vs. RHE after 50 cycles between 0.27 and 1.27 V.It is clearly seen that based on the geometric current density, the smaller Pd NPs exhibit significantly higher currents for ethanol oxidation than the larger NPs. To differentiate if this observation is a sole consequence of the different surface area, the electrochemically active surface area (ECSA) has been estimated based on “blank” CVs (without ethanol) recorded between 0.27 and 1.27 V vs. RHE in pure 1 M KOH aqueous electrolyte and integrating the area of the reduction peak at ca. 0.68 V, Fig. S5.^† As conversion factor, 424 μC cm⁻² was used.²⁸Using this method, the smaller NPs with a size around 2.5 nm exhibit an ECSA of 92 m² g⁻¹ whereas the larger NPs with a size in the range 5–25 nm exhibit an ECSA of 47 m² g⁻¹, consistent with a larger size. Normalising the ethanol electro-oxidation to these ECSA values instead of the geometric surface area, Fig. 3b, still indicates a size effect. It is clearly seen that the smaller Pd NPs exhibit higher surface specific ethanol oxidation currents, in particular at low electrode potentials. Furthermore, a clear difference in the peak ratios in the CVs is observed. The ratio in current density of the forward anodic peak (j_f, around 0.9 V) and the backward cathodic peak (j_b, around 0.7 V vs. SCE) is around one for the smaller NPs, whereas it is about 0.5 for the larger NPs. The forward scan corresponds to the oxidation of chemisorbed species from ethanol adsorption. The backward scan is related to the removal of carbonaceous species not fully oxidised in the forward scan. The higher j_f/j_b ratio therefore confirms that the smaller NPs are more active for ethanol electro-oxidation and less prone to poisoning, e.g. by formation of carbonaceous species that accumulate on the catalyst surface.^29,30 This observation is further supported by a chronoamperometry (CA) experiment, Fig. 3c, at 0.71 V performed after continuous cycling (50 cycles between 0.27 and 1.27 V at a scan rate of 50 mV s⁻¹). In the CA testing of the thus aged catalysts at 0.71 V, the ethanol oxidation current on the two catalysts starts at around the same values, however, its decay rate is significantly different. The Pd mass related oxidation currents for the smaller NPs are after 30 minutes almost twice as high (ca. 200 A g_Pd⁻¹) as for the larger ones (ca. 130 A g_Pd⁻¹), confirming that the small Pd NPs are less prone to poisoning. In particular a factor up to 4 in the Pd mass related ethanol oxidation currents after 1800 s of continuous operation is achieved compared to a recently characterised commercial Pd catalyst on carbon,²⁰ Table S1.^† Despite different testing procedure reported in the literature, it can be concluded from these investigations that the surfactant-free synthesis presented shows promising properties for electrocatalytic ethanol oxidation even after extended cycling.The extended cycling, however, has different consequences for the two catalysts. For the small (2.5 nm) NPs of the Pd/C catalyst, a massive particle loss, but only moderate particle growth is observed as highlighted in Fig. 2 (see also Fig. S6^†). TEM micrographs of the two Pd/C samples recorded before and after the complete testing (CVs and CAs, for details see Fig. S7^†) show that the catalyst with small Pd NPs exhibits a pronounced particle loss as well as a particle growth to ca. 6 nm probably due to sintering. By comparison, for the Pd/C catalyst with the large Pd NPs, no significant influence of the testing on particle size or particle density is apparent.     

Highly efficient hydrogen peroxide direct synthesis over a hierarchical TS-1 encapsulated subnano Pd/PdO hybrid

We report a hierarchical TS-1 encapsulated subnano Pd/PdO hybrid catalyst that shows unprecedented activity in H₂O₂ direct synthesis from H₂ and O₂. The macro reaction rate in 30 min is up to 35 010 mmol g_Pd⁻¹ h⁻¹ at ambient temperature. Such high catalytic activity is achieved due to the hierarchical porous structure of TS-1 and the formation of the encapsulated subnano Pd/PdO hybrid after oxidation/reduction/oxidation treatment.

A hierarchical TS-1 encapsulated subnano Pd/PdO hybrid catalyst that shows unprecedented activity in H₂O₂ direct synthesis from H₂ and O₂.

Hydrogen peroxide as a clean and strong oxidant is one of the commonly used chemicals in various fields of chemical industry, such as the pulp and paper industry, the textile industry, wastewater treatment, green chemical synthesis metallurgy, electronics manufacture, propulsion and the food industry.¹ Compared to the traditional anthraquinone process (sequential hydrogenation and oxidation of alkyl anthraquinone), the direct synthesis of hydrogen peroxide (DSHP) from hydrogen and oxygen was recognized as an efficient and environmental alternative process owing to its remarkable adherence to green chemistry perspectives, such as low energy consumption, minimized toxicity and infrastructure investment.^2–5Pd supported catalysts were the most extensively and earliest studied catalysts for the DSHP since 1914.⁶ Both DFT and experimental results indicated that subnano Pd particles were most effective for the selective oxygen hydrogenation to hydrogen peroxide,⁷ and the activity and selectivity are also highly dependent upon the oxidation state of the Pd particles.⁸ However, there were limitations in applying Pd nanoparticles catalyst to the reaction due to the thermal vulnerability in a calcination and reduction activation process.⁹ To solve this problem, many preparation methods have been adopted to stabilise Pd nanoparticles and control the particle size and morphology, such as yolk–shell structure,¹⁰ core–shell structure¹¹ and other encapsulation structure supports. But there were still problems that the size of metal particles is larger than 2.5 nm. Encapsulation of Pd species by mercaptosilane-assisted dry gel conversion (DGC) synthesis method can provide a precise control over the nanoparticle size as well as limitating the aggregation under high temperature during activation.¹² However, active sites deep inside the encapsulated nanoparticles were often hardly accessible since the internal configuration diffusion limitations of reactants and products in micropores, leading to low H₂ conversion and decomposition of the long residence time of synthetic H₂O₂.¹³ So, the role of the porous structured catalyst was essential for encapsulated metal nanoparticles.Titanium silicalite-1 (TS-1) has already been used as an excellent catalyst for a variety of selective oxidation reactions employing hydrogen peroxide as oxidant.^14,15 Moreover, in situ H₂O₂ generation coupled with these selective oxidation reactions leading to the desired products such as propylene,^16,17 benzyl alcohol,¹⁸ cyclohexene¹⁹ was a desirable, green and lower cost route. More importantly, the Ti–OOH species formed on the TS-1 during selective oxidation might improve the stability of OOH, which is a key reaction intermediate during the DSHP.²⁰ Hutchings et al. reported that hierarchical titanium silicalite supported Au–Pd catalysts showed high peroxide production rate and benzaldehyde production rate for oxidation of benzyl alcohol by in situ generated H₂O₂.²¹ In this report, the encapsulation of subnano-sized Pd metal particles within conventional (Pd@TS-1) and hierarchical titanium silicalite-1 (Pd@HTS-1) has been achieved (see Scheme 1). The Pd@HTS-1 catalyst after oxidation–reduction–oxidation pre-treatment showed unprecedented activity in direct synthesis of hydrogen peroxide from hydrogen and oxygen under ambient temperature without any promoter.Open in a separate windowScheme 1Schematic diagram of the preparation method for Pd@HTS-1.The TS-1 and HTS-1 encapsulated Pd sub-nanoparticles were first synthesized via solvent evaporation-assisted dry gel conversion method, where the Pd was encapsulated in situ through hydrothermal crystallization in assistance of 3-mercaptopropyl-trimethoxysilane (Scheme 1). The results of ICP analysis confirmed that total Pd contents in Pd@TS-1 and Pd@HTS-1 were 0.094 and 0.106 wt%, respectively. The characteristic diffraction “finger peak” on the X-ray diffraction in Fig. S1^† proved that the TS-1, Pd@TS-1 and Pd@HTS-1 had a well-crystallized MFI structure,²² which was further confirmed by the asymmetric stretching of Si–O–Ti in the spectra of Fourier Transform Infrared Spectroscopy (FT-IR, see Fig. S2^†). For all of the samples, the diffraction peak at 2θ of 25.4° was not observed. Meanwhile, the diffraction peak of crystalline Pd was also not detected for Pd@HTS-1 and Pd@TS-1, indicating that the Pd particles were well dispersed in the zeolite.⁷ Besides, the diffuse reflectance UV-vis spectra of the TS-1, Pd@TS-1 and Pd@HTS-1 were shown in Fig. S3.^† The band at 210 nm in three samples confirmed the tetrahedral structural geometry of Ti in these silicates, and the weak band at 280 nm was assigned to small amounts of penta/hexacoordinated Ti species.²³ Moreover, the absorption band around 300 nm indicated that the three samples contain anatase TiO₂.²⁴The textural properties of the synthesized Pd@TS-1 and Pd@HTS-1 were characterized by N₂ adsorption/desorption and the results were shown in Fig. 1 and Table S1.^† Notably, typical irreversible type IV adsorption isotherms with an H1 hysteresis loop were observed over the Pd@HTS-1 sample (Fig. 1b), indicating the presence of a mesoporous structure. The mesopore size of Pd@HTS-1, obtained through the BJH method, and the obtained graph peaked at about 7.0 nm. Volume of the micropores was around 0.14 cm³ g⁻¹ for both Pd@TS-1 and Pd@HTS-1, but the surface area of Pd@HTS-1 (509.9 m² g⁻¹) was 48.9 m² g⁻¹ larger than that of Pd@TS-1 (461.0 m² g⁻¹) due to its mesoporous structure, which is beneficial for the diffusion of reactants and products through the catalysts.²⁵Open in a separate windowFig. 1Nitrogen adsorption–desorption isotherms of the synthesized TS-1: (a) Pd@TS-1 and (b) Pd@HTS-1.Comparison between the experimentally obtained results from ammonia temperature-programmed desorption (NH₃-TPD) analysis (Fig. S4^†) and the previously reported data showed that the peaks observed were corresponding to weak acid sites, medium acid sites, and strong acid sites of the catalysts.²⁶ Furthermore, pyridine adsorption peak on the FT-IR spectra of these samples (Fig. S5^†) revealed that titanium silicate (TS-1) was an acidic support with a large number of Lewis acid segments and few Brønsted acid segments. As shown in scanning electron microscopy (SEM) image (Fig. 2), Pd@TS-1 particles were crystallites with a morphology close to cuboids and a mean particle size of about 3–5 μm, while the Pd@HTS-1 has spherical morphology with a particle size of about 1.3 μm.Open in a separate windowFig. 2SEM images of the synthesized Pd-modified TS-1: (a) Pd@TS-1 and (b) Pd@HTS-1.The synthesized TS-1 and HTS-1 encapsulated Pd sub-nanoparticles were then subjected to oxidation/reduction/oxidation treatment to adjust the valence states of Pd.²⁷ Such heat treatment cycle can switch off the sequential hydrogenation and decomposition reactions in the DSHP. However, Ostwald ripening, thus the migration and coalescence of metal clusters, will occur at a higher temperature. Therefore, high temperature treatments was used to emulate the conditions used in the literature mentioned before,^28,29 and the thermal stability of the encapsulated Pd@TS-1 catalysts before and after the treatments were also evaluated and compared to investigate the effect of high temperature and the thermal treatments on the catalysts. The Pd@TS-1 and Pd@HTS-1 samples after an air/H₂/air thermal treatments at 500/400/500 °C for 4/2/6 h were denoted as Pd@TS-1-O, Pd@TS-1-OR, Pd@TS-1-ORO, Pd@HTS-1-O, Pd@HTS-1-OR, Pd@HTS-1-ORO respectively with O denoting oxidation and R denoting reduction. The Pd particle size distribution after such treatments was first released by the high-resolution transmission electron microscopy (HRTEM) image in Fig. 3 and S6.^† The Pd particles encapsulated within microporous TS-1 zeolites were well dispersed and uniformly distributed throughout the zeolite crystals. The average sizes of Pd particles encapsulated in the TS-1 and HTS-1 were in the range of 1–2 nm, which, however, was bigger than those of the MFI topology channels (0.53 × 0.56 nm) and intersectional channels (∼0.9 nm). Nevertheless, the successful encapsulation of the Pd particles in the TS-1 zeolites was verified by comparing the hydrogenation rates of a mixture of nitrobenzene and 1-nitronaphthalene. As shown in Fig. S7,^† the reaction rate for the hydrogenation of nitrobenzene and 1-nitronaphthalene was much higher over the Pd@HTS-1-OR compared to the Pd@TS-1-OR. We anticipated that the slightly larger Pd size than the zeolite channels might reflect the local disruption of the crystal structures near the location of the particles during the in situ synthesis. More detailed size distributions of Pd particles encapsulated in the TS-1 and HTS-1 zeolites after air, Ar/H₂ and air treatments were shown in Fig. 3d–f and j–l, respectively. The particle sizes of most of the Pd species still remain below 2 nm on average, which indicated the absence of metal clusters migration and coalescence by Ostwald ripening even after such higher temperature treatments. The high thermal stability of the Pd subnano particles resulted from the embedding confinement.³⁰Open in a separate windowFig. 3HRTEM images and metal particle size distributions of the Pd@TS-1 and Pd@HTS-1 before and after high-temperature oxidation–reduction–oxidation treatments. (a, d Pd@TS-1-O. b, e Pd@TS-1-OR. c, f Pd@TS-1-ORO. g, j Pd@HTS-1-O. h, k Pd@HTS-1-OR. i, l Pd@HTS-1-ORO.)The Pd dispersion and average Pd nanoparticle size for Pd@TS-1 and Pd@HTS-1 after the air/H₂ treatment were further determined by CO chemisorption measurements (see Table S2^†). The dispersions of Pd in Pd@TS-1 and Pd@HTS-1 are 85% and 81%, respectively. The average Pd particle sizes for Pd@TS-1 and Pd@HTS-1 calculated by CO adsorption measurements are 1.06 nm and 1.17 nm, respectively, which was smaller than that estimated from the TEM analysis. This was probably due to the presence of Pd nanocluster or single atoms, which cannot be directly observed by HRTEM.We now turn to the Pd valence states of the catalysts after the oxidation/reduction/oxidation treatment by the XPS (see Fig. S8^†). The Pd3d spectra signals were hardly observed when the concentration of Pd atoms was low, the binding energy peaks for different oxidation states of Pd atoms were collected after peak fitting by prolonging the scanning time.³¹ The XPS results demonstrated the presence of both metallic Pd and PdO. The binding energy of peaks for Pd⁰3d_5/2 and Pd⁰3d_3/2 correspond to 335.5 and 340.6 eV, respectively, while the binding energy for Pd²⁺3d_5/2 and Pd²⁺3d_3/2 were at 337.8 and 341.9 eV, respectively.³¹ The transformation of valence state could be observed in Fig. S8a–c,^† which was derived from XPS measurements. Moreover, the ratios for Pd⁰ and Pd²⁺ atoms in Pd@TS-1 and Pd@HTS-1 were approximately 2 and 1, respectively. On the basis of these results, we proposed a reaction mechanism for the synthetic process of the catalysts, subnano-sized Pd particles might be oxidated from Pd⁰ to Pd²⁺ to form PdO on the surface of the catalysts during reoxidation.The catalytic performance of the TS-1 and HTS-1 encapsulated subnano-sized Pd/PdO hybrid in the direct synthesis of hydrogen peroxide from H₂ and O₂ were tested at ambient temperature without any promoters. Compared to the Pd supported by the active carbon, the selectivity of hydrogen peroxide was higher, the reason might be the formation of Ti–OOH³² and the confinement effect of the Pd encapsulated in the channel of the zeolite (Scheme 2). Both HTS-1 zeolite and Pd@zeolites showed significant amount of O₂ adsorption according to the O₂-TPD (Fig. S9^†), which might be the reason for high activity/selectivity. The selectivity for hydrogen peroxide on Pd@TS-1-OR is lower than that on Pd@TS-1-O, while the degradation rate of hydrogen peroxide on Pd@TS-1-OR are higher than that on Pd@TS-1-O (Fig. 4 and Table S3^†), which was attributed to the change in oxidation state from Pd²⁺ to Pd⁰ after reductive treatment, in agreement with previous reports.^27,33 The selectivity of hydrogen peroxide over Pd@TS-1 increased after an oxidation/reduction/oxidation cycle, the reason might be the weaker adsorption of O₂ and H₂, the intermediate OOH and the production H₂O₂ and the suppression of H₂O₂ decomposition.²⁰Open in a separate windowScheme 2Schematic of the mechanism for DSHP by Pd@TS-1.Open in a separate windowFig. 4H₂O₂ selectivity of DSHP over Pd@TS-1 with different oxidation states for 5 min reaction. Reaction conditions (same as Fig. 5 and ​and6):6): H₂/Ar (2.9 MPa) and air (1.35 MPa), 8.5 g solvent (2.9 g water, 5.6 g MeOH), 0.02 g catalyst, RT, 1200 rpm.The productivity of DSHP over Pd@TS-1 increased with oxidation, reduction and reoxidation treatment in 30 minutes (Fig. 5 and Table S3^†), demonstrated that PdO layer on monometallic Pd catalysts could suppress oxygen dissociation and H₂O₂ degradation,¹² the appropriate PdO formed on the surface of the catalysts after reoxidation can optimize the H₂O₂ production. The hierarchical Pd@TS-1 (35 010 mmol g_Pd⁻¹ h⁻¹) is remarkably higher than those of conventional Pd@TS-1 (3210 mmol g_Pd⁻¹ h⁻¹), the superior hydrogen peroxide production rate of Pd@HTS-1-ORO indicating that the Pd encapsulated by uniformed topology structure of TS-1 highly limited by the effect of pore-diffusion resistance.¹¹ Compared to Pd@TS-1, it was noteworthy that Pd@HTS-1 with only 0.1 wt% Pd content and subnano size after oxidative treatments showed famous reaction activity without any promoters under mild condition, which could be mainly ascribed to the presence of internal diffusion limitation within encapsulated micropore zeolites. The micropore structure limited the use of Pd metal because a part of the Pd crystal surface was blocked by zeolite supports, the hydrogen and oxygen were restricted by the configurational diffusion of zeolite to the Pd surface. Moreover, the formed and desorption H₂O₂ was also constrained by the micropore and thereby resulted in prolonged residence time of the product leading to degradation of H₂O₂. The intracrystal diffusion no longer limited the mass transport process of the hierarchical zeolite due to the presence of additional porosity. Although the physical and structural properties (including the primary particle size, the properties of the external surface and so on) were different between Pd@HTS-1 and Pd@TS-1, we may still draw a conclusion that the excellent catalytic activity is mainly attributed to the presence of mesopore favours diffusion of both reactants and products to and off the active sites in micropores.Open in a separate windowFig. 5Macro reaction rate for H₂O₂ production over Pd@TS-1 and Pd@HTS-1. ^aPd/C#C&Pd/C#Ex from Young-Min Chung;^34bPd–Sn/TiO₂ from Hutchings.²⁹The TON of H₂O₂ production at different reaction time over the six different Pd@TS-1 and Pd@HTS-1 catalysts were shown in Fig. 6. The TON increases with increasing reaction time, however, the slop of the TON–time curves (d_TON/d_t) seems decreased with increasing time, which revealed that the net productivity rate of hydrogen peroxide synthesis declined slightly with increasing time, especially for the Pd@HTS-1-OR at the reaction period of 30–60 min. The accumulative productivity of hydrogen peroxide slowed down, the reason might be the rapid decrease of hydrogen partial pressure in the medium and the ongoing H₂O₂ degradation.Open in a separate windowFig. 6The TON of H₂O₂ production with different reaction time over Pd@TS-1 and Pd@HTS-1 catalysts. TON (turnover number) = mol (H₂O₂)/mol (surface Pd).In summary, successful encapsulation of subnano-size Pd metal particles within titanium silicate (TS-1) voids was achieved via the mercaptosilane-assisted DGC synthesis method. The subnano-size Pd nanoparticles encapsulated in HTS-1 zeolites exhibited superior thermal stability after the oxidation/reduction/oxidation heat treatment process adjusting Pd/PdO hybrid owing to the embedding confinement. The synthesized high-efficiency Pd@HTS-1-ORO showed the famous hydrogen peroxide synthesis productivity, a hydrogen peroxide production rate as high as about 35 010 mmol H₂O₂ g_Pd⁻¹ h⁻¹. Our strategy brings about a finely tailored method to control particle size down to the subnano level and eliminate the diffusion inside metal encapsulated microporous zeolites, which is advantageous for catalytic activity and selectivity in direct synthesis of hydrogen peroxide. Thus, our approach opens up the possibility that the titanium-containing zeolites encapsulated noble metal catalyst can be extended further to selective oxidation reactions with H₂O₂ generated in situ from H₂ and O₂.     

Facile synthesis of high-performance SiO2@Au core–shell nanoparticles with high SERS activity

This study proposes a facile and general method for fabricating a wide range of high-performance SiO₂@Au core–shell nanoparticles (NPs). The thicknesses of Au shells can be easily controlled, and the process of Au shell formation was completed within 5 min through sonication. The fabricated SiO₂@Au NPs with highly uniform size and SERS activity could be ideal SERS tags for SERS-based immunoassay.

This study proposes a facile and general method for fabricating a wide range of high-performance SiO₂@Au core–shell nanoparticles (NPs).

The design and controlled fabrication of Au nanocomposites have attracted extensive attention because of their outstanding chemical and optical properties and wide applications in various fields, such as catalysis,¹ drug delivery,² photothermal cancer therapy,³ sensing,⁴ and surface-enhanced Raman scattering (SERS).⁵ However, small Au nanocomposites tend to aggregate, which seriously affects their stability and usability. The combination of silica nanoparticles (SiO₂ NPs) and Au shells provides a good alternative to Au nanocomposites.^6,7 These SiO₂ NPs are ideal core materials due to their high stability, easy preparation, uniform spherical shape, and large particle size range.^8,9Many synthesis methods have been explored for the fabrication of SiO₂@Au core–shell NPs; these methods include electroless plating,¹⁰ self-assembly,¹¹ layer-by-layer synthesis,¹² and seed growth.¹³ The seed growth method is the most commonly used to coat the Au shell on the surface of the SiO₂ core and involves two steps: deposition of nucleus seeds on the functionalized SiO₂ surface and Au shell growth. Although this method is beneficial for the synthesis of nanostructures with narrow size distribution, it exhibits two major shortcomings. First, the surface of SiO₂ NPs must be functionalized with various organosilanes containing amino (–NH₂) or mercapto (–SH) groups for adsorption or deposition of metal seeds on the SiO₂ NPs before subsequent growth of Au shells.^14,15 However, full surface amino/mercapto modification is often difficult to achieve; in this regard, dense metal seed layer formation on the surface of SiO₂ NPs cannot be achieved, eventually affecting the uniform and complete Au shell coating. Second, the formation of complete Au shell on the SiO₂ NPs is frequently achieved using a slow-growth approach through slow or multiple addition of HAuCl₄ to the seed-coated SiO₂ NPs suspension containing reducing agents.^16,17 The application of these slow-growth methods is restricted by its complex procedure and time-consuming preparation. Thus, a facile method must be developed for synthesis of Au coated SiO₂ NPs with controllable Au shell, good dispersibility, and fast preparation.In this work, we report a sonochemically assisted seed growth method for facile synthesis of monodisperse SiO₂@Au core–shell NPs for the first time. Cationic polyethyleneimine (PEI) was used to form a cationic thin interlayer with numerous primary amine groups for easy adsorption of dense Au seeds on the silica surface and keeping the nanostructure stability during shell growth. Sonication was used instead of traditionally used mechanical stirring to shorten the reaction time. The entire reaction process for Au shell formation was completed within 5 min. Moreover, the thickness of the Au shell was easily controlled outside the silica cores of different sizes. To the best of our knowledge, the proposed method is the most convenient synthesis route for preparation of high-performance SiO₂@Au core–shell NPs to date. Our results further demonstrate that the fabricated SiO₂@Au NP could be an ideal SERS tag for SERS-based lateral flow immunoassay (LFA). The method was validated for detection of human immunoglobulin M (IgM) and showed a detection limit as low as 0.1 ng mL⁻¹. The details of the experiments including SiO₂@Au NPs preparation, SERS-based LFA strip preparation, SERS detection protocol, and sensitivity test were provided in the ESI^† section.The synthesis principle of monodisperse SiO₂@Au NPs is presented in Fig. 1a. SiO₂ NPs were first prepared by using a modified Stöber method as the core. The SiO₂ NPs were ultrasonically treated with PEI solution to form PEI-coated SiO₂ NPs (SiO₂@PEI). The positively charged PEI effectively attached to the negatively charged SiO₂ NPs and formed a stable polymer layer via electrostatic self-assembly. SiO₂–Au seed NPs were prepared by adsorbing small Au NPs (3–5 nm) on the PEI layer of SiO₂ NPs densely and firmly through covalent binding between the –NH₂ groups of PEI and Au NPs. Finally, monodisperse SiO₂@Au NPs were quickly obtained through the reduction of HAuCl₄ by hydroxylamine hydrochloride (NH₂OH·HCl) under the stabilization of PVP. The uniform Au shells outside the SiO₂ NPs were formed within 5 minutes through the isotropic growth of all Au seeds under sonication.Open in a separate windowFig. 1Synthesis principle of SiO₂@Au NPs (a). TEM images of (b) SiO₂ NPs, (c) SiO₂–Au seed NPs, (d) SiO₂@Au NPs and their corresponding elemental mapping in (g), (h), and (i) respectively. (e) HRTEM picture and (f) bright-field TEM image of a single SiO₂@Au NP.The morphology of the as-synthesized products in different stages were characterized through transmission electron microscopy (TEM). The as-prepared SiO₂ NPs were uniform in size and had a diameter of approximately 140 nm (Fig. 1b). After coating the SiO₂@PEI NPs with Au seeds, many small seeds homogeneously adhered to the surface of the silica core (Fig. 1c). The dense Au seeds acted as randomly oriented crystalline sites for subsequent seed-mediated growth of the Au shell. Fig. 1d and e show the low- and high-magnification TEM images of the final SiO₂@Au core–shell NPs, respectively. Continuous and rough edges were detected around the SiO₂@Au NPs. The HRTEM image (Fig. 1e) indicated that large adjacent Au NPs covered the entire surface of the SiO₂ NPs, forming a complete and rough Au shell. The average particle size increased from 140 nm to 190 nm after the Au shell formation, indicating that the thickness of the Au shell was approximately 25 nm. Additionally, the SEM images (Fig. S1^†) showed that the SiO₂@Au NPs were successfully fabricated on a large scale and exhibited a rough surface and uniform size. The elemental composition of SiO₂@Au NPs was also confirmed through X-ray mapping (Fig. 1f–i). The results indicated that a layer of Au shell was uniformly coated on the surface of the SiO₂ NPs. The zeta potentials of SiO₂, SiO₂@PEI, SiO₂–Au seeds, and SiO₂@Au NPs in aqueous solution were found to be −46.7, +41.9, −7.4, and −21.1 mV, respectively (Fig. S2^†). The significant change in the zeta potential revealed the successive completion of PEI coating, Au seed adsorption, and Au shell formation. Fig. 2a shows the typical XRD patterns of the as-synthesized SiO₂–Au seed (blue line) and SiO₂@Au NPs (red line). The specific XRD pattern of Au is characterized by five peaks positioned at 2θ values of 38.3°, 44.3°, 64.5°, 77.4°, and 81.6°, which correspond to the reflections of the (111), (200), (220), (311), and (222) crystalline planes of Au (JCPDS no. 04-0784), respectively.^18,19 The intensity of the diffraction peaks of SiO₂@Au NPs increased when the Au shells were coated. No peaks of SiO₂ and PEI were detected in the XRD pattern because of their amorphous form.²⁰Open in a separate windowFig. 2Typical XRD patterns (a) and UV-visible spectra (b) of the as-synthesized products. Fig. 2b illustrates the UV-vis spectra of the as-synthesized products dispersed in deionized water in different stages. SiO₂ and SiO₂@PEI NPs had no obvious absorption peaks in the UV-vis spectra (curves a and b). SiO₂–Au seed NPs displayed a clear absorption peak at about 568 nm (curve c), which confirms the formation of the Au seed layer onto the surface of SiO₂ NPs. As the Au shell formed, the UV-vis spectral peak obviously red-shifted, and the intensity increased significantly (curve d). This result could be due to the strong interaction between and the coupling of the large adjacent Au NPs of the Au shells outside the SiO₂ NPs.²¹The strategy for Au shell formation is essentially seed-mediated growth. Thus, the surface morphology of SiO₂@Au NPs can be easily controlled by adjusting the Au³⁺ concentration by using a constant amount of SiO₂–Au seed. Fig. 3a–d shows the representative TEM images of SiO₂@Au NPs synthesized with different concentrations of HAuCl₄ while the other parameters remained constant. As the concentration of the HAuCl₄ increased from 0.01 mM to 0.04 mM, the Au seeds absorbed outside the SiO₂ NPs gradually increased in size and finally intersected with each other and formed a continuous and Au shell of a different thickness.Open in a separate windowFig. 3TEM images of SiO₂@Au NPs synthesized with different HAuCl₄ concentrations: (a–d) 0.01, 0.02, 0.03, and 0.04 mM HAuCl₄. (e) UV-vis spectra of SiO₂@Au synthesized with different HAuCl₄ concentrations: curves (a–e) 0, 0.01, 0.02, 0.03, and 0.04 mM HAuCl₄ and the corresponding Raman spectra of DTNB (f). Fig. 3e shows the UV-vis spectra of the synthesized SiO₂–Au seed and SiO₂@Au NPs with different Au shell thicknesses. The absorption peak of the obtained products red shifted gradually from 568 nm to 700 nm, and the peak width became broader with increasing concentration of HAuCl₄. Thus, the absorption peak of SiO₂@Au NPs can also reflect the formation and thickness of the Au shell. Fig. 3f shows the SERS activity of SiO₂@Au NPs prepared with different HAuCl₄ concentrations. 5,5-Dithiobis-(2-nitrobenzoic acid) (DTNB) was used as Raman molecule because it contains a double sulfur bond, which can be chemically coupled to the Au shell to form Au–S chemical bond and could produce strong and concise SERS peaks located at 1062, 1148, 1331, and 1556 cm⁻¹.^22,23 Moreover, DTNB molecules can provide free carboxyl groups as sites to conjugate antibodies.²⁴ As shown in the Raman spectra in Fig. 3f, the SiO₂–Au seed showed fairly weak SERS ability (spectra a), whereas the SiO₂@Au NPs exhibited gradually enhanced SERS activity as the HAuCl₄ concentration increased (spectra b–d). However, the overgrowth of the Au shell decreased the SERS activity of SiO₂@Au NPs (spectra e). This phenomenon could be attributed to the fully continuous Au shell formation, which reduced the nanogaps and hot spots on the surface of SiO₂@Au NPs. Hence, we chose SiO₂@Au NPs prepared with 0.02 mM HAuCl₄ as the optimal material for SERS application because of their nearly complete Au shell and optimal enhancement effect.PEI can be self-assembled on the surface of SiO₂ NPs of any size under sonication conditions. Thus, our proposed PEI-assisted seed growth method is a general route for preparing monodisperse SiO₂@Au core–shell particles with different sizes, ranging from nanoscale to microscale levels. Fig. 4a–c shows the TEM images of single SiO₂–Au seed NPs with different sizes (70–300 nm), and Fig. 4d–f clearly shows their corresponding fabricated SiO₂@Au NPs, respectively. The TEM images of multiple SiO₂@Au NPs with different sizes are displayed in Fig. S3.^† All of the obtained SiO₂@Au NPs possess homogeneous nanostructure, uniform Au nanoshell, and good dispersity. We further examine the dependence of SERS activity on the SiO₂@Au NPs size up to 350 nm. Fig. S4^† shows a set of SERS spectra of DTNB (10⁻⁵ M) adsorbed on the SiO₂@Au NPs of different sizes. The SERS intensity presented in the figure is the average intensity from 10 spots for each sample. Obviously, all the SiO₂@Au NPs exhibited excellent SERS abilities, and the signal intensities were gradually enhanced as the particle size increased. In fact, the Au shells of SiO₂@Au NPs were made of large sized Au NPs. This experimental result indicates that the larger the size of the Au NPs of shell, the higher the SERS activity achieved.Open in a separate windowFig. 4(a–c) TEM images of single SiO₂–Au seed with different sizes: (a) 70, (b) 150, and (c) 300 nm and their corresponding fabricated SiO₂@Au NPs (d), (e), and (f), respectively.For the determination of the SERS sensitivity of the 80 nm SiO₂@Au NPs, a series of DTNB ethanol solution (with concentration ranging from 10⁻⁴ M to 10⁻¹¹ M) was prepared. Each tube of DTNB solution was mixed with 10 μL of SiO₂@Au NPs (1 mg mL⁻¹) and sonicated for 1 h. After separation and washing, the final precipitate was dropped on a Si chip and analyzed with Raman signals. The spectra and calibration curve of DTNB absorbed on the SiO₂@Au NPs are shown in Fig. 5a and b, respectively. The SERS signal significantly decreased as the concentration of DTNB decreased, and the main Raman peak at 1331 cm⁻¹ remained evident at DTNB concentrations as low as 10⁻¹⁰ M. Thus, the limit of detection (LOD) of DTNB is 10⁻¹⁰ M. These results indicate that the SiO₂@Au NPs have good potential to be active SERS substrate for greatly enhancing the SERS signal of molecules adsorbed on them.Open in a separate windowFig. 5(a) SERS spectra of DTNB measured with different concentrations on the SiO₂@Au NPs. (b) Calibration curve for DTNB at a concentration range of 10⁻⁴ M to 10⁻¹¹ M obtained using SERS intensity at 1331 cm⁻¹. The error bars represent the standard deviations from five measurements.Upon modification with Raman report molecules and detection antibodies, the monodisperse SiO₂@Au NPs must be efficient SERS tags for highly reproducible SERS immunoassays due to the integration of high SERS activity of the Au nanoshell and the homogeneity and stability of SiO₂ NPs (Fig. 6a). SERS-based LFA strip is a recently reported analytical technique to overcome the shortcomings of conventional lateral flow assay, such as poor sensitivity and semiquantitative ability on the basis of colorimetric analysis.^25–27 The fundamental principle of SERS-based strip is the use of functional SERS tags instead of Au NPs. High-sensitivity and quantitative detection can be achieved by Raman spectroscopy because the intensity of the SERS signal is directly proportional to the number of SERS tags on the test line.Open in a separate windowFig. 6(a) Synthesis route for SiO₂@Au SERS tags. (b) Schematic of SERS-based LFA strips for quantitative detection of human IgM. Fig. 6b represents the operating principle of the monodisperse SiO₂@Au NPs (80 nm) based SERS-LFA strip. Human IgM was selected as the model target antigen to explore the sensitivity of the proposed method. The representative SERS-LFA strip is composed of a sample loading pad, a conjugate pad, a NC membrane containing test line and control line, and an absorption pad. In our system, goat anti-human IgM antibody-labeled SiO₂@Au/DTNB NPs were dispensed onto the glass fiber paper as the conjugate pad, and the goat anti-human IgM antibody and donkey anti-goat immunoglobulin G (IgG) antibody were dispensed onto the NC membrane to form the test line and control line, respectively. When the sample solution containing the target human IgM passed through the conjugation pad, immunocomplexes (human IgM/SERS tags) were formed and continued migrating along the NC membrane until they reach the test line where they were captured by the previously immobilized anti-human IgM antibodies. Excess antibody-conjugated SiO₂@Au tags continued to migrate to the control line and were immobilized by the donkey anti-goat IgG antibody. Consequently, two dark bands appeared in the presence of the target human IgM (positive), whereas only the control line turned to a dark band in the absence of human IgM (negative). Quantitative detection of human IgM could be realized by detecting the SERS signal on the test line.Human IgM was diluted within 10 000 ng mL⁻¹ to 0.1 ng mL⁻¹ as the sample solution, and PBST solution (10 mM PBS, 0.05% Tween-20) was used as blank control. As shown in Fig. 7a, the color of SERS tags captured by the test line was visualized and gradually decreased with decreasing human IgM concentration. The LOD of colorimetric method for detection of human IgM was found to be 10 ng mL⁻¹. Quantitative analysis was also conducted by measuring the characteristic Raman signals of the SERS tags on the test lines, and the Raman spectra are displayed in Fig. 7b. The Raman spectra were analyzed by plotting the intensity at 1331 cm⁻¹ of DTNB as a function of the logarithm of the target human IgM concentration to generate a calibration curve (Fig. 7c). The LOD of the SERS-LFA strips based on the SiO₂@Au tags is 0.1 ng mL⁻¹, which was calculated as a 3 : 1 threshold ratio with respect to the blank control measurement. Using SiO₂@Au tags-based SERS LFA strip offers a 100-fold improvement in the detection limit compared with colorimetric analysis. Basing on these results, we demonstrated the high efficiency and great potential of monodisperse SiO₂@Au NPs as suitable SERS tags for SERS-based LFA strips. The specificity of the SERS-LFA strips was tested by a high concentration (1 μg mL⁻¹) of other proteins including human IgG and BSA. Fig. S5^† shows the result of the specificity test. IgG and BSA did not show significant interference signals both in visualization and Raman spectrum analyses, whereas 100 ng mL⁻¹ human IgM exhibited a strong signal. Hence, the SiO₂@Au tags-based SERS-LFA strip has good selectivity.Open in a separate windowFig. 7(a) Photographs of SERS-based LFA strips in the presence of different concentrations of human IgM. (b) SERS spectra measured in the corresponding test lines. (c) Plot of the Raman intensity at 1331 cm⁻¹ as a function of the logarithmic concentration of human IgM. The error bars represent the standard deviations from five measurements.In summary, this work proposes a sonochemically assisted seed growth method for facile synthesis of monodisperse SiO₂@Au core–shell NPs with a complete Au shell. This method is a general route for preparing SiO₂@Au particles with sizes ranging from nanoscale to microscale levels. High-performance SiO₂@Au NPs were obtained from the intermediate product (SiO₂–Au seed) within 5 min through sonication. The obtained SiO₂@Au NPs were highly uniform in size and shape and exhibited satisfactory SERS activity. Hence, these NPs could be ideal SERS tags for various SERS based immunoassays. The small SiO₂@Au NPs (80 nm) with light weight and good dispersibility were also successfully applied to SERS-based LFA strip for human IgM rapid detection, with limit of detection as low as 0.1 ng mL⁻¹. We expect that high-performance SiO₂@Au NPs SERS tags can be used for actual detection.     

Correction: Compositional effect on the fabrication of AgxPd1−x alloy nanoparticles on c-plane sapphire at distinctive stages of the solid-state-dewetting of bimetallic thin films

Correction for ‘Compositional effect on the fabrication of Ag_xPd_1−x alloy nanoparticles on c-plane sapphire at distinctive stages of the solid-state-dewetting of bimetallic thin films’ by Puran Pandey et al., RSC Adv., 2017, 7, 55471–55481.

Errors were present in the published article and ESI. The errors in the article are in the plots of SAR and coverage in Fig. 6(m) and (n) and the corrected figure is shown below. At the same time, Fig. 6(l) has been edited in order be consistent with Fig. 6(m) and (n). Specifically, the blue lines denote “Pd_0.25Ag_0.75” and the black lines “Pd_0.75Ag_0.25”. In the ESI summarized values of R_q in Table S7 were incorrect and the ESI document is now replaced.Open in a separate windowFig. 6Evolution of Ag–Pd bimetallic nanostructures by the variation of annealing durations between 0 and 3600 s at 650 °C with the deposition thickness of 10 nm and distinct bilayer composition as labelled. (a)–(f) AFM top-views of 3 × 3 μm². (g) and (h) Summary of EDS intensities of Ag Lα1 and Pd Lα1 with respect to the annealing durations. (i) and (j) Reflectance spectra of Ag–Pd nanostructures. (k) Summary plot of average reflectance. (l)–(n) Summary plot of R_q, SAR and coverage plots with respect to the annealing duration.The Royal Society of Chemistry apologises for these errors and any consequent inconvenience to authors and readers.     

Integrating plasmonic nanostructures with natural photonic architectures in Pd-modified Morpho butterfly wings for sensitive hydrogen gas sensing

This work reports a bioinspired three-dimensional (3D) heterogeneous structure for optical hydrogen gas (H₂) sensing. The structure was fabricated by selective modification of the photonic architectures of Morpho butterfly wing scales with Pd nanostrips. The coupling of the plasmonic mode of the Pd nanostrips with the optical resonant mode of the Morpho biophotonic architectures generated a sharp reflectance peak in the spectra of the Pd-modified butterfly wing, as well as enhancement of light–matter interaction in Pd nanostrips. Exposure to H₂ resulted in a rapid reversible increase in the reflectance of the Pd-modified butterfly wing, and the pronounced response of the reflectance was at the wavelength where the plasmonic mode strongly interplayed with the optical resonant mode. Owing to the synergetic effect of Pd nanostrips and biophotonic structures, the bioinspired sensor achieved an H₂ detection limit of less than 10 ppm. Besides, the Pd-modified butterfly wing also exhibited good sensing repeatability. The results suggest that this approach provides a promising optical H₂ sensing scheme, which may also offer the potential design of new nanoengineered structures for diverse sensing applications.

Three-dimensional heterogeneous nanostructures that integrate plasmonic nanostructures of Pd with photonic architecture of Morpho butterfly wings can achieve sensitive hydrogen gas detection.

Hydrogen gas (H₂) is widely used in many industrial processes and also considered as a sustainable and environmentally friendly energy source that holds promise as a replacement for fossil fuels.^1,2 H₂, however, is highly volatile with a low flammability point of ∼4 vol% in air, which gives rise to risk of explosion. Due to the colorless and odorless nature of H₂, accurate and sensitive H₂ sensors with rapid response are highly demanded for leakage detection in various applications.³Palladium (Pd) can rapidly absorb large amounts of H₂ into its crystal lattice and form palladium hydride reversibly under ambient conditions, which induces expansion as well as changes in electrical and dielectric properties.^4,5 Therefore, Pd is considered as an effective H₂ sensing material. A large number of H₂ sensors, typically including electrical and optical ones, have been demonstrated by employing Pd as the active material.^6–11 Compared with electrical H₂ sensors, optical sensors have particular advantages in practical applications since they are immune to electromagnetic interference and also inherently safe as no electric spark can be generated.^11,12In recent years, intensive researches have been conducted to develop optical H₂ sensors with Pd nanostructures, which offer the potential of miniaturization and the fast reaction kinetics originated from the short diffusion lengths for H atoms.^13–20 Pd nanostructure strongly interacts with light and shows intriguing optical phenomenon due to the localized surface plasmon resonance (LSPR). Such LSPR is associated with the resonant excitation of collective oscillations of the free electrons by incident light and can generate large electromagnetic field confinement at nanometer scale.²¹ The position and intensity of the plasmon resonance peak of Pd nanostructure change with the hydrogen-induced changes of volume and dielectric property, which can be used for the readout of the direct nanoplasmonic sensing scheme.^22–24 Nevertheless, the sensing performance of those plasmonic H₂ sensors is restrained from a fundamental limitation, in which the LSPR of Pd nanostructure generally exhibits broad resonance peak owing to the interband electronic transitions.^25,26 To overcome this problem, several efforts have been devoted to achieve other possible sensing schemes containing Pd nanostructures. For instance, H₂ sensor based on perfect absorption in the visible wavelength range was designed, utilizing a coupled plasmonic system that consisted of Pd nanowires arranged on the top of a thick gold (Au) film separated with a spacer layer of MgF₂.²⁷ Indirect LSPR sensors were proposed by precisely placing Pd nanoparticles (NPs) in the vicinity of other metallic NPs that possessed superior LSPR properties and acted as optical antennas to enhance the response of Pd NPs in the stimulation with H₂.^28–32 Besides, some Au–Pd core–shell nanostructures with various morphologies have also been synthesized through wet chemistry methods in order to take advantage of the local field enhancement of Au core to improve the optical response to H₂.^33–37Here we explored a different H₂ sensing approach that is based on three-dimensional (3D) photonic architectures of the Morpho butterfly wing scales modified by Pd nanostructures, as shown in Fig. 1. The iridescent scales of Morpho butterfly wing have unique multilayered air–chitin structures and produce a sharp reflectance peak in the blue region of the spectrum. Such structures have been demonstrated in many high-performance optical sensors.^38–42 We take advantage of the narrow-band resonance and sensitive feature of the photonic crystal to facilitate H₂ sensing through incorporating Pd nanostructures into such biological photonic crystal structures. Specifically, the 3D heterogeneous structures of Morpho butterfly wing containing Pd nanostrips distributed on the edge portion of the lamella layers of the wing scale were generated through physical vapor deposition (PVD) of Pd (Fig. 1). Owing to coupling between the plasmonic mode of Pd nanostrips and the optical resonant mode of the biophotonic nanostructures, the Pd-modified butterfly scales showed a sharp reflection peak, and light–matter interaction in Pd nanostructure was enhanced. We demonstrated that the synergetic effect of Pd nanostrips and biophotonic structures played a role in H₂ sensing, which resulted in the sensitive response of the Pd-modified butterfly scales upon exposure to H₂. This work should provide some stimulation for the design of sensing platforms that combine plasmonic nanostructures with photonic crystals.Open in a separate windowFig. 1Schematic illustration showing optical H₂ sensing based on the 3D heterogeneous structures that are consisted of Pd nanostrips and the photonic architectures of the Morpho butterfly wing scales.The Morpho sulkowskyi butterfly was chosen as model for the fabrication of H₂ sensing platform. As shown in Fig. 2a, the Morpho butterfly wing displays brilliant blue iridescence originated from the elaborate hierarchical photonic structures of the scales. The scales are regularly arranged on the wing surface, as presented in Fig. 2b. They are ∼200 μm in length and ∼50 μm in width (Fig. 2c). On each scale, there are ordered arrays of ridges running along the longitudinal direction and adjacent ridges are connected each other by cross-ribs in the transversal direction, as presented in Fig. 2d. The high magnification scanning electron microscopy (SEM) image in the inset of Fig. 2d shows that the ridges contain multilayered lamellae folds. The details of the multilayered structures are shown by a cross-section transmission electron microscopy (TEM) image of ridges in Fig. 2e, which reveals the Christmas tree-like structures. A stem with width of 50–120 nm stands at the middle of each ridge. Approximately eight lamellae that are separated by air decorate at both sides of the stem, with the width of lamella gradually decreasing from the bottom to the top of the stem. The average thicknesses of lamella and the spacing between the lamella are ∼65 nm and ∼150 nm, respectively. Multilayer interference of light from the lamella layers, in combination with the light diffraction from the arrays of the ridges, contributes to the iridescence color appearance of the butterfly wing.^42,43 We used the PVD method to deposit Pd coating on the butterfly wing structures. Considering the formation of the continuous coating and the response time of the 3D structure, we set the thickness of the Pd coating at 15 nm. During the deposition of Pd, the butterfly wing was placed directly under a Pd source so that the Pd coating was deposited vertically onto the butterfly wing structures. After coating the Pd layer, the edge portions of the lamellae were covered with Pd, resulted in selective modification of photonic structures of butterfly wing scales with Pd nanostrips, as shown in Fig. 2f and S1.^†Open in a separate windowFig. 2Morphologies and structures of the Morpho butterfly wing before and after selective modification with Pd. (a) The photo of the Morpho sulkowskyi butterfly. (b) Optical microscopy image of the stacked scales on the wing surface. (c) SEM image of a single butterfly wing scale supported on a silicon substrate. (d) Top SEM view of the photonic architecture of the scale. The inset figure is a high magnification SEM showing the multilayered lamella structures. (e) TEM image of a transverse section of the scale showing ridges with lamella structures. (f) TEM image showing the selectively modified ridges, where the edge of each lamella was coated with Pd nanostrip.We investigated the optical reflectance of the Pd-modified butterfly wing at normal incidence. Compared with the reflectance spectra of original scales, the reflectance of Pd-modified butterfly wing scales with a Pd layer of about 15 nm in thickness exhibited a main peak, which underwent blue shift from ∼485 nm to ∼460 nm and also showed smaller full width at half maximum than the original scales, as shown in Fig. 3a and b. Moreover, decorating of the Morpho scales with Pd nanostrips produced a decrease in the reflectance along with a minor peak at the wavelength near the violet end of the reflection spectra as well as a weak peak at ∼560 nm. These spectral changes were most likely due to the interaction of the reflection band originated from butterfly wing structures with the plasmonic absorption from the Pd nanostrips on the wing structures. In order to study this optical interaction, the reflectance spectra for the original butterfly structures and the butterfly structures decorated with Pd nanostrips were calculated through finite-difference time-domain (FDTD) methods, as presented in Fig. 3c and d, respectively. The simulated reflectance spectra exhibited good matching with the corresponding experimental results. We also performed simulations to separately analyze the plasmonic absorption of Pd nanostrips, as shown in Fig. S2.^† The calculated absorption cross-section revealed the broad plasmonic absorption band of Pd nanostrip, which overlaps with the reflection band of butterfly wing structures. Besides, we computed the electric-field distributions of the Pd nanostrips on a planar dielectric substrate (Fig. S3^†) and the butterfly wing structures with Pd nanostrips (Fig. 3e) at resonance, respectively. The light was concentrated at the tips of the Pd nanostrips and larger enhancement was observed in the case of the Pd nanostrips distributed in the butterfly wing structures, which indicated that the interaction of light with Pd nanostrips were enhanced by coupling to the optical cavity of Morpho butterfly scales. As such, the plasmonic mode of Pd nanostructures could effectively interplay with resonant mode of the biophotonic nanostructures, which thus was expected to enhance the sensitivity in H₂ sensing.Open in a separate windowFig. 3Optical properties of the original butterfly wing scales and the modified scales with Pd nanostrips. (a) Measured reflectance spectra of the original Morpho butterfly wing. (b) Measured reflectance spectra of the modified butterfly wing scales with Pd nanostrips. (c) Calculated reflectance spectra of the original Morpho butterfly scales. (d) Calculated reflectance spectra of Pd-modified butterfly scales. (e) The simulated electric field distribution at the wavelength of 500 nm showing the enhanced light–matter interaction in the Pd nanostrips distributed in the butterfly wing structures.To examine the response of Pd-modified butterfly wing scales to the exposure of H₂, the butterfly wing sample was placed in a glass chamber with an optical fiber mounted through the top hatch at room temperature, as shown in Fig. S4.^† The chamber was connected to a gas inlet channel for H₂ and nitrogen (N₂) carrier gas that were premixed before flowing into the chamber. The concentration of H₂ was regulated through changing the gas flow of N₂ and H₂ with two mass flow controllers. We recorded the reflectance spectra of the sample at different H₂ concentrations. To evaluate the changes in reflectance spectra following H₂ exposures, we calculated the relative reflectance ΔR(λ) according to⁴⁴ΔR(λ) = 100% × [R(λ)/R₀(λ)]1where R₀(λ) is the spectrum collected from the Pd-modified butterfly wing scales in pure N₂ and R(λ) is the spectrum collected upon exposure to H₂ in N₂ carrier gas. Fig. 4a showed relative reflectance of the Pd-modified butterfly wing scales with respect to different H₂ concentrations over a range from 0.001% to 4%. Exposure to H₂ caused the reflectance increase, and the pronounced ΔR response was over the wavelength region of 400–550 nm, which took place in the same wavelength range of the strong interaction between the plasmonic mode of Pd nanostructures and resonant mode of the biophotonic nanostructures. The reflectance of the original Morpho butterfly wing scales with respect to different H₂ concentrations over a range from 0.1% to 4% also was measured, which suggested that there was no obvious response (Fig. S5^†). In the presence of H₂, the Pd nanostrips dissociated H₂ molecules and absorbed the H atoms to form Pd hydride, typically accompanied by lattice expansion and the change of dielectric properties to less metallic. As a result, the plasmonic absorption of the Pd nanostrip was altered and the spectral variations of the heterogeneous structures were induced. The increase in the reflectance response of the Pd-modified butterfly wing was observed with the increase in H₂ concentration, which was attributed to the larger change of dielectric function and volume expansion.Open in a separate windowFig. 4Response of the Pd-modified photonic architectures to H₂ gas. (a) Relative reflectance measured in different H₂ concentrations. (b) The simulated relative reflectance of the Pd-modified butterfly scales due to expansion and change in dielectric function when Pd is converted into β-PdH_x. (c) Normalized reflectance change at wavelength of 500 nm in different H₂ concentrations showing a detection limit below 10 ppm. (d) Normalized reflectance change at wavelength of 500 nm in different H₂ concentrations from 0.1% to 4%.To further understand the sensing mechanism, we carried out simulations to compare the reflectance of Pd-modified butterfly scales as well as the Pd nanostrips on a planar dielectric substrate before and after H₂ uptake, as shown in Fig. 4b and S6.^† For simplicity, we assumed that the Pd was completely converted into β-phase (β-PdH_x) after the exposure of H₂ gas. In the simulations, we examined the contributions of two factors: volume expansion and the change in dielectric function. The volume expansion of Pd nanostrips alone led to a little decrease in the ΔR(λ) spectra of the heterogeneous structures over the short-wavelength range and slight increase over the long-wavelength range. The change in dielectric function led to a clear increase in the ΔR(λ) spectra over the 430–510 nm. The calculated results of the combined effects of volume expansion and the change in dielectric function suggested that the change of dielectric function has more influence on the optical response of the Pd-modified biophotonic structures. The calculated ΔR(λ) spectra in Fig. 4b are in general agreement with the experimental spectra (Fig. 4a). The pronounced feature in calculated and experimental spectra was the increase in the ΔR(λ) spectra over the wavelength range of the strong interaction between the plasmonic mode and the resonant mode of the biophotonic nanostructures.The dynamic change of reflectance for the Pd-modified butterfly wing sample at the wavelength of 500 nm were shown in Fig. 4c and d. The optical response clearly showed the dependence on the H₂ concentration. For comparison, the temporal reflectance was normalized to the reflectance of the sample in pure N₂. As the concentration of H₂ increased, the normalized reflectance at wavelength of 500 nm increased, and the sensor exhibited a lowest detectable response at concentrations of 10 ppm. The low noise level relative to the response indicated a detection limit less than 10 ppm, which is among the lowest detection limit for optical H₂ sensing based on plasmonic Pd nanostructures.^13,19,30 The high sensitivity of the Pd-modified butterfly wing scales mainly arose from the synergetic effect between Pd nanostrips and biophotonic structures.Furthermore, the reflectance change at the wavelength of 500 nm as a function of H₂ concentration in the range from 0.001% to 4% suggested a positive relationship (Fig. 5a). It is worth to note that a linear relationship between the reflectance change and the H₂ concentrations was observed within the high concentration range from 0.75% to 4% (inset of Fig. 5a), which is approximately corresponding to the regimes of mixed α + β-phase and the β-phase Pd hydride.^4,30 We also measured the response time of the Pd-modified butterfly wing scales at different H₂ concentrations (Fig. 5b). The response time was defined as the time needed to reach 90% of the equilibrium ΔR. As depicted in Fig. 5b, when H₂ concentration increased, the response time decreased first, and then gradually became relatively stable. The observed response time is within the range of response time reported so far.^27,33 The sensing speed could be further improved by using different materials such as the Pd alloys.⁴⁵Open in a separate windowFig. 5Sensing performance of the Pd-modified Morpho butterfly scales to H₂. (a) The plot showing relative reflectance at wavelength of 500 nm for different H₂ concentrations. The inset figure shows a good linear relation between the relative reflectance and the H₂ concentration at the range of 0.75–4%. (b) The response time for different H₂ concentrations. (c) Five cycles of H₂ exposure for concentrations of 0.5% and 1%.In addition, we monitored the optical response of our sensor during five on/off H₂ cycles at the concentration of 0.5% and 1%, respectively. The temporal response of normalized reflectance at the wavelength of 500 nm was shown in Fig. 5c. The Pd-modified butterfly wing scales exhibited a rather consistent response for each repeated cycle, which suggested a good repeatability and stability of H₂ sensing.For investigation of the selective response of the Pd-modified Morpho butterfly wing scales, the relative reflectances upon exposure to H₂ and several potential interfering gases, including O₂, CO₂, and CH₄ were compared, as shown in Fig. S7.^† The responses from the interfering gases were much less than that observed for H₂ at similar concentration, indicating that the 3D heterogeneous structure has a little cross-sensitivity to these interfering gases.     

Graphene reinforced carbon nanofiber engineering enhances Li storage performances of germanium oxide

The rational design of electrode materials with high power and energy densities, good operational safety, and long cycle life remains a great challenge for developing advanced battery systems. As a promising electrode material for rechargeable batteries, germanium oxide (GeO₂) shows high capacity, but suffers from rapid capacity fading caused by its large volume variation during charge/discharge processes and poor rate performance owing to low intrinsic electronic conductivity. In this study, a novel one-dimensional (1D) carbon/graphene-nanocable–GeO₂ nanocomposite (denoted as GeO₂/nanocable) is rationally designed and prepared via a facile electrospinning method. Specifically, amorphous carbon and graphene spontaneously construct a nanocable structure, in which graphene acts as the “core” and amorphous carbon as the “shell”, and GeO₂ nanoparticles are encapsulated in the nanocable. The graphene “core” promises good electrical conductivity while the amorphous carbon “shell” guarantees fast Li ions diffusion. When tested as an anode material for rechargeable lithium ion batteries, the GeO₂/nanocable exhibits remarkable Li storage performance, including high reversible capacity (900 mA h g⁻¹), high capacity retention (91% after 100 cycles), and good rate performance (595 mA h g⁻¹ at 5000 mA g⁻¹).

In the GeO₂/nanocable, amorphous carbon and graphene spontaneously construct a nanocable structure, graphene “core” promises the good electrical conductivity while the amorphous carbon “shell” guarantees the fast Li ions diffusion.

Lithium ion batteries (LIBs) require longer cycle lifetimes, and higher energy density and rate capability in order to satisfy the increasing popularity of electric vehicles (EVs) and hybrid vehicles (HEVs). Nevertheless, the current commercial LIBs using graphite anode materials are unable to meet this ever-growing demand because of their relatively low capacity (372 mA h g⁻¹) and safety issues due to their low Li intercalation potential.^1–7GeO₂ is considered as a good alternative for graphite as an anode material for LIBs because of its many advantages, including a high theoretical capacity of 1125 mA h g⁻¹, low operating voltage and rapid Li⁺ diffusion rate.^8–15 In practical use, GeO₂ anodes suffer from fast capacity degradation and poor rate performance caused by their large volume variations during lithiation/delithiation cycles and low intrinsic electronic conductivity.^{6,9,11,13,16–19} The hybridization of GeO₂ with conductive buffer materials such as graphene, amorphous carbon, and carbon nanotubes are effective strategies to address these shortcomings.^{5,6,10,20–26} In particular, electrospinning methods that tailor GeO₂ anode materials into one-dimensional (1D) carbon nanofibers have attracted the attention of many researchers, because carbon nanofibers with short Li ion diffusion pathways are recognized as good architectures for energy storage applications.^27–30 However, the electrochemical performances of these GeO₂/carbon nanofibers are still unsatisfied because: (i) carbon nanofibers typically could not withstand the large volume change of GeO₂ due to their structural fragility,²⁷ thus lead to the poor cycling performance; (ii) carbon nanofibers usually exhibit relative low electronic conductivity compared to that of graphitized carbon,³⁰ therefore the rate performance of these electrodes is still not satisfactory.In order to overcome the above mentioned two drawbacks that widely existed in carbon nanofibers, in this work, we tailored graphene in the internal structure of carbon nanofibers to form a nanocable structure via a facile electrospinning method. Benefitting from the favorable mechanical properties, and electronic conductivity of graphene, the as-prepared carbon/graphene nanocable successfully mitigates the drawbacks of carbon nanofiber electrodes. As illustrated in Scheme 1, after the electrospinning and the following calcination processes, a ternary nanocomposite–amorphous carbon/graphene-nanocable-encapsulated GeO₂ (denoted as GeO₂/nanocable) was obtained. In this unique nanocable architecture, graphene acts as the “core” and amorphous carbon as the “shell”, and simultaneously GeO₂ was also encapsulated into the “nanocable”. When tested as an anode material for LIBs, GeO₂/nanocable exhibits enhanced cycling and rate performances compared to those of GeO₂/carbon nanofibers (denoted as GeO₂/CNF, prepared with the absence of graphene) electrodes.Open in a separate windowScheme 1Schematic illustration of GeO₂/nanocable.Scanning electron microscopy (SEM) images show as-prepared products possess a 1D fiber-like morphology with a typical length on the order of 10–100 μm and an average diameter of ∼300 nm (Fig. 1a and b). The microstructure of the GeO₂/nanocable was further investigated by transmission electron microscopy (TEM) (Fig. 1c–e) accompanied by selective area electron diffraction (SAED). As shown in Fig. 1c, the graphene “core” was clearly embedded within an amorphous carbon “shell”, judging by the distinct contrasts in the TEM images. The “shell” has a thickness of ∼100 nm while the “core” has a diameter of approximately 200 nm. Graphene enhanced the flexibility of the GeO₂/nanocable. As depicts in Fig. S1,^† after bending, the structure of GeO₂/nanocable could remain intact while the GeO₂/CNF collapsed. The formation mechanism of the GeO₂/nanocable prepared by a single-hole needle should be the conductivity difference between graphene and the electrospinning solution (PAN dissolved in DMF). Driven by a high voltage electrostatic force, graphene nanosheets with good electrical conductivity may join together to form the nanocable''s “core”, and the corresponding PAN solution forms the amorphous carbon “shell”. As shown in Fig. 1d and e, higher-magnification images show that many nanoparticles of diameter < 20 nm were attached to the “core”. The inset of Fig. 1d shows the SAED rings of GeO₂, where the inner and outer diffraction rings correspond to the diffractions of the (100) and (101) planes, respectively.³¹ Therefore, the above nanoparticles may be reasonably attributed to GeO₂ primary nanoparticles. Fig. 1f shows the dark field scanning transmission electron microscopy (STEM) image of GeO₂/nanocable, where the bright contrast further confirms the nanocable structure of the product. Energy-dispersive spectroscopy (EDS) elemental mapping analysis was employed to investigate the elemental distribution of the GeO₂/nanocable. As shown in Fig. 1g–i, the C, O, and Ge elemental maps match well with the STEM image (Fig. 1f). In Fig. 1g, as is consistent with the TEM image, the C elemental map is consisted of light red “shell” and dark red “core”. Combined with the above TEM analysis, the light red “shell” is recognized as amorphous carbon, because the texture of the amorphous carbon is the same as that obtained without graphene (as depicted in Fig. S2^†). The dark red “core” is supposed as graphene based on the fact that GO is the only possible carbon source except PAN. From Fig. 1h–i, Ge and O are not uniformly distributed over the entire area of the nanocable but are concentrated in the “core” area. Because when the GO solution was mixed with Ge⁴⁺, Ge⁴⁺ would be selectively bonded with the oxygenated groups by electrostatic forces due to GO nanosheets contained epoxyl and hydroxyl groups on the basal planes and carboxylic acid groups.³² This is another evidence that support graphene is the “core” of the nanocable.Open in a separate windowFig. 1(a and b) SEM images of GeO₂/nanocable at low and high magnifications. (c and d) TEM and (e) HRTEM images of GeO₂/nanocable, inset of (d) is the corresponding SAED patterns; (f) dark field STEM image and (g–i) EDS-elemental mapping images of a single GeO₂/nanocable (images g, h and i represent C, O and Ge elements, respectively).X-ray photoelectron spectroscopy (XPS) curves of the GeO₂/nanocable shown in Fig. 2a indicate the existence of Ge, C, and O elements. The corresponding high-resolution spectrum shows that there is a sharp XPS peak of Ge 3d at a binding energy at 32.8 eV, confirming the presence of Ge⁴⁺ in the GeO₂/nanocable (Fig. 2b).^11,33 Moreover, a high resolution O 1s peak is displayed in Fig. 2c at 531.8 eV, suggesting that oxygen exists in the O²⁻ oxidation state.^34,35 The high-resolution C 1s spectrum shows one primary and one shoulder peak centered at 284.7 and 286.7 eV corresponding to C–C and C–N, respectively (Fig. 2d).³⁶Open in a separate windowFig. 2XPS spectra of GeO₂/nanocable: (a) the full XPS spectrum of the GeO₂/nanocable; (b–d) high-resolution spectra Ge 3d, C 1s and O 1s, respectively. Fig. 3a shows the X-ray diffraction (XRD) patterns of the GeO₂/nanocable. The sharp diffraction peaks centered at 20.5°, 26.3°, and 38.2° corresponded to the (100), (101), and (102) planes of the crystalline GeO₂, respectively, thereby confirming the presence of GeO₂.³⁷ No carbon and graphene-related peaks were observed because of their relatively low crystallinity compared with that of GeO₂.¹²Fig. 3b shows the Raman spectra of commercial GeO₂ and GeO₂/nanocable. The sharp peak at 443 cm⁻¹ corresponds to the characteristic peak of GeO₂ (red). The absence of GeO₂ peak in GeO₂/nanocable (green) implies most of GeO₂ was beneath the amorphous carbon “shell” and its content in the “shell” was very low, this result is consistent with the above EDS mapping analysis. A 2D band, which is the characteristic band of graphene can be observed at 2600–3000 cm⁻¹ in the Raman spectra of GeO₂/nanocable further confirms the existence of graphene.³⁸ Two sharp peaks at 1332 and 1590 cm⁻¹ are present in the GeO₂/nanocable spectrum, which could be assigned to the defect (D) and graphitized (G) bands of carbon, respectively.³⁹ The intensity ratio of the D band is obviously higher than that of the G band, which indicates that higher amounts of disordered carbon were formed with numerous defects in the amorphous carbon layer (Raman spectra of nanomaterials primarily yield surface information). Amorphous carbon has two effects on the rate performance of LIBs. On the one hand, disordered carbon would enhance the Li ion diffusion kinetics, thus improving the high-rate performance during charge/discharge cycles of the LIBs.^40,41 On the other hand, excessive amorphous carbon (or thick coating layer) would reduce the electronic conductivity of the electrode, which is harmful to its rate performance.⁴² In the GeO₂/nanocable, the graphene “core” promises the good electrical conductivity while the amorphous carbon “shell” guarantees the fast Li ions diffusion, thus the high power density of the anode material could be anticipated.Open in a separate windowFig. 3(a) XRD pattern of GeO₂/nanocable, (b) Raman spectra of GeO₂/nanocable (green line) and commercial GeO₂ powder (red line), (c) TG curve of GeO₂/nanocable in oxygen atmosphere, (d) nitrogen adsorption and desorption isotherms of GeO₂/nanocable, inset image is the corresponding pore size distribution plots.The GeO₂ content in the GeO₂/nanocable was determined by thermal gravimetric analysis (TGA). In the GeO₂/nanocable, the weight ratio of GeO₂ is 53.56 wt%, and the weight ratio of graphene and amorphous carbon is 46.44 wt% based on the weight loss on carbon combustion and the fact that GeO₂ is stable in air. The weight loss that commences at 500–600 °C could be attributed to the graphene and the amorphous carbon combustion reaction. The specific surface area of the GeO₂/nanocable, which is calculated using Brunauer–Emmett–Teller (BET) measurements, is 28.3 m² g⁻¹. The nitrogen adsorption–desorption isotherm exhibits a typical IV-type isotherm with an H3 type hysteresis loop (Fig. 3d).¹² These surface area values indicate that the GeO₂/nanocable possesses a porous nanostructure, which may be caused by the amorphous carbon layer. According to the above structural characterization, we believe that the rationally designed GeO₂/nanocable could be presented an ideal anode material for high-performance LIBs.To systematically study the electrochemical performance of the GeO₂/nanocable, various electrochemical tests including cyclic voltammetry (CV), electrochemical impedance spectroscopy (EIS) and galvanostatic charge/discharge were performed. GeO₂/CNF was also tested for comparison. Initially, the Li storage mechanism of the GeO₂/nanocable was investigated by using CV and the corresponding CV curves are shown in Fig. 4a. The sample was tested at a scan rate of 0.2 mV s⁻¹ from 0.0 to 3.0 V vs. Li⁺/Li. During the first cathodic scan, the peak at around 0.65 V arose from the decomposition of the electrolyte, the irreversible reaction between electrode and electrolyte to form a stable solid electrolyte interface (SEI) layer, and the irreversible reaction of Li and GeO₂ to form Li₂O (GeO₂ + 4Li⁺ → Ge + 2Li₂O).^8,14,43 The sharp cathodic peak below 0.30 V corresponded to a series of Li_xGe phases.^13,43 During the anodic scan, the peak at around 0.35 V was caused by the dealloying reaction of Li_xGe alloys.^20,44,45 The broad peak located at approximately 1.15 V arose from the reoxidation of Ge to GeO₂, thus result in the partial reversibility of the GeO₂ conversion reaction.^14,15 After the first cycle, the CV curves of the GeO₂/nanocable overlapped well, suggesting good stability and reversibility of the GeO₂/nanocable electrode for Li ions insertion and extraction.Open in a separate windowFig. 4(a) CV curves of the GeO₂/nanocable in the voltage window 0.0–3.0 V at a scan rate of 0.1 mV s⁻¹. (b) Charge–discharge curves of GeO₂/nanocable at the 1st, 10th, 30th, 50th and 100th cycles (current: 200 mA g⁻¹). (c) Comparison of the cycling performance of GeO₂/nanocable and GeO₂/CNF 200 mA g⁻¹. (d) Rate performance the GeO₂/nanocable at different current densities. (e) EIS spectra of GeO₂/nanocable and GeO₂/CNF electrodes.The galvanostatic charge–discharge profiles of the GeO₂/nanocable electrodes were recorded in the voltage window of 0.0–3.0 V versus Li/Li⁺ at a current rate of 200 mA h g⁻¹ over 100 cycles (Fig. 4b). In the first discharge profile, a voltage plateau at approximately 0.4 V and a subsequent long continuous voltage drop down to 0.0 V could be observed, which match well with the CV data and are indicative of Li-alloying reactions. The GeO₂/nanocable electrode displays an initial discharge/charge capacity of 1470/900 mA h g⁻¹; the high initial irreversible capacity is related to the formation of the SEI layer, electrolyte decomposition and the irreversible reaction of GeO₂ with Li. After the first cycle, the reversibility of the GeO₂/nanocable significantly improved and the coulombic efficiency increased up to 97% after the second cycle. Fig. 4c compares the cycling performance of the GeO₂/nanocable and the GeO₂/CNF electrodes at a current density of 200 mA h g⁻¹. For the GeO₂/nanocable electrode, the capacity stabilized at above 819 mA h g⁻¹ after 100 cycles. The capacity loss between the 1st and 100th cycles was only 9%, thus showing the superior cyclability of GeO₂/nanocable (calculated based on the reversible charge capacities). In contrast, the declining capacity plots of the GeO₂/CNF electrode indicates its poor cycling performance. In fact, the GeO₂/CNF electrode showed a capacity retention of only 12.5% with a final reversible capacity of 110 mA h g⁻¹. The excellent structural strength and flexibility of graphene led to good cycling stability of the GeO₂/nanocable electrode, and this assumption could be further verified by the SEM images that obtained at the end of cycles (see ESI, Fig. S3^†). Fig. S3^† compares the SEM images of both electrodes after 100 cycles. From these images, it is clear that most of the GeO₂/nanocables maintain their original 1D structures, while GeO₂/CNF shows obvious fracture phenomena.As shown in Fig. 4d, the rate capacities of GeO₂/nanocable electrodes were also tested. The performed current increased over every 5 cycles in step from 200 mA h g⁻¹ to 5000 mA h g⁻¹ and back to 200 mA h g⁻¹ at the last 5 cycles. At the currents of 200, 1000, 2000, 3000 and 5000 mA h g⁻¹, the corresponding reversible charge capacities were approximately 890, 825, 760, 690 and 595 mA h g⁻¹, respectively. When the specific current was returned back to 200 mA h g⁻¹, the capacity rose to 865 mA h g⁻¹, which is very close to the initial charge capacity. These results demonstrate that the GeO₂/nanocable electrode exhibits good tolerance to variable charge/discharge currents, which is an important characteristic required for high-power applications. Since the rate capability is dominated by the kinetics of lithium-ion diffusion and electronic conductivity, the better electrochemical performance of the GeO₂/nanocable electrode was further verified using EIS measurements with a GeO₂/CNF electrode for comparison. As shown in Fig. 4e, the EIS plots consisted of a semicircle at medium to high frequency and a straight line at low frequency. The inset of Fig. 4e shows the Randles equivalent electrical circuit model of both electrodes, it can be observed that the experimental data could be well fitted using the equivalent circuit model. As is shown, the GeO₂/nanocable electrode shows a considerably lower charge-transfer resistance (135 Ω) compared to that of the GeO₂/CNF electrode (331 Ω) (Fig. 4e and ESI Table S1^†), indicating a faster charge-transfer reaction for the GeO₂/nanocable anode.⁴⁶ This would lead to a good rate capability of the GeO₂/nanocable electrode.     

Pd nanoparticle supported reduced graphene oxide and its excellent catalytic activity for the Ullmann C–C coupling reaction in a green solvent

An efficient and easy route to synthesize reduced graphene oxide with well dispersed palladium (Pd) nanoparticles (Pd(0)-RGO) is described. The synthesized materials were fully characterized by different techniques such as: XRD, FTIR, Raman, SEM, and TEM. An average particle size of 7.5 nm for the metal particles was confirmed by TEM analysis. Pd(0)-RGO demonstrated outstanding catalytic activity for Ullmann coupling with 97% yield and good reusability (4 cycles).

An efficient and easy route to synthesize reduced graphene oxide with well dispersed palladium (Pd) nanoparticles (Pd(0)-RGO) is described.

Nanoparticles (NPs) proved to be efficient materials with wide applications in the fields of energy, environment, fine chemical synthesis, adsorption and sensors.^1–3 Nano catalysts are more efficacious than traditional catalysts because of their higher surface to volume ratio, as well as increased number of active sites.^4,5 Among all the nano catalysts, palladium and palladium based nanoparticles with many applications have gained importance in the last decade.^6–8 Various types of Pd nanoparticle have been employed as catalysts for different coupling reactions. Due to the recovery and reusability limitations of pure nanoparticles, the use of supported Pd nanoparticles is more cost effective and eco-friendly.Over the period, numerous carbon based materials such as activated carbon, carbon nanotubes, graphite and graphene have been explored as supports appropriate to the catalyst loading. Among the carbon supports, reduced graphene oxide has evolved as attractive option to be active support, due to its remarkable properties like optical properties, high surface area and electrical conductivity.^9–17 Graphene and graphene oxide (GO) have also gained importance as prospective support materials for palladium-catalysed C–C coupling reactions.¹⁸ Ullmann C–C coupling reaction, involving two aryl halides yielding biphenyl as a selective product, has attracted researchers'' attention in the recent past. Wang et al. reported that palladium modified ordered mesoporous carbon (Pd/OMC) as catalyst gave 43% biphenyl yield at 100 °C in water medium at 6 h.¹⁹ Yuan et al. reported excellent yields (96%) towards C–C coupling reaction at 80 °C in 20 h using Pd/MIL-101 as catalyst.²⁰ Liyu et al. showed that MOF-253·0.05PdCl₂ as catalyst, the reaction gives 99% yield in DMSO/EtOH (20 : 1) at 120 °C in 10 h.²¹ Karimi et al., obtained 95% yield of biphenyl by using Au supported mesoporous silica at 100 °C for 16 h.²² Varadwaj et al., obtained 96% yield of biphenyl in water medium at 80 °C in 6 h, employing Pd(0) nanoparticles supported organ functionalized clay.²³In this communication, we describe a facile and efficient route for synthesis of Pd nanoparticles supported on reduced graphene oxide and its efficacy as catalyst for Ullmann reaction in water with exceptional yields (97%). Reusability test confirms that the material is perpetual and recyclable up to four cycles.The graphene oxide was prepared according to modified Hummers'' method.²⁴ For the preparation of Pd reduced graphene oxide (Pd(0)-RGO) catalyst: 1.0 g of GO and 50 ml of distilled water was taken in a flask and sonicated for 30 min. Then palladium nitrate was added in the solution with GO, to prepare 5 and 7 wt% of Pd loaded materials. The mixture was stirred for 2 h. Then, 12 mmol of NaBH₄ with tetrahydrofuran (10 ml) solvent was added to the mixture, which was constantly stirred for 1 h. The material was filtered and washed, followed by drying at 100 °C overnight in a vacuum oven to obtain 5-Pd(0)-RGO and 7-Pd(0)-RGO materials. Fig. 1 illustrates the XRD spectra of GO (a) and 5-Pd(0)@RGO (b). In Fig. 1(a), the spectrum represents 2θ ≈ 10.75 corresponding to (002) plane of GO.^25,26 In case of the Pd(0) metal modified graphene oxide material converted to Pd(0) reduced graphene oxide [Fig. 1(b)], plane (002) at 2θ ≈ 23.11 is due to the reducing agent.²⁷ The angle at 2θ ≈ 40.11 and 46.79 correspond to (111), (200) planes of Pd metal particles.²⁶ This confirms the presence of Pd(0) metal particles on the RGO surface.Open in a separate windowFig. 1XRD spectra of GO (a) and 5-Pd(0)-RGO (b) samples.The stretching and bending frequencies of the FT-IR spectra of GO (a) and 5-Pd(0)-RGO (b) samples are illustrated in ESI Fig. S1.^† In these spectra, 3400 cm⁻¹, 1740 cm⁻¹ and 1385 cm⁻¹ represent the O–H stretching, O–H bending vibration of C–OH and C O stretching of –COOH groups respectively, which were clearly attributed to graphene oxide material.²⁸ After modification of Pd metal on the GO surface, maximum number of functional groups disappeared, which is because of the reducing agent. Fig. 2 illustrates the Raman spectra of GO (a) and 5-Pd(0)-RGO (b) samples. In the Raman spectra, all the samples showed the characteristic D-bands at 1342 cm⁻¹ and G-bands at 1595 cm⁻¹.²⁹ The intensity of the I_D/I_G ratio of normal GO sample is 0.63, but in case of 5-Pd(0)-RGO sample, the intensity of the I_D/I_G ratio increased to 0.69.Open in a separate windowFig. 2Raman spectra of GO (a) and 5-Pd@RGO (b) samples.The SEM, TEM and particle size distribution images of 5-Pd(0)-RGO sample are shown in Fig. 3. The SEM and TEM images give the details about the layered sheets of the Pd(0)-RGO sample. The uniform distribution of Pd nanoparticles on the RGO surface was confirmed by transmitted electron microscope monograph. The average particle size of the nanoparticles was 7.5 nm as calculated from TEM image (Fig. 3(d)).Open in a separate windowFig. 3SEM image of GO, scale bar = 2 μm (a), 5-Pd(0)-RGO, scale bar = 1 μm (b) and TEM image of GO, scale bar = 200 nm (c), 5-Pd(0)-RGO, scale bar = 200 nm (d) catalyst.The SEM/EDX analysis provides information on elements present on the material. Fig. 4 illustrates the SEM/EDX and colour mapping images of 5-Pd(0)-RGO catalyst. The images validate the presence of Pd, C and O on the catalyst, which is also highlighted through their colour mapping. Fig. S2, in the ESI^† shows the binding energy of Pd. Binding energy of Pd 3d^5/2 and Pd 3d^3/2 were 335.7 eV and 341.08 eV, which represent the zero-oxidation state of Pd metal. The exact amount of the Pd metal loaded on the support surface was confirmed by ICP-MS analysis indicating the Pd content in materials was 4.5 wt% and 5.1 wt% respectively.Open in a separate windowFig. 4SEM/EDX with color mapping of 5-Pd(0)-RGO, scale bar = 1 μm catalyst.Ullmann C–C coupling is one of the valuable procedures to produce biaryls and biaryls derivatives. C–C coupling reactions are known to be accelerated by various Pd-based catalysts together with organic solvents³⁰ and aqueous inorganic bases.³¹ As reported by Li et al., Pd/Ph-SBA-15 catalyst gave 75% yield towards coupling product at 100 °C for 10 h.³² Wan et al. reported that, silica-carbon supported palladium catalyst gave 64% yields towards coupling product and also the reaction was performed under water medium for 6 h.³³ Gadda et al. reported 46% conversion and 91% selectivity towards biphenyl at 150 °C in water medium with Pd/C catalyst for Ullmann coupling reaction.³⁴ The inherent drawbacks in these reports were essentially long reaction times and high temperature requirement, which have negative impact both fiscally and environmentally.We report an efficient Ullmann C–C coupling reaction of two molecules of iodobenzene with excellent yields, using potassium carbonate as base and Pd(0)-RGO as catalyst. No reaction was observed in absence of catalyst. In the preliminary studies, when the coupling reaction was performed in presence of graphene oxide (GO) for 5 h at 80 °C, reaction gave 4% yield. Results with reduced graphene oxide (RGO) and different wt% of Pd(0) modified RGO as catalysts, under similar conditions are summarized in

Correction: Dipyrrolyl-bis-sulfonamide chromophore based probe for anion recognition

Correction for ‘Dipyrrolyl-bis-sulfonamide chromophore based probe for anion recognition’ by Namdev V. Ghule et al., RSC Adv., 2014, 4, 27112–27115, DOI: 10.1039/C4RA04000G.

The authors regret that an incorrect version of Fig. 1 was included in the original article. The correct version of Fig. 1 is presented below.Open in a separate windowFig. 1Color changes of receptor DPBS in chloroform upon addition of 5 equiv. of F⁻, Cl⁻, Br⁻, I⁻, H₂PO₄⁻, HSO₄⁻, ClO₄⁻ and AcO⁻ (tetrabutylammonium salts).The Royal Society of Chemistry apologises for these errors and any consequent inconvenience to authors and readers.     

In situ approach of cementite nanoparticles encapsulated with nitrogen-doped graphitic shells as anode nanomaterials for Li-ion and Na-ion batteries

Novel Fe₃C nanoparticles encapsulated with nitrogen-doped graphitic shells were synthesized by floating catalytic pyrolysis. Due to the short synthesis time and controllable pyrolytic temperature, the diameters of Fe₃C core nanoparticles ranged from 5 to 15 nm (Fe₃C@NGS900 prepared at 900 °C) and the average thickness of N-doped graphitic shells was ∼1.2 nm, leading to their high electrochemical performance: specific capacity of 1300 mA h g⁻¹ at current density 0.2 A g⁻¹, outstanding rate capability of 939 mA h g⁻¹ at 3 A g⁻¹, improved initial coulombic efficiency (Fe₃C@NGS900: 72.1% vs. NGS900 (pure graphitic shells): 52%) for lithium ion batteries (LIBs), and impressive long-term cycle performance (1399 mA h g⁻¹ maintained at 3 A g⁻¹ after 500 cycles for LIBs; 214 mA h g⁻¹ maintained at 1 A g⁻¹ after 500 cycles for sodium ion batteries).

Novel Fe₃C nanoparticles encapsulated with nitrogen-doped graphitic shells were synthesized by floating catalytic pyrolysis.

Because of the fast development of portable electronic devices and hybrid electric vehicles, lithium ion batteries (LIBs)^1–8 and sodium ion batteries (NIBs)^9–16 with high energy/power density, good cycling performance, and lack of memory effects are in ever-increasing need. Due to the low theoretical capacity of carbon materials (graphite: 372 mA h g⁻¹),^8,17–22 optimizing the morphology of graphitic electrode materials has been important to improve specific capacity.^1–7,23,24 On the other hand, chemical doping (e.g., N, S, B, P) is an effective strategy to raise their specific capacity by increasing conductivity or active sites for Li⁺ or Na⁺ storage.^3,4,25–28 Additionally, metallic compounds (e.g., Fe₃C) have been also introduced into improving electrochemical performance of carbon anodes, because such materials are proposed to activate some components for reversible transformation of the solid electrolyte interface (SEI) and further benefit reversible capacity.^13,16,29,30 Unfortunately, most of them have been prepared by complex methods including tedious synthetic steps or long-time annealing,^6,17–19 from which, it is hard to prepare Fe₃C particles with desirable small sizes.^8,13,15,16 Thus, developing appropriate Fe₃C/C electrode materials still requires further research.In this work, Fe₃C nanoparticles encapsulated with nitrogen-doped graphitic shells (Fe₃C@NGS) were in situ approached from floating catalytic pyrolysis. Due to the short annealing time of the pyrolysis, Fe₃C@NGSs was prepared with controllable sizes. Furthermore, the in situ approach led the graphitic shells just grew on the surface of Fe₃C core nanoparticles, which improved electron transfer between the cores and the shells. Thus, such nanoparticles might be a superb electrode material towards high performance applications of LIBs and NIBs.For preparing the Fe₃C@NGSs with controllable sizes, floating catalytic pyrolysis was carried out to shorten synthetic time: the gas mixture was introduced into quartz pipe furnace, which was set at 700–1100 °C. For a typical experiment, nitrogen (flow rate: 80 L h⁻¹), acetylene (10 mL min⁻¹) and ammonia (100 mL min⁻¹) gases were embedded into iron pentacarbonyl held at 10 °C to form the gas mixture. After the pyrolysis, Fe₃C@NGS was collected at the other end of the quartz pipe. The details of materials characterization and electrochemical measurements can be found in ESI.^†As shown in transmission electron microscope (TEM) images of the prepared Fe₃C@NGSs (Fig. 1a–c and S1a of ESI^†), the metallic cores (dark section) are encapsulated with their shell (light section). The XRD results (Fig. 1d) shows the metallic core nanoparticles are Fe₃C. Thus, the sample prepared at 900 and 1100 °C have been marked with Fe₃C@NGS900 and Fe₃C@NGS1100, respectively. However, the sample prepared at 700 °C, which has been marked with FN@NGS700, has been oxidized in air at room temperature, because of its poor graphitic layers. Compared with XRD pattern of Fe₃C@NGS900, those peaks of Fe₃C@NGS1100 are much sharper indicating much bigger ferrous cores of Fe₃C@NGS1100. According to the TEM and high resolution TEM (HRTEM) images for Fe₃C@NGS900 (Fig. 1a–c, S1a, and b^†), the diameter of the core nanoparticles is ranged from 5–15 nm and the average thickness of their shells is ∼1.2 nm, respectively. Moreover, the spacing of the lattice fringes (Fig. S1a and b^†) is ∼0.34 nm corresponding to the characteristic (002) peak of graphite implying high graphitization of those shells.^25,28,31 In the HRTEM images, every Fe₃C cores is found to be a single crystal and encapsulated with the graphitic shell. The boundary between Fe₃C cores and N-doped graphitic shells is continuous and distinguished, indicating Fe₃C cores are tightly encapsulated with the graphitic shell.Open in a separate windowFig. 1(a) TEM images of core-shells nanoparticles (FN@NGS700) prepared at 700 °C, TEM and HRTEM (inset) images of nanoparticles (Fe₃C@NGS900) prepared at 900 °C (b) and nanoparticles (Fe₃C@NGS1100) prepared at 1100 °C (c) and (d) XRD patterns of core-shells nanoparticles prepared at different temperatures.Fe₃C@NGS samples have been also analyzed by X-ray photoelectron spectroscopy (XPS), which suggests Fe₃C@NGS samples contain Fe, C, O and N atoms (FN@NGS700: C content of 16.3 wt%, Fe content of 54.6 wt%, N content of 1.6 wt% and O content of 27.5 wt%; Fe₃C@NGS900: C content of 26 wt%, Fe content of 67 wt%, N content of 2 wt% and O content of 5 wt%; Fe₃C@NGS1100: C content of 33.7 wt%, Fe content of 52 wt%, N content of 2.3 wt% and O content of 7 wt%). According to XPS results of Fe₃C@NGS900, the weight ratio of Fe₃C cores and graphitic shells is 3.39 : 1. Additionally, the element contents of Fe₃C@NGS samples are different due to their morphology and structure (Fig. 1). Higher O content of FN@NGS700 is because its ferrous cores have been oxidized, and thick-walled graphitic shells (∼3 nm) of Fe₃C@NGS1100 leads to its higher C content, as XPS can only measure element contents of surface of samples.The electrochemical properties of the Fe₃C@NGS-based electrodes have been shown in Fig. 2 and ​and3.3. Since FN@NGS700 sample has been completely oxidized, electrochemical properties of ferrous oxide has not been measured. The cyclic voltammetry (CV) curves of Fe₃C@NGS900-based electrode show details of possible lithium storage process. Besides the similar peak (0.5 V) in the initial cycle for the SEI formation, two reduction peaks are located at 0.7 and ∼1.5 V, corresponding to the reduction of some SEI components (Li⁺ insertion). During the Li⁺ extraction process, the corresponding oxidation peaks is found to shift from ∼1.7 to ∼1.9 V. The almost overlapped oxidation peaks demonstrate good reversibility and cycling stability of core–shell Fe₃C@NGS.¹⁵ As shown in Fig. 2b, the galvanostatic discharge–charge (GDC) profiles of Fe₃C@NGS900 in a voltage range of 0.005–3 V (vs. Li⁺/Li) exhibits the typical shape of Fe₃C@NGS-based anodes, and Fe₃C@NGS900 delivers initial charge and discharge capacities of 1246.7 and 1729.1 mA h g⁻¹, respectively. The initial columbic efficiency (CE) reaches up to 72.1% which is higher than 52% of pure graphitic shells.³ For Fe₃C@NGSs electrode, the phenomenon of capacity increment is related to growing reversibly SEI film via the decomposition of electrolyte due to the catalysis of Fe₃C.^16,29 From the second cycle, the shape of the discharge profiles changes with respect to that of the first cycle, which may be due to the modification of the SEI film.¹²Open in a separate windowFig. 2Electrochemical performance of prepared nanoparticles as anodes for LIBs: CV profile (a) of Fe₃C@NGS900 at a scan rate of 0.1 mV s⁻¹ between 0.01–3 V vs. Li⁺/Li for the 1st–2nd charge/discharge cycles; (b) the galvanostatic charge–discharge profiles of Fe₃C@NGS900; (c) charge–discharge cycling performance of Fe₃C@NGS core–shell nanoparticles at different current densities from 0.2 to 3 A g⁻¹ at room temperature; (d) cycling performance of Fe₃C@NGS at 3 A g⁻¹.Open in a separate windowFig. 3Electrochemical performance of the prepared nanoparticles as anodes for NIBs: (a) CV profile of Fe₃C@NGS900 at a scan rate of 0.1 mV s⁻¹ between 0.01–2.5 V vs. Na⁺/Na for the 1st–3rd charge/discharge cycles. (b) The galvanostatic charge–discharge profiles of Fe₃C@NGS900; (c) cycling performance of Fe₃C@NGS at a current density of 1 A g⁻¹, and the corresponding columbic efficiency.The excellent rate capability of Fe₃C@NGS-based anodes have been investigated by testing charge/discharge at current densities of 0.2, 0.5, 1 and 3 A g⁻¹ for every 5 cycles (Fig. 2c). At the corresponding rates, the reversible capacities are 1300, 1101, 1062 and 939 mA h g⁻¹ for Fe₃C@NGS900; 925, 721, 690 and 663 mA h g⁻¹ for Fe₃C@NGS1100, indicating the smaller size of the Fe₃C nanoparticles might enhance their electrochemical capability. Compared with reported pure graphitic shells (NGS900 prepared by removing Fe₃C cores of Fe₃C@NGS900),³ the core–shell nanoparticles (Fe₃C@NGS900) exhibits impressive rate performances (NGS900: 760 mA h g⁻¹ at 0.5 A g⁻¹, 620 mA h g⁻¹ at 1 A g⁻¹ and 340 mA h g⁻¹ at 5 A g⁻¹), which might be caused by Fe₃C, as a good conductor of electricity, can effectively improve electrical conductivity of carbon electrode material.¹³ Calculated from eqn (1) of ESI,^† specific capacity of Fe₃C cores of Fe₃C@NGS900 can be evaluated (1199 mA h g⁻¹ at 0.5 A g⁻¹ and 1193 mA h g⁻¹ at 1 A g⁻¹, respectively). Based on the conversion mechanism for lithium storage, if possible, Fe₃C can store only 1/6 Li per unit (∼26 mA h g⁻¹),¹⁵ which is negligible regarding to the high capacity of ∼1300 mA h g⁻¹. The specific capacity of Fe₃C is larger than what it should be, which might be resulted from the pseudocapacity on the interface between the material and the electrolyte.¹⁶ For evaluating N doping structure in the graphitic shells, Fe₃C@NGS samples have been prepared with different percent of doping content at 900 °C by introducing ammonia with different flow rates (0, 30, 100 or 500 mL min⁻¹). As a result, Fe₃C@GS900 prepared without ammonia has graphitic shells without N-doping leading to its poor electrochemical performance: at current densities of 0.2, 0.5, 1 and 3 A g⁻¹, its reversible capacities are 575, 492, 458 and 402 mA h g⁻¹ (Fig. S4^†); the performances of Fe₃C@NGS900 (ammonia flow rate: 100 mL min⁻¹; content of N: 2 wt%) and Fe₃C@NGS900A (ammonia flow rate: 30 mL min⁻¹; content of N: 1.5 wt%) are similar; the sample prepared with ammonia flow rates of 500 mL min⁻¹ (FN@NGS900B) was violent oxidized to ferrous oxide in the air. Hence, N doping structure in graphitic shells has been confirmed to enhance diffusion.The long-term cycling performance of these two electrodes also has been investigated in Fig. 2d. The Fe₃C@NGS-based anode exhibits a favorable reversible capacity, which can reach 1399 mA h g⁻¹ after 500 discharge/charge cycles at 3.0 A g⁻¹, showing high capacity retention with CE of ∼100%. The capacities of Fe₃C samples increases with cycle number rising (from 120 to 450), which might be attributed to the pseducapacity presented by the Fe₃C.¹⁶To better study the kinetic properties of Fe₃C@NGS900 and Fe₃C@NGS1100, Fig. S3 of ESI^† shows the Nyquist plots and equivalent circuit obtained from electrochemical impedance spectroscopy (EIS) measurements. Here, R_s represents the ohmic resistance of the battery. Constant phase element (CPE) represents the double layer capacitive reactance between the electrode materials and the electrolyte. The semicircles and straight lines correspond to the electrochemical polarization impedance (R_p) and Warburg resistance (W), respectively.³² Both the fitted R_s value (5.526 Ω) and R_p value (20.826 Ω) for Fe₃C@NGS900 electrode is much lower than that Fe₃C@NGS1100 electrode (R_s: 8.501 Ω; R_p: 43.665 Ω), indicating the superior redox kinetics in the Fe₃C@NGS900 composite.In order to study the electrochemical properties of the Fe₃C@NGSs electrode as anodes for NIBs, CV analysis has been carried out at a scanning rate of 0.1 mV s⁻¹ between 0.005 and 2.5 V vs. Na⁺/Na. As shown in Fig. 3a, there are two irreversible reduction peaks around 2.02 V and 0.83 V found during the initial cathodic scan, which could be ascribed to the interaction of Na ions with specific functional groups and the decomposition of electrolyte along with the formation of SEI film on the electrode surfaces.^33,34 For the subsequent cycles, the peak at 2.02 V disappears and the peak at 0.83 V shifts to 0.60 V. For the anodic scan, the main oxidation peak ranging from about 0.47 V to 1.28 V is supposed to be the desodiation reactions.³⁵Fig. 3b depicts the GDC curves of the Fe₃C@NGS900-based electrode for the 1st–5th cycle at current density of 0.1 A g⁻¹. The large irreversible capacity in the 1st cycle is attributed to the SEI formation and the irreversible insertion of sodium ion with a relatively large ionic radius.^22,23 Following the first cycle, the charge–discharge curves become more linear which exhibits a higher and more stable CE indicating a stable SEI layer formed in the first cycle.Meanwhile, cycling performance of Fe₃C@NGS-based anodes for NIBs have been shown in Fig. 3c. In the extended cycling test at 1 A g⁻¹, a reversible capacity 214 mA h g⁻¹ of Fe₃C@NGS900 electrode is still maintained after 500 cycles, which is ∼2 times the capacity delivered by the Fe₃C@NGS1100 indicating excellent cycling stability of Fe₃C@NGS900. Thus, Fe₃C@NGS-base anodes holds great potential as a promising candidate compared with other carbonaceous anode materials for NIBs (113 mA h g⁻¹ at 1 A g⁻¹ (modified PFR/C),³⁶ 188.6 mA h g⁻¹ at 0.1 A g⁻¹ after 300 cycles (S/C),³³ 150 mA h g⁻¹ at 1 A g⁻¹ after 200 cycles (S/graphene)³⁵).It is noticed that the excellent electrochemical performance of the prepared Fe₃C@NGSs is apparently ascribed to their novel structure. First, because of in situ growing graphitic shells on the surface of Fe₃C during floating catalytic pyrolysis, contact between Fe₃C cores and graphitic shells effectively increases, leading to lower contact resistance and faster electron transfer between the cores and the shells, comparing with traditional ferrous/carbon composites generated by multistep carbonization approach.^8,13,15,16 Second, due to floating catalytic pyrolysis, sizes of Fe₃C core nanoparticles have been under control (Fe₃C@NGS900: 5–15 nm vs. Fe₃C@NGS1100: 15–40 nm). Smaller size of Fe₃C core nanoparticles might increase surface of Fe₃C nanoparticles, leading to distinguished improvement of their electrochemical performance (Fig. 2 and ​and3),3), due to active sites for Li⁺ or Na⁺ storage rising. For comparison, sizes of reported Fe₃C composites have been listed: Fe₃C@C: 60 nm nanoparticles encapsulated with 4 nm carbon shells;⁸ Fe₃C@PC: 29 ± 5 nm nanoparticles embedded in 300 nm porous carbon;¹⁶ Fe₃C/C: 300 nm;¹³ Fe@Fe₃C/C: 28–58 nm Fe@Fe₃C nanoparticles.¹⁵ Third, introducing Fe₃C to electrode material can promote the reversible formation/decomposition of the SEI film, causing improvement of initial CE (72.1% vs. 52%) for LIBs, due to the catalysis function of Fe₃C.^16,29 Fourth, due to in situ N-doping (∼2 wt%) during their floating catalytic pyrolysis, such defects of the graphitic shells might offer lots channels for fast diffusion of electrolyte and Li⁺/Na⁺ into those nanoparticles. Fifth, such prepared core–shell nanoparticles have ultra thin-walled graphitic shells (∼1.2 nm shown in Fig. 1d), which shorten the diffusion route of ions and electrolyte. All above confirm Fe₃C@NGS900 has novel structure towards the electrochemical applications, compared with the reported works: Fe@Fe₃C/C sample was prepared by sol–gel and carbonization approach, and whether its pure Fe cores was good for Li⁺ storage was doubtful;¹⁵ the Fe₃C@C nanoparticles were prepared with no doping carbon shells;⁸ the graphitization of carbon structure was doubtful, when ferrous/carbon composites were prepared by polymerization-carbonization of iron phthalocyanine¹³ or hydrothermal method-carbonization.¹⁶ Thus, the Fe₃C@NGS900 performs better (1300 mA h g⁻¹ at current density of 0.2 A g⁻¹; 939 mA h g⁻¹ at 3 A g⁻¹) than many reported ferrous/carbon composite anode materials for LIBs (0.2 A g⁻¹: <382 (Fe@Fe₃C/C),¹⁵ ∼480 (Fe₃C/C),⁸ 787.9 (Fe₂O₃/C),⁵ ∼850 (Fe₃O₄/C),¹⁹ 873 (N,S/C),³⁶ and 881 (Fe₃O₄/C)⁹ mA h g⁻¹; 3 A g⁻¹: ∼300 (Fe₂O₃@C),¹⁸ ∼370 (FeS@C)²¹ and ∼612 (Fe₂O₃/C)¹⁴ mA h g⁻¹).     

Coaxial heterojunction carbon nanofibers with charge transport and electrocatalytic reduction phases for high performance dye-sensitized solar cells

Novel coaxial heterojunction carbon nanofibers, fabricated by electro-spinning a mixture of hydro-pitch and polyacrylonitrile, served as the counter electrode for dye-sensitized solar cells. Their high power conversion efficiency, being comparable to that of Pt CE, was achieved due to their good conductivity and high heteroatom content.

Coaxial CNFs featured high conductivity derived from HP, and high N content and defective sites derived from PAN.

The counter electrode (CE) in dye-sensitized solar cells (DSSCs) has multiple functions, including the catalytic reduction of I₃⁻ to I⁻ and the regeneration of dye molecules, and is one of the key dominant components that governs the practical applications of DSSCs to a great degree.¹ Noble metal platinum (Pt) with its low electrical resistance and excellent electrocatalytic activity was the first to be applied as such and is now widely used as CEs in DSSCs. However, there is limited availability of Pt sources, and this has led to the high cost of CEs, which has hindered the practical applications of DSSCs.² To date, various efforts have been made to develop techniques for the production of inexpensive yet high performance CEs to replace the Pt CEs. Of the alternatives now available, carbon materials that are categorized as zero-dimensional (0D) to three-dimensional (3D) structures have attracted much attention as candidates for Pt-free electrodes in view of the advantages of low cost as well as good electrochemical stability; these include carbon black,³ carbon nanoparticles,⁴ carbon nanotubes,^5–9 graphene,^10–14 and graphite¹⁵ and their composites.^16–19 It is believed that a large quantity of defects in the carbon CEs may lead to high catalytic activities, while good electric conductivity can result in fast charge transportation. Nevertheless, how to combine and integrate the high electrical conductivity and abundant defects into one carbon electrode in a balanced way to fabricate high performance CEs with tuned structure remains a major challenge.Recently, one dimensional (1D) core–shell nanofibers have received much attention because a good combination of electrical conductivity and catalytic activity can be realized and achieved in one electrode material.²⁰ However, the core and shell in these fibers are two separate phases, which limits the rapid charge transition and increases the electrical resistance when such nanofibers are used as CEs in DSSCs. Moreover, the tedious and complicated fabrication process greatly limits their large-scale production. As such, it is necessary to explore a new approach for this kind of material, in which electrical conductivity and catalytic activity are combined well.Herein, we present 1D coaxial carbon nanofibers (CNFs) fabricated by the electrospinning method from two kinds of carbon precursors: hydrogenated pitch and polyacrylonitrile (PAN). The adopted hydro-pitch (HP) features planar aromatic hydrocarbon molecules and is more easily transformed into an optically anisotropic, graphitizable carbon structure,²¹ thus leading to high conductivity (Table S1, ESI^†). PAN tends to form a carbonaceous structure with high nitrogen (N) content and abundant defects; it is widely acknowledged that high electrochemical activities could be achieved by the different nitrogen species and defects.^{7,16,17,22–24} Benefiting from different nucleation mechanisms, a novel coaxial core/shell structure has been produced within the matrix, in which hydro-pitch contributes to the core phase, while PAN is responsible for the shell phase. Such CNFs with two phases and heterogeneous characteristics are very unique when they are applied as CEs in DSSCs, with the core playing the role of cable tunnel for charge transporting and the shell providing active sites for catalyzing the reduction of I₃⁻ to I⁻, as shown in Scheme 1.Open in a separate windowScheme 1Illustration of the structure and functions of heterojunction carbon nanofibers as CEs in DSSCs.The process of fabricating our CNFs (denoted as 0.15-HCNF and 1-HCNF, in which 0.15 and 1 are the weight ratios between hydro-pitch and PAN) is illustrated in Fig. S5,^† and the morphologies of different CNFs and their corresponding diameter distributions are shown in Fig. 1a–c and a₁–c₁. The average diameter is 0.33 μm for PAN CNFs (PCNF), 0.36 μm for 0.15-HCNF, and 0.76 μm for 1-HCNF, indicative of a size-increased behavior with the increase in the ratio between the hydro-pitch and PAN. The obtained CNFs in sample 1-HCNF are cross-linked or joined together, as can be seen in Fig. 1c, leading to the inner parts being exposed and the observation of different phases.Open in a separate windowFig. 1Scanning electron microscopy (SEM) images of (a) PCNF, (b) 0.15-HCNF and (c) 1-HCNF.The HCNFs were examined in much detail by transmission electron microscopy (TEM) to yield information about the inner structure, and the typical TEM images are shown in Fig. 2. It can be clearly seen that two phases in the axial direction are obviously observed in the HCNF samples (Fig. 2b and c) in comparison to PCNF (Fig. 2a). Energy-dispersive X-ray spectra (EDX) of the shell region show the high nitrogen content of 9.3 at% (Fig. 2e), which implies that the carbon shell is derived from PAN. To further gain insight into the contributions of these two precursors, sample 1-HCNF was treated with KOH at high temperature to remove the shell, and the corresponding sample was further characterized by TEM and EDX, the detailed results of which are shown in Fig. 2d and f. It can be seen that the shell layer was removed, and the content of nitrogen was only 1 at% in the core phase (Table S2, ESI^†), in comparison to there being only 0.4 wt% (Table S3, ESI^†) of nitrogen in the hydro-pitch.Open in a separate windowFig. 2TEM images of (a) PCNF, (b) 0.15-HCNF, (c) 1-HCNF, and (d) the core of the 1-HCNF fiber after KOH treatment for 1 h at 700 °C in N₂. (e and f) The EDX spectra of the fiber shell and core.With all of this information in mind, it was deduced that the core phase was mainly derived from the hydro-pitch, while the shell was mainly from PAN.It is interesting that this result is different from that reported by Yang.²⁵ In general, the low molecular weight pitch tends to be pushed to the outer surface during the solvent evaporation process.^26,27 Nevertheless, in the present system the hydro-pitch is more easily fixed in the inner part, which was evidenced by density functional theory (DFT) calculations (Fig. 3). As shown in Fig. 3a, the binding energy between the hydro-pitch and two PAN units (E_HP) is 5.6 kcal mol⁻¹, which is 1.3 kcal mol⁻¹ higher than that between pitch and PAN units (E_P, 4.3 kcal mol⁻¹). When three PAN units were applied, an increase of 25% was observed for E_HP, which reached 7.0 kcal mol⁻¹. The newly formed hydrogen bonding between aliphatic hydrogen in hydro-pitch and the cyano group in the PAN unit is responsible for the remarkable increase in binding energy, while the number of E_P only slightly increased from 4.3 to 4.4 kcal mol⁻¹, and no new bonding was observed. As a result, compared to the pitch, the hydro-pitch could be more easily fixed in PAN units due to the strong binding caused by the multi-interaction between the aliphatic hydrogen and cyano groups. In addition, the hydro-pitch is always wrapped by several PAN long chain molecules, as shown in Fig. S5.^† Therefore, a kind of heterojunction core–shell structure was finally produced, in which the core was attributed to the hydro-pitch while the shell was attributed to PAN.Open in a separate windowFig. 3The spatial configurations and binding energies between two kinds of pitch molecules and PAN units. (a) Hydro-pitch with two PAN units, (b) molecular pitch and two PAN units, (c) hydro-pitch and three PAN units, (d) molecular pitch and three PAN units.The HCNF samples were also analyzed by X-ray photoelectron spectroscopy (XPS) and Raman spectra to reveal the changes in the surface configuration caused by the addition of hydro-pitch (Fig. S2, ESI^†). Compared to PCNF (7.6 at%, Table S4, ESI^†), sample 0.15-HCNF maintained 4.6 at% nitrogen content, and the same I_D/I_G ratio (1.07) was also obtained for the two samples. With an increase in hydro-pitch in the case of 1-HCNF, a slightly lower nitrogen content (4.2 at%) and I_D/I_G number (1.01) were observed, which could be attributed to the exposure of the hydro-pitch-based core, as mentioned in Fig. 1c. The XPS spectrum of N 1s is shown in Fig. S3 in the ESI^† and the ratios of different nitrogen species are summarized in Table S5.^† It is well known that these nitrogen species in the carbon network could produce highly electrocatalytically active sites,^22,23 while the defects in carbon materials could also play the same role.The HCNFs exhibited unique characteristics in which the shell retained the high nitrogen content and defects, while the core derived from hydro-pitch featured high conductivity. Such an integrated structure in one electrode is so attractive that it could be used as a high-performance CE in DSSCs, as shown in Scheme 1. In this case, it has great potential as an electrocatalyst for Pt replacement in DSSCs. Benefiting from this, the device performance for DSSCs with PCNF, Pt, and two HCNF CEs was determined and the results are shown in Samples V _oc/V J _sc/mA cm⁻²FF η/%0.15-HCNF0.7614.160.656.92 ± 0.151-HCNF0.7513.480.636.32 ± 0.15PCNF0.7612.250.464.26 ± 0.15Pt0.7313.280.656.34 ± 0.15Open in a separate windowAs shown in Fig. 4a, our HCNF CEs exhibit similar or even better performance compared with Pt CE, and the electrochemical performance varies with the ratio of the hydro-pitch to PAN. The sample 0.15-HCNF CE shows a higher J_sc of 14.16 mA cm⁻², a fill factor (FF) of 0.65 and a power conversion efficiency of 6.92%. In contrast, the J_sc, FF and power conversion efficiency (η) for the DSSCs with PCNF CE are only 12.25 mA cm⁻², 0.46 and 4.26%, respectively; in particular, an increase in efficiency of 62.4% was observed. Nevertheless, when the ratio of hydro-pitch and PAN increased to 1, the performance of DSSCs with 1-HCNF electrode was reduced to J_sc of 13.48 mA cm⁻², FF of 0.65 and η of 6.32%. The reason for this is attributed to the lower contents of nitrogen and defects, nitrogen and defects, and the loss of the Brunauer Emmet Teller (BET) surface (Table S6, ESI^†).Open in a separate windowFig. 4(a) Current density–voltage (J–V) characteristics with different CEs: PCNF, Pt, 1-HCNF, and 0.15-HCNF. (b) Illustration of charge transport and diffusion in different CNFs.To further demonstrate the unique effects of the coaxial CNF, to further demonstrate the unique effects of the coaxial CNF, the possible mechanism involved in the process was proposed and shown in Fig. 4b. As is known, the general consensus for the reaction mechanism can be described as follows:²⁸I₃⁻ (sol) ↔ I₂ (sol) + I⁻ (sol)1I₂ (sol) + 2* → 2I*2I* + e⁻ → I⁻ (sol)3After the desorption of the solvated I⁻ (sol) into the electrolyte, the activated site lost one electron and high-rate regeneration was demanded for the active site. The electrons from the external circuit are typically transferred from FTO to CE and then transmitted in CEs to the activated catalytic sites. The key in this process is the abundant supply of electrons to the activated sites within a short diffusion time. For our HCNF CE, the 1D tunnel structure minimizes the loss of electrons for its low electrical resistance, and high-rate regeneration of the active site is also realized by direct electron diffusion from the core to the activated catalytic site, due to its heterojunction structure. Benefiting from the unique structure, highly efficient utilization of the activated catalytic sites and fast charge transfer derived from short transport distance are achieved, and high power-conversion efficiency is yielded, which is comparable to or even better than other metal-free CEs reported in the literature (Table S8^†).To further investigate the electrochemical characteristics of HCNF CEs, Nyquist plots on an asymmetric dummy were obtained for Pt, PCNF and 0.15-HCNF electrodes (Fig. S4, ESI^†).²⁹ In contrast to Pt, both kinds of CNF CEs have the same lower R_s value, 1.6 Ω cm² (Table S7, ESI^†), indicating that good adhesion between CNFs and FTO was realized. High R_ct values of 6.4 Ω cm² for PCNF and 4.3 Ω cm² for 0.15-HCNF CE (Table S7, ESI^†) were also observed. In the future, further surface modification needs to be carried out to reduce the R_ct value,^17–19 thus optimizing and improving the electrochemical performance of CNFs.In summary, the coaxial heterojunction CNFs with two phases were fabricated from PAN/hydro-pitch blend precursors using the electrospinning method, in which the shell and core were constructed by PAN based carbon and hydro-pitch based carbon, respectively. The as-made coaxial CNFs exhibited high conductivity derived from hydro-pitch, and high nitrogen content and defects derived from PAN. Such HCNFs were employed as CEs in DSSCs, and high power-conversion efficiencies were delivered, which are comparable to that of the Pt CE constructed under the same conditions. The tunnel along the axis direction separates the fibers into the charge transport phase and the electrocatalytic phase, and high utilization of catalytic sites is realized by the abundant charge supply and fast electron–hole recombination. In addition, this work shows that low cost and large-scale production of high performance CEs in DSSCs can be realized using the electrospinning method.     

Thiophene-containing tetraphenylethene derivatives with different aggregation-induced emission (AIE) and mechanofluorochromic characteristics

Four thiophene-containing tetraphenylethene derivatives were successfully synthesized and characterized. All these highly fluorescent compounds showed typical aggregation-induced emission (AIE) characteristics and emitted different fluorescence colors including blue-green, green, yellow and orange in the aggregation state. In addition, these luminogens also exhibited various mechanofluorochromic phenomena.

Four thiophene-containing AIE-active TPE derivatives were synthesized. Furthermore, these luminogens exhibited various mechanofluorochromic phenomena.

High-efficiency organic fluorescent materials have attracted widespread attention due to their potential applications in organic light-emitting devices and fluorescent switches.^1–8 Meanwhile, smart materials sensitive to environmental stimuli have also aroused substantial interest. Mechanochromic luminescent materials exhibiting color changes under the action of mechanical force (such as rubbing or grinding) are one important type of stimuli-responsive smart materials, which can be used as pressure sensors and rewritable media.^9–18 Bright solid-state emission and high contrast before and after grinding are very significant for the high efficient application of mechanochromic fluorescence materials.^19–28 However, a majority of traditional emissive materials usually exhibit poor emission efficiency in the solid state due to the notorious phenomenon of aggregation caused quenching (ACQ), and the best way to solve the problem is to develop a class of novel luminescent materials oppositing to the luminophoric materials with ACQ effect. Fortunately, an unusual aggregation-induced emission (AIE) phenomenon was discovered by Tang et al. in 2001.²⁹ Indeed, the light emission of an AIE-active compound can be enhanced by aggregate formation.^30–32 Obviously, it is possible that AIE-active mechanochromic fluorescent compounds can be applied to the preparation of high-efficiency mechanofluorochromic materials. Numerous luminescent materials exhibiting mechanochromic fluorescent behavior have been discovered up to now.³³ Whereas, examples of fluorescent molecules simultaneously possessing AIE and mechanofluorochromic behaviors are still limited, and the exploitation of more AIE-active mechanofluorochromic luminogens is necessary. Organic solid emitters with twisted molecular conformation can effectively prevent the formation of ACQ effect, thus exhibiting strong solid-state luminescence. Tetraphenylethene is a highly twisted fluorophore. Meanwhile, it is also a typical AIE unit, which can be used to construct high emissive stimuli-responsive functional materials.^34–37The design and synthesis of novel organic emitters with tunable emission color has become a promising research topic at present. Only a limited number of organic fluorescent materials with full-color emission have been reported to date.^38,39 For example, in 2018, Tang et al. reported six tetraphenylpyrazine-based compounds. Interestingly, in film states, these luminogens exhibited different fluorescence colors covering the entire visible range, and this is the first example of realizing full-color emission based on the tetraphenylpyrazine unit.⁴⁰ It is still an urgent challenge to develop novel organic luminophors with tunable emission color basing on the same core structure.In this study, four organic fluorophores containing tetraphenylethene unit were successfully synthesized (Scheme 1). Introducing the thiophene and carbonyl units into the molecules possibly promoted the formation of weak intermolecular interactions such as C–H⋯S or C–H⋯O interaction, which was advantageous to the exploitation of interesting stimuli-responsive fluorescent materials. Indeed, all these compounds showed obvious AIE characteristics. Furthermore, these luminogens emitted a series of different fluorescent colors involving blue-green, green, yellow and orange in the aggregation state. In addition, these luminogens also exhibited reversible mechanofluorochromic phenomena involving different fluorescent color changes.Open in a separate windowScheme 1The molecular structures of compounds 1–4.To investigate the aggregation-induced properties of compounds 1–4, the UV-vis absorption spectra of 1, 2, 3 and 4 (20 μM) in DMF–H₂O mixtures of varying proportions were studied initially (Fig. S1^†). Obviously, level-off tails were obviously observed in the long-wavelength region as the water content increased. This interesting phenomenon is generally associated with the formation of nano-aggregates.⁴¹ Next, the photoluminescence (PL) spectra of 1–4 in DMF–H₂O mixtures with various water fraction (f_w) values were explored. As shown in Fig. 1, almost no PL signals were noticed when a diluted DMF solution of luminogen 1 was excited at 365 nm, and thus almost no fluorescence could be observed upon UV illumination at 365 nm, and the corresponding absolute fluorescence quantum yield (Φ) was as low as 0.04%. However, when the water content was increased to 50%, a new blue-green emission band with a λ_max at 501 nm was observed, and a faint blue-green fluorescence was noticed under 365 nm UV light. As the water content was further increased to 90%, a strong blue-green emission (Φ = 30.81%) could be observed. Furthermore, as shown in Fig. S2,^† the nano-aggregates (f_w = 90%) obtained were confirmed by dynamic light scattering (DLS). Therefore, the compound 1 with bright blue-green emission caused by aggregate formation showed typical AIE feature.Open in a separate windowFig. 1(a) Fluorescence spectra of the dilute solutions of compound 1 (2.0 × 10⁻⁵ mol L⁻¹) in DMF–water mixtures with different water contents (0–90%). Excitation wavelength = 365 nm. (b) Fluorescence images of 1 (2.0 × 10⁻⁵ mol L⁻¹) in DMF–water mixtures with different f_w values under 365 nm UV light.Similarly, as can be seen in Fig. 2–4, compounds 2–4 also showed obvious aggregation-induced green emission, aggregation-induced yellow emission, and aggregation-induced orange emission, respectively. When the water content was zero, the quantum yields of compounds 2–4 were 0.04%, 0.05% and 0.46%, respectively, while as the water content increased to 90%, the corresponding quantum yields of compounds 2–4 also increased to 30.67%, 45.57% and 26.53%, respectively. Hence, luminogens 2–4 were also AIE-active species. In addition, as shown in Fig. 5, the DFT calculations for the compounds 1–4 were performed. The calculated energy gaps (ΔE) of four compounds were 3.6178416 eV (compound 1), 3.276084 eV (compound 2), 3.3073755 eV (compound 3) and 3.0766347 eV (compound 4) respectively. Therefore, the various numbers and the various kinds of the substituents had slight effects on their molecular orbital energy levels of 1–4.Open in a separate windowFig. 2(a) Fluorescence spectra of the dilute solutions of compound 2 (2.0 × 10⁻⁵ mol L⁻¹) in DMF–water mixtures with different water contents (0–90%). Excitation wavelength = 365 nm. (b) Fluorescence images of 2 (2.0 × 10⁻⁵ mol L⁻¹) in DMF–water mixtures with different f_w values under 365 nm UV light.Open in a separate windowFig. 3(a) Fluorescence spectra of the dilute solutions of compound 3 (2.0 × 10⁻⁵ mol L⁻¹) in DMF–water mixtures with different water contents (0–90%). Excitation wavelength = 365 nm. (b) Fluorescence images of 3 (2.0 × 10⁻⁵ mol L⁻¹) in DMF–water mixtures with different f_w values under 365 nm UV light.Open in a separate windowFig. 4(a) Fluorescence spectra of the dilute solutions of compound 4 (2.0 × 10⁻⁵ mol L⁻¹) in DMF–water mixtures with different water contents (0–90%). Excitation wavelength = 365 nm. (b) Fluorescence images of 4 (2.0 × 10⁻⁵ mol L⁻¹) in DMF–water mixtures with different f_w values under 365 nm UV light.Open in a separate windowFig. 5(a) HOMO and LUMO frontier molecular orbitals of molecule 1 based on DFT (B3LYP/6-31G*) calculation. (b) HOMO and LUMO frontier molecular orbitals of molecule 2 based on DFT (B3LYP/6-31G*) calculation. (c) HOMO and LUMO frontier molecular orbitals of molecule 3 based on DFT (B3LYP/6-31G*) calculation. (d) HOMO and LUMO frontier molecular orbitals of molecule 4 based on DFT (B3LYP/6-31G*) calculation.Subsequently, the mechanochromic fluorescent behaviors of compounds 1–4 were surveyed by solid-state PL spectroscopy. As shown in Fig. 6, the as-synthesized powder sample 1 exhibited an emission band with a λ_max at 444 nm, corresponding to a blue fluorescence under 365 nm UV light. Intriguingly, a new blue-green light-emitting band with a λ_max at 507 nm was observed after the pristine solid sample was ground. After fuming with dichloromethane solvent vapor for 1 min, the blue-green fluorescence was converted back to the original blue fluorescence. Therefore, luminogen 1 exhibited reversible mechanochromic fluorescence feature. Furthermore, this reversible mechanofluorochromic conversion was repeated many times by grinding-exposure without showing signs of fatigue (Fig. 10).Open in a separate windowFig. 6(a) Solid-state PL spectra of compound 1 before grinding, after grinding, and after treatment with dichloromethane solvent vapor. Excitation wavelength: 365 nm. Photographic images of compound 1 under 365 nm UV light: (b) the as-synthesized powder sample. (c) The ground sample. (d) The sample after treatment with dichloromethane solvent vapor.Open in a separate windowFig. 10Repetitive experiment of mechanochromic behavior for compound 1.Similarly, as evident from Fig. 7–9, luminogens 2–4 also exhibited obvious mechanofluorochromic characteristics. Moreover, the repeatabilities of their mechanochromic behaviors were also satisfactory (Fig. S3^†). Hence, all the compounds 1–4 showed reversible mechanofluorochromic phenomena involving different fluorescent color changes, and the various numbers of the substituents could effectively influence the mechanofluorochromic behaviors of 1–4. Obviously, luminogen 3 or 4 after grinding exhibited more red-shifted fluorescence in comparison with that of the corresponding luminogen 1 or 2 after grinding.Open in a separate windowFig. 7(a) Solid-state PL spectra of compound 2 before grinding, after grinding, and after treatment with dichloromethane solvent vapor. Excitation wavelength: 365 nm. Photographic images of compound 2 under 365 nm UV light: (b) the as-synthesized powder sample. (c) The ground sample. (d) The sample after treatment with dichloromethane solvent vapor.Open in a separate windowFig. 8(a) Solid-state PL spectra of compound 3 before grinding, after grinding, and after treatment with dichloromethane solvent vapor. Excitation wavelength: 365 nm. Photographic images of compound 3 under 365 nm UV light: (b) the as-synthesized powder sample. (c) The ground sample. (d) The sample after treatment with dichloromethane solvent vapor.Open in a separate windowFig. 9(a) Solid-state PL spectra of compound 4 before grinding, after grinding, and after treatment with dichloromethane solvent vapor. Excitation wavelength: 365 nm. Photographic images of compound 4 under 365 nm UV light: (b) the as-synthesized powder sample. (c) The ground sample. (d) The sample after treatment with dichloromethane solvent vapor.In order to further explore the possible mechanism of mechanofluorochromism of 1–4, the powder X-ray diffraction (PXRD) measurements of various solid states of 1–4 were carried out. As depicted in Fig. 11, the pristine solid powder 1 showed many clear and intense reflection peaks, suggesting its crystalline phase. However, after the pristine powder sample was ground, the sharp and intense diffraction peaks vanished, which indicated the crystalline form was converted to the amorphous form. Interestingly, when the ground solid sample was fumigated with dichloromethane solvent vapor for 1 min, the corresponding sample powder exhibited the PXRD pattern of the initial crystalline form. Meanwhile, the structural transformations of the solid samples of 2–4 were similar to that of 1 (Fig. S4–S6^†). Obviously, the morphological changes of solid samples of 1–4 from crystalline state to amorphous state and vice versa could be attributed to the reversible mechanical switching in compounds 1–4, and the mechanofluorochromic phenomena observed in 1–4 were related to the morphological transition involving the ordered crystalline phase and the disordered amorphous phase.Open in a separate windowFig. 11XRD patterns of compound 1: unground, ground and after treatment with dichloromethane solvent vapor.Fortunately, single crystals of compounds 1 and 2 were obtained by slow diffusion of n-hexane into a trichloromethane solution containing small amounts of 1 or 2. As shown in Fig. 12 and ​and13,13, the molecular structures of 1 and 2 exhibited a twisted conformation due to the existence of tetraphenylethene unit. Meanwhile, some weak intermolecular interactions, such as C–H⋯π interaction (d = 2.866 Å) for 1, π⋯π interaction (d = 3.371 Å) for 1, C–H⋯S interaction (d = 2.977 Å) for 2, and π⋯π interaction (d = 3.189 Å) for 2, were observed. These weak intermolecular interactions gave rise to a loose packing motif of 1 or 2, which indicated their ordered crystal packings might readily collapse upon exposure to external mechanical stimulus. Therefore, their solid-state fluorescence could be adjusted by mechanical force.Open in a separate windowFig. 12The structural organization of compound 1.Open in a separate windowFig. 13The structural organization of compound 2.In summary, four fluorescent molecules containing thiophene and tetraphenylethene units were successfully designed and synthesized in this study. All these compounds showed obvious AIE characteristics. Furthermore, these luminogens emitted various fluorescence colors involving blue-green, green, yellow and orange in the aggregation state. Meanwhile, these luminogens basing on the same core structure also exhibited reversible mechanofluorochromic phenomena involving different fluorescent color changes. The results of this study will be beneficial for the exploitation of novel luminophors with full-color emission.     

Novel continuous flow synthesis of Pt NPs with narrow size distribution for Pt@carbon catalysts

In this work, we report a novel continuous flow synthesis method to achieve ultra-small Pt NPs (2.3 to 2.5 nm) with narrow size distribution. This method expedited the synthesis of Pt NPs without any harsh reducing agent or capping agent. Further these Pt NPs were immobilized on a carbon support in a single step procedure for its application as an electrocatalyst for fuel cells. The synthesized Pt NPs and Pt NPs supported on carbon were analyzed using transmission electron microscopy affirming uniform distribution of Pt NPs throughout the carbon support without aggregation.

A novel continuous flow synthesis method was performed to achieve ultra-small Pt NPs (2.3 to 2.5 nm) with narrow size distribution. This method expedites the synthesis of Pt NPs without any harsh reducing agent or capping agent.

Platinum nanoparticles are widely used in the field of catalysis,^1,2 sensors,^3,4 fuel cell technology^5,6 and other integrative areas. The expanding interest in these particular nanoparticles is because of their well acknowledged properties and potential in various fields of energy especially as an electrocatalyst for fuel cell applications.^7,8 Nonetheless, their structural parameters like size,^9,10 morphology^11,12 and distribution are some of the limiting factors to their widespread application. Therefore, it is highly important to synthesize small nanoparticles with controlled structural parameters.Various researchers have synthesized Pt NPs with small size and narrow size distribution using various methods. But mostly the use of stabilizers like PVP,¹³ PEG¹⁴ and CTAB¹⁵ or harsh reducing agent like sodium borohydride¹⁶ is employed to achieve small nanoparticles without aggregation. Although, for its application in fuel cell technology where it is supported on the various carbon material^17–19 it is preferred that minimum organic compounds are used for the synthesis of Pt NPs to achieve high activity. Therefore, the preparation of these Pt@carbon catalysts require tedious post-treatment method to remove organics before it is subjected for electrochemical analysis. However, some researchers pursued alternative pathway to reduce Pt salt using homogenous deposition method and mild reducing agents like ethylene glycol in batch synthesis.^20–24 In this work a continuous flow method has been demonstrated for the synthesis of ultra-small Pt NPs.In a continuous flow method for the preparation of Pt NPs, PFA tubing was used with inner diameter of 1.59 mm and the total length of 1.8 m. The PFA tube was rolled in spiral fashion before subjecting it to the heating furnace with the two ends out. The one end of the tube was connected with glass syringe fitted in syringe pump for the reactant infusion and the colloidal solution of Pt NPs was collected from the other end of the tube (Fig. 1).Open in a separate windowFig. 1Experimental set-up showing the continuous flow synthesis of Pt NPs.The Pt NPs were synthesized by facile polyol synthesis using ethylene glycol as reducing agent and solvent. To prepare the reactant mixture H₂PtCl₆·6H₂O (0.015 M) was dissolved in ethylene glycol and water mixture in 4 : 1 ratio. Further the concentration of NaOH/Pt was maintained around 7.5 molar ratio in the EG : water mixture to control a nanoparticle size. The resulting solution was stirred for 30 min at room temperature in open air before injecting. The flow rate for the continuous flow synthesis was maintained at 3.0 mL h⁻¹ and the temperature of the heating furnace was set to 180 °C. The colloidal solution of NPs was collected in the centrifuged tube for immobilization on carbon support.Carbon black (Vulcan XC72) (CB) was pretreated by 6 M HCl solution.^25–27 And after the decantation of supernatant, the precipitate was washed thoroughly with ultrapure water (18.7 MΩ cm). The acid treated CB was then filtered and kept for overnight drying at room temperature. To prepare Pt@C material the pretreated CB (30 mg) was dispersed in water (30 mL) using ultra-sonication for 1 hour to obtain homogenous dispersion. The as-synthesized colloidal NPs solution (10 mL) was added to the carbon black solution dropwise under vigorous stirring for overnight mixing. Then 5 M HCl solution was added to the above suspension to adjust the pH ∼2 under constant stirring condition at room temperature. The Pt@C material was then filtered using membrane filter and washed with Milli-Q-water and ethanol to remove Cl⁻ ions from the material. The prepared Pt@C was dried at 50 °C overnight in vacuum oven. The synthesized Pt NPs were decorated on as received Vulcan XC72 for comparison. It was, however, observed that the deposition on without acid treated Vulcan XC72 show irregularity in the shape due to aggregation as shown in Fig. S1.^† In this process we can synthesize ∼158 mg of Pt NPs in 1 day and ∼375 mg of Pt@C catalyst in 3 days including washing and drying.The size distribution and dispersibility of the Pt NPs synthesized using continuous flow method were analyzed using transmission electron microscopy (TEM, JEOL Ltd., JEM-2100F) at 200 kV. Fig. 2 shows the TEM images and the corresponding particle size distribution histograms of Pt NPs and Pt@C catalyst. The average size of the platinum NPs was around 2.3–2.4 nm. As shown in Fig. 2(B) the Pt NPs are uniformly dispersed all over the carbon surface with narrow size distribution. And no obvious change was observed in the size of NPs after loading on to the carbon support. The dispersibility and narrow size distribution attained using continuous synthesis is much better compared to the commercial catalyst. The TEM results showed that the present continuous flow synthesis was highly favorable to achieve the well dispersed Pt NPs over the carbon support.Open in a separate windowFig. 2TEM images and size distribution histograms for (A) Pt NPs and (B) Pt@C catalyst.The amount of platinum in the prepared Pt@C material was determined by inductively coupled plasma atomic emission spectroscopy (ICP-AES, PerkinElmer Co., Ltd. Optima 2100DV). The ICP-AES sample was prepared by dissolving the Pt@C in aqua regia (1 mL) followed by heating at 170 °C for 30 min. The solution was then diluted to optimal concentration for the analysis. The Pt loading on the carbon support was also confirmed using thermogravimetric analysis (TGA, Shimadzu Co., DTG-60A) as shown in Fig. 3. The prepared Pt@C material was weighed ∼5 mg and was heated under Ar/air = ¼ (v/v) flow over the temperature range of 40 to 790 °C at 20 °C min⁻¹ for TGA analysis. The amount of Pt content in the prepared catalyst was 33.6%. A gentle weight loss observed in TGA and corresponding sharp peak observed in DTA at 195 °C originated from ethylene glycol which has a boiling point of 197.6 °C. The wide peak originating above 370 °C arises from CB in Pt@C catalyst. Most of ethylene glycol in Pt@C may present on the surface of carbon not on Pt nanoparticles since the TGA curve of pure Pt NPs shows only ∼0.5 wt% loss at 195 °C (see Fig. S2^†).Open in a separate windowFig. 3TGA and DTA curves for Pt@C catalyst. Fig. 4 shows the XRD pattern of Pt@C catalyst synthesized by continuous flow synthesis. The four strong peaks at 2 theta values 39.9, 46.5, 67.8, and 81.2° corresponds to the [111], [200], [220], [311] crystal plane of Pt respectively revealing the fcc structure. The presence of fcc structure planes showed the crystalline nature of Pt NPs decorated over carbon. The Pt [111] plane was used to evaluate the particle size of Pt NPs by Sherrer''s equation: D = 0.89λ/(B cos θ) where, λ = 0.154 nm and B is the full width at half maximum (FWHM). The calculated particle size of 2.7 nm was in close agreement with the TEM results.Open in a separate windowFig. 4X-ray diffraction pattern of Pt@C catalyst.Chemical status of platinum in Pt@C was evaluated by X-ray photoelectron spectroscopy (KRATOS, AXIS Ultra DLD) by using monochromatic Al Kα radiation (1486.6 eV, 150 W) as exciting source. Commercial catalyst was also measured for comparison. Binding energy was externally calibrated by the Au 4f_7/2 peak at 83.8 eV. Baseline subtraction and curve fitting was done using Casa XPS software (Version 2.3.17). XP spectra of Pt 4f and C 1s are shown in Fig. 5 as well as the deconvolution of experimental line profile. The Pt 4f peak shows a doublet peak representing Pt 4f_7/2 and Pt 4f_5/2 from spin–orbital splitting. The peak top of Pt 4f_7/2 locates at 71.2 eV for both Pt@C and commercial catalyst indicating almost similar oxidation state. To examine the oxidation state of Pt in more detail, relative atomic concentration was estimated by the peak area of Pt(0), Pt(ii), and Pt(iv). The atomic concentration of Pt(0), Pt(ii), and Pt(iv) are 48.9%, 32.6%, and 18.5%, respectively, for Pt@C, and 42.9%, 33.0%, and 24.1%, respectively, for commercial catalyst. The atomic concentration of Pt(0) in Pt@C is higher than that of commercial catalyst, and the platinum nanoparticle of Pt@C is more reduced state. This result can be explained by the difference in nanoparticle size of catalysts. In a small nanoparticle, the surface atom increase per unit weight and it becomes easy to oxidize by air. In fact, Pt@C shows a larger Pt nanoparticle size (2.7 nm) by XRD spectrum compared to commercial catalyst (2.0 nm). XRD spectrum of commercial catalyst is shown in Fig. S3.^† Deconvolution of the C 1s peak of Pt@C shows sp² carbon, sp³ carbon, C–OH, and C O peaks and the peaks of sp³ carbon, C–OH, and C O show higher intensity than those of commercial catalyst. The appearance of sp³ and C–OH peaks is consistent with the TG curve indicating the presence of ethylene glycol. The peak assigned C O can derive from oxidation product of ethylene glycol. Under the presence of sodium hydroxide, the ethylene glycol reduces the metal precursor by undergoing oxidation to form aldehydes and carboxylic acid.²⁸ During the catalyst synthesis, the Pt@C was treated with HCl and precipitated from ethylene glycol solution. Considering the effect of Cl adsorption on Pt surface on electrocatalytic properties, we also performed XPS measurement of Cl 2s peak (see Fig. S4^†). The Cl 2s peak was not detected from Pt@C and the effect of Cl species on electrocatalytic properties is considered to be small.Open in a separate windowFig. 5XPS of Pt (a) and C (b) of Pt@C and commercial catalyst. Deconvolution of the Pt 4f and C 1s peak of catalyst into single chemical components. The experimental data is represented by black circles, whereas the fit is represented by blue line.Electrochemical surface area (ECSA) of Pt@C and commercial catalyst were evaluated using cyclic voltammetry (CV) technique. Commercial catalyst (30% Pt/CB, Tanaka Kikinzoku Kogyo K. K., TEC10V30E) was used as standard catalyst for comparing the electrochemical performance of synthesized Pt@C catalyst. All electrochemical experiments were used a three compartment electrochemical cell. CV measurements were carried out in argon saturated 0.1 M HClO₄ (aq.) using Pt wire as counter electrode, glassy carbon (GC) electrode coated with catalysts as working electrode and reversible hydrogen electrode was used as the reference electrode.^29,30 For the preparation of the electrode, a GC (diameter, 5 mm; area, 0.196 cm²) was first polished with alumina (0.05 μm) and rinsed with Milli-Q water twice under sonication. Further, synthesized Pt@C and commercial catalyst (ca. 1 mg) was dispersed using ultrasonification in 0.05% Nafion solution (water : 2-propanol = 3.2 : 1 (v : v)) at 0 °C for 45 min. The calculated amount of catalyst was then casted over the GC electrode fixed on the inverted RDE rotator. To obtain a uniform layer of catalyst over GC electrode surface the electrode was rotated at 700 rpm until complete dry then it was placed in oven for 30 min at 60 °C. The platinum loading density on the prepared electrodes was 10 μg cm⁻². Further, these electrodes were used for electrochemical measurements. Fig. 6 shows the cyclic voltammogram of Pt@C in comparison with commercial catalyst. CV measurements were performed at a scan rate of 50 mV s⁻¹. The electrochemically active surface areas (ECSAs) were calculated by characteristic hydrogen adsorption peak below 0.4 V. The observed ECSA values of commercial catalyst (64.0 m² g_Pt⁻¹) is slightly higher than the prepared Pt@C catalyst (53.5 m² g_Pt⁻¹). The lower ECSA values can arise because of the presence of organics at the surface of Pt NPs as evident in TGA and XPS results.Open in a separate windowFig. 6Cyclic voltammograms for Pt@C and commercial catalysts at 50 mV s⁻¹ in an argon saturated 0.1 M HClO₄.To test the electrocatalytic activity, rotatory disk experiment was performed in an oxygen saturated 0.1 M HClO₄ solution. Fig. 7 show the comparison of two primitive materials at a rotation speed of 1600 rpm. This result indicated that the electrochemical activity of the synthesized Pt@C catalyst was comparable to the commercial catalyst. The mass activity (i_km) and specific activity (i_ksp) at 0.9 V vs. RHE was evaluated after the normalization based on Pt loading density for both the catalysts (9 These electrochemical results suggested that prepared Pt@C catalyst with comparable electrochemical activity arising from continuous flow synthesis can be of great advantage for the development of next generation PEMFCs.Open in a separate windowFig. 7ORR polarization curves for Pt@C and commercial catalysts at 20 mV s⁻¹ in an oxygen saturated 0.1 M HClO₄ at a rotation rate of 1600 rpm.Comparison of electrochemical properties for Pt@C and commercial catalysts

The TiO2 topotactic transformation assisted trapping of an atomically dispersed Pt catalyst for low temperature CO oxidation

Atomically dispersed Pt catalysts are synthesized on TiO₂ with high activity and strong high temperature resistance by loading Pt in the process of converting the NH₄TiOF₃ precursor to TiO₂ by a topotactic transformation process. The atomically dispersed Pt catalyst displayed high catalytic activity for the low temperature CO oxidation reaction.

Atomically dispersed Pt catalysts are deposited on the rough surface of TiO₂, which is synthesized via topotactic transformation from a NH₄TiOF₃ mesocrystal.

Supported noble metal catalysts are widely used in industrial processes on account of their high activity and/or selectivity for many key chemical reactions.¹ Usually, the noble metals are finely dispersed on a support to give a high specific surface area to effectively use the catalytically active component and increase the amount of active sites.^2,3 Atomically dispersed metal/metal oxide catalysts have attracted widespread interest in diverse research areas, such as chemistry, material science and environmental science.⁴ Due to their low-coordination, unsaturated atoms often function as active sites in catalytic processes, suggesting that downsizing the particles or clusters to single atoms is ideal for catalytic reactions.⁵ Single atoms tend to aggregate and grow into clusters or nanoparticles during the catalytic reaction processes due to the significant surface free energy increases with decreases in particle size.⁶ Although several methods have been explored for the preparation of atomically dispersed noble-metal catalysts in the past decade, the fabrication of stable atomically dispersed noble metal catalysts is still a great challenge.⁷ It is well known that the interaction between single atoms and supported substrate is essential for stabilizing active single atoms.³ Many previous reports on oxide supported metal clusters show that surface defects of the carriers can be used as anchor points for metal clusters or even single atoms.⁸ Besides noble metal atoms as the active sites, metal oxide carriers can also play an important role in the catalytic process. The interaction between oxide carrier and metal atoms can change the electronic properties of metal atoms, so it is of great significance to the activity and selectivity of the catalysts,⁹ especially in the catalytic oxidation of CO.^2,10 The different properties of the catalyst may be caused by different coordination environments around the single atom metal center, which is similar to the so-called “support effect” in traditional heterogeneous catalysis.Herein, we report the synthesis of atomically dispersed Pt catalysts on TiO₂ supports during a topotactic transformation process of NH₄TiOF₃ mesocrystals. Topotactic transformation is a useful method for preparing crystals with a required morphology through conversion from a precursor or mother crystal, in which the crystal orientations of the precursors and the target crystals have a certain topotactic correspondence.¹¹ NH₄TiOF₃ is a typical mesocrystal, which exists in similar structures to anatase TiO₂ with an average lattice mismatch of only 0.02%.¹² The NH₄TiOF₃ mesocrystals can be topotactically transformed into TiO₂ mesocrystals by either washing with aqueous H₃BO₃ or calcination at high temperature.^13–15 The topotactic transformation can be applied to anchor the single atom via atom trapping by the rough surface,¹⁶ crystalline defects and active vacancies.⁸As illustrated in Fig. 1, NH₄TiOF₃ mesocrystals were firstly impregnated with H₂PtCl₆ solution for a period of time, resulting in adsorption of [PtCl₆]²⁻ ions on the surface and an internal pore of NH₄TiOF₃ mesocrystals. Then, NH₄TiOF₃ began to form hollow anatase TiO₂ with ordered arranged nanothorns by topotactic transformation under an aqueous H₃BO₃ environment, in which the rough surface of TiO₂ can prevent the aggregation of Pt atoms and result in the formation of highly dispersed Pt single atoms on the TiO₂ substrate (Fig. S1^†). We compared the effects of topotactic transformation, noble metal loading content and calcination. Pt/TiO₂-T catalysts were prepared with various Pt loading contents of 0.1%, 0.5%, 1% and 3%. The above samples were calcined in N₂ at 450 °C for 4 h and the final products were labeled as Pt/TiO₂-TC (experimental details are given in the ESI^†).Open in a separate windowFig. 1Schematic diagram of the synthesis of Pt/TiO₂-TC.The dispersion and configuration of atomically dispersed Pt catalysts ware characterized by atomic-resolution high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM), which can identify the heavy atoms in the actual catalyst. A panoramic SEM image of the NH₄TiOF₃ precursors clearly showed the well-defined uniform NH₄TiOF₃ nanobricks with an average length of 400 nm and a thickness of 70 nm (Fig. 2A). After topotactic transformation and calcination, the resulting Pt/TiO₂ turned rough on the surface and became a mesoporous hollow structure, as shown in Fig. 2B. From common TEM images, shown in Fig. 2C (also see Fig. S2^†), we clearly see that the sample is a hierarchical structure composed of regularly arranged nanocrystals. The inset of Fig. 2C shows a lattice spacing of 0.35 nm, which can be assigned to the (101) plane of TiO₂ nanocrystals. No Pt cluster or particle can be observed on the surface in the normal TEM observation, suggesting that the resulting Pt is highly dispersed and has a sub-nanometer size. In order to verify that Pt atoms have been successfully deposited on TiO₂, we performed atomic-resolution HAADF-STEM observations of the 1% Pt/TiO₂-TC. Large amounts of marked bright points show that individual Pt atoms (marked by the red circles) uniformly dispersed on the surfaces of the TiO₂ nanocrystals (Fig. 2D). Examination of different areas showed that only Pt single atoms existed in the sample 1% Pt/TiO₂-TC (Fig. S3^†). Fig. 2D clearly shows that each Pt atom (red circles) occupies the lattice position of a Ti atom. The statistical size distributions of Pt in 1% Pt/TiO₂-TC and 3% Pt/TiO₂-TC are 0.9 nm and 1.09 nm (Fig. S4^†), respectively, indicating that only subnanometer clusters and single atoms of Pt are formed on the TiO₂ substrate after the topotactic transformation and calcination. Energy dispersive spectroscopy (EDS) shows that Pt is uniformly dispersed on the surface of TiO₂ (Fig. 2E).Open in a separate windowFig. 2SEM images of (A) NH₄TiOF₃ mesocrystals and (B) 1% Pt/TiO₂-TC catalysts. (C) A TEM image of 1% Pt/TiO₂-TC; the inset shows the crystal lattice of anatase TiO₂. (D) High-resolution HAADF-STEM images of 1% Pt/TiO₂-TC. (E) EDS elemental mapping of a single 1% Pt/TiO₂-TC crystal.The structures of Pt/TiO₂-TC catalysts with certain Pt contents were analyzed by using powder X-ray diffraction (XRD) in order to assess the impact of Pt addition upon the nanoparticles. XRD analysis shows the structure transformation process from the NH₄TiOF₃ (Fig. S5^†) to TiO₂ (Fig. 3A). All the diffraction peaks are attributed to the anatase TiO₂. No Pt diffraction peaks were detected in these samples, indicating the presence of highly dispersed clusters or single atoms on TiO₂. Even after 450 °C of calcination in N₂ for 4 h, the catalysts exhibited no Pt diffractions, suggesting that the calcination did not cause Pt atom aggregation. The N₂ adsorption/desorption isotherms display that the BET specific areas of the 1% Pt/TiO₂-TC and 3% Pt/TiO₂-TC were 130.06 m² g⁻¹ and 132.14 m² g⁻¹, respectively (Fig. 3B).Open in a separate windowFig. 3(A) XRD patterns of Pt/TiO₂-TC catalysts with different Pt loading content. (B) N₂ adsorption–desorption isotherms and pore size distributions (inset) of Pt/TiO₂-TC catalysts. (C) O 1s XPS spectra and (D) Pt 4f XPS spectra for Pt/TiO₂ catalysts.X-ray photoelectron spectroscopy (XPS) analysis was carried out to evaluate the surface composition and valence states of the Pt/TiO₂ catalyst. The representative XPS survey scan spectrum indicates the existence of Ti, O, and Pt elements (Fig. S6^†). Ti 2p spectra at 458.6 and 464.3 eV belong to the Ti 2p_3/2 and Ti 2p_1/2 peaks of Ti⁴⁺ (Fig. S7^†).¹⁷ The binding energy of Ti 2p did not change after calcination. Fig. 3C shows O 1s core-level XPS spectra of Pt/TiO₂-T and Pt/TiO₂-TC; the catalysts contained two kinds of O species. The main peak centered at 530 eV is considered to be the oxygen band of Ti–O–Ti that can be assigned to the lattice oxygen of bulk TiO₂ and the shoulder peak at 531.35 eV can be ascribed to the surface OH species (Ti–OH) which could be correlated with an oxygen vacancy.¹⁸ It is obvious that the Pt/TiO₂-TC catalyst has more surface OH groups than the Pt/TiO₂-T catalyst. It is reported that surface OH groups are formed through water dissociation on oxygen vacancies or on metal surfaces by water–oxygen interaction.¹⁹ After calcination at 450 °C, peak shifts occurred and the two peaks corresponding to O 1s core-level XPS spectra were 529 and 531.6 eV, respectively, indicating that electron transfer occurred. The oxidation state of Pt in the catalyst is shown in Fig. 3D. The spectra collected for the 1% Pt/TiO₂-TC and 3% Pt/TiO₂-TC catalysts show two peaks at the Pt 4f edge with binding energies of 70.2 and 73.7 eV, which are assigned to the 4f_7/2 and 4f_5/2 states of Pt⁰, respectively. For 3% Pt/TiO₂-TC, the Pt 4f bimodal peak shows a downshift by 0.2 eV in binding energy. Deconvolution analysis reveals that there are two additional peaks at 71.4 and 75.8 eV, in addition to the two peaks related to Pt⁰, which can be attributed to the same spin-orbital split of Pt²⁺. The spectra collected for the 1% Pt/TiO₂-T and 3% Pt/TiO₂-T samples also show four peaks at the Pt 4f edge, in which the binding energies of 70.6 and 73.8 eV are assigned to the 4f_7/2 and 4f_5/2 states of Pt⁰, and the binding energies of 72.5 and 76 eV are assigned to the 4f_7/2 and 4f_5/2 states of Pt²⁺.²⁰ For Pt/TiO₂-T and Pt/TiO₂-TC catalysts, the binding energies of Pt²⁺ 4f_7/2 and Pt⁰ 4f_5/2 decreased by 0.3 eV and 0.8 eV, respectively, which indicates that the surface reconstruction occurred during the calcination pre-treatment and enhanced the interaction between the metal and the support. The peak area of Pt⁰ increased obviously after calcination, which may be the reason for the decomposition of the oxidation state Pt at a high temperature.²¹ The existence of Pt²⁺ suggests a strong interaction between Pt and O, which is of benefit for the thermal stability of a Pt single atom, and also propitious for the charge transfer during CO oxidation.^4b According to the XPS spectra, it is inferred that the actual content of Pt in the catalyst is very low (Table S1^†) and is favorable for atom-level dispersion.CO oxidation was chosen as a probe reaction to study the catalytic performance of single Pt atom supported on TiO₂ because such a reaction is highly sensitive to the chemical environment of the metal centers. Fig. 4 shows the catalytic performance of catalysts with different Pt loading contents. The data show that a CO oxidation reaction onset at near 50 °C and a total convention at near 130 °C, indicating an excellent catalytic performance for low temperature CO oxidation. Controlled experiments were carried out with catalysts prepared by non-topotactic transformation and non-calcined methods (Fig. S8^†). Even using a high-performance commercial TiO₂ photocatalyst, P25, as the supporting substrate, the resulting Pt/TiO₂–P25 catalyst also shows a much lower activity on CO oxidation in comparison with Pt/TiO₂-TC (Fig. S9^†). All of the samples of Pt/TiO₂-T (non-calcined) and Pt/TiO₂-NC (non-topotactic transformed) did not show good catalytic activity for CO oxidation. We summarize the CO oxidation reaction temperature in Table S2,^† in which T100 denotes the temperature at which 100% of CO was converted into CO₂ and T50 is the temperature required for a 50% CO conversion. The T50 and T100 of 1% Pt/TiO₂-TC were 108 °C and 130 °C, respectively, which are lower than those all other catalysts. The atomically dispersed Pt/TiO₂ catalysts show a higher catalytic performance in the CO oxidation reaction in comparison with previously reported single atom catalysts, such as Pt/La–Al₂O₃,¹⁶ Pt/θ-Al₂O₃ (ref. 10a) and Pd/La-γ-Al₂O₃ (ref. 10b) (Table S3^†). To illustrate the high activity of Pt/TiO₂-TC, Arrhenius plots are depicted in Fig. 4B. The corresponding Arrhenius plots of the CO reaction rate ln(TOF) show an approximate linear relation versus 1/T for the CO oxidation reaction. The apparent activation energy (E_a) is ∼10 kJ mol⁻¹, much lower than for Pt/Al₂O₃ or Pt/CeO₂ (E_a = 90–100 kJ mol⁻¹).²² 1% Pt/TiO₂-TC has the smallest apparent activation energy, which is one of the main reasons that it shows the highest catalytic activity. The TOFs with different Pt SACs are summarized in Open in a separate windowFig. 4(A) CO conversion and (B) corresponding Arrhenius plots of the reaction rate ln(TOF) versus 1/T for the CO oxidation reaction using Pt/TiO₂-TC catalysts with different Pt loading content.Specific rates and TOFs of Pt/TiO₂ catalysts compared with reported Pt SACs catalysts^a

Insulating 3D-printed templates are turned into metallic electrodes: application as electrodes for glycerol electrooxidation

We turned printed plastic pieces into a conductive material by electrochemical polymerization of aniline on the plastic surface assisted by graphite. The conductive piece was then turned into a metallic electrode by potentiodynamic electrodeposition. As a proof-of-concept, we built indirect-3D-printed Pd, Pt and Au electrodes, which were used for glycerol electrooxidation.

Insulating printed plastics are turned into metallic pieces by electrochemical polymerization of aniline followed by metal electrodeposition.

Additive manufacturing, or 3D-printing, is a revolutionary technique which allows for the easy fabrication of three-dimensional objects, enabling the construction of complex devices with minimum waste and very low cost, which previously could only have been fabricated with sophisticated equipment and facilities.^1,2 This technology has attracted attention from industry and academia due to the myriad architectures, materials and applications.^2–4 Among the various 3D-printing methods, extrusion through fused deposition modeling (FDM) is the most commonly used. This method, in which a thermoplastic is forced through a heated nozzle head, was invented in 1989 by Scott Crump,⁵ and is currently the most affordable and commercialized way of 3D-printing.The field of electrochemistry has benefited from 3D-printing technologies. Ambrosi and Pumera have detailed the advances brought by additive manufacturing to electrochemistry, including instruments, sensors and materials.² In particular, the ability to rapidly prototype new parts and devices boosted the advancement in instrumentation used for electrochemistry.^2,6,7Our group printed a three-parts electrolyzer with an approximate cost of US $5.⁸ This device was used with Pd-nanocubes-modified glassy carbon as working electrode to convert glycerol into tartronate.⁸ Concerning microfluidics, 3D-printing is already considered a revolution in microfluidics, due to the ability of partially or completely replacing poly(dimethylsiloxane) pieces built by soft lithography.⁹FDM technique has recently opened up the possibility of printing conductors, which can be used as electrodes for a variety of applications. The biggest challenge is to print ready-to-use electrodes, since the extrusion of composite thermoplastic/metal is still rare. In this sense, much effort has been spent to build and print plastic/carbon materials^10–12 and to modify existing printed objects.^4,11 Wei et al. produced graphene oxide (GO) blended with acrylonitrile butadiene styrene (ABS) or polylactic acid (PLA) to build a GO-thermoplastic filament for the first time.¹⁰ The authors also reduced GO to rGO with hydrazine hydrate in order to make a rGO-thermoplastic filament.¹⁰ Another remarkable achievement was made by Rymansaib et al.¹² These authors made a new filament of polystyrene/carbon-nanofiber/graphite simultaneously printed embedded in an insulating material, composing a monolithic object, which is electrochemically stable and prospective for electroanalysis.¹²Palenzuela et al. developed a simple protocol to achieve the theoretical electrical conductivity predicted for printed graphene/PLA electrodes.¹¹ They exposed the printed electrodes to dimethylformamide in order to dissolve the fused polymer on the electrode surface. This procedure allows the authors to access the active sites of the electrode, which is used for electroanalysis.¹¹ Aside from thermoplastic/carbon materials, there is an impending need to build metallic 3D-printed electrodes. A successful post-printed modification was made by Díaz-Marta et al. to produce Pd- and Cu-based materials for heterogeneous organic synthesis.⁴ By following complex steps, the authors printed a SiO₂ material, which was chemically treated in sequential steps to be finally modified with CuI and Pd(AcO)₂ to produce Cu and Pd catalysts.⁴ The development of metal-based 3D-printed materials for a wide range of electrochemistry applications using simple protocols is still necessary.Here, we used FDM to print PLA templates, which are modified with precursors of a conducting polymer and graphite (GR). This material is submitted to electrochemical polymerization, allowing further metallic electrodeposition. As proof-of-concept, we prepared Pd, Pt and Au electrodes built on 3D-printed templates. These electrodes were then used for glycerol electrooxidation to investigate their applicability as anodes of glycerol fuel cells and electrolyzers.^13–15We used. PLA as the thermoplastic filament in a commercially available 3D-printer (Sethi 3D, model S3). A CAD model of the electrode was drawn using software Autodesk Inventor 2017 and further sliced using software Simplify 3D. The electrodes are chosen to have dimensions of 3 × 3 × 5 mm³ so that they could be connected to a regular alligator connector, as shown in Fig. 1. The template is modified in aniline and HCl solution and further modified with graphite 99.69% (ashes 0.31%; humidity 0.06%; 13.33% retained in 100# grid and 97.40% in 325# grid). The procedures for such modifications are detailed through the text.Open in a separate windowFig. 1Illustrative scheme of the graphite-assisted electrochemical polymerization of polylactic acid template electrodes. The PLA template is covered by aniline precursor and graphite, which is further electrochemically polymerized to emeraldine (and other PANI structures in different oxidation states).The printing parameters of the working electrode template are summarized in Table S1.^† The total building time is only 1 min and material cost is 0.01 US $, which is a very low price and a fast way of obtaining the templates. It is worth noting that the printing parameters found in Table S1^† change whether one changes the size and shape of the desired template.The post-printing treatment developed here is based on a surface polymerization reaction for growing polyaniline (PAni) on PLA. The 3D-printed template is immersed in a mixture of 1 mL aniline with 2 mL 1 mol L⁻¹ HCl in order to place the polymer precursor, aniline chloride (C₆H₅NH₃⁺Cl⁻) on the PLA surface. The aim is to polymerize the precursor on the template to become conducting. The electrochemical synthesis of PAni has been previously described.^16,17 This conducting polymer is structurally different depending on the oxidation state. The reduced form, which has a yellow color, is leucoemeraldine. This form can be electrooxidized to emeraldine, which is blue in an alkaline medium or green in an acidic medium, as shown in Fig. S1.^† Emeraldine can be further oxidized to pernigraniline (violet color). In order to maintain structural stability during the electrochemical measurements, the reduced form leucoemeraldine/emeraldine is desirable, while further oxidation of emeraldine may accelerate degradation. This process has been widely discussed in the literature, as those for metal polymerization via electrochemistry.^18,19 For these previous studies, the substrate of the polymerization was already a conducting material, facilitating the electric contact with the voltage source (e.g. potentiostat/galvanostat), represented by an “aligator” in Fig. 1. However, in the present case, the challenge is that the widely used 3D-printed PLA is an insulating material.To surpass this issue, we developed GR-assisted polymerization. Immersing the template in the aniline chlorite solution smoothly melts the PLA surface. The template is then transferred to a glass container with GR and gently shaken until the PLA is completely covered. The GR-modified printed piece is finally dried at room temperature for 20 h. After dried, the template is washed with D.I. water under stirring for 10 min. This process is repeated five times and it is imperative to spread out unattached GR. Finally, the electrode is transferred to an electrochemical cell for the electrochemical polymerization of aniline in 1 mol L⁻¹ HCl. Once impossible electrochemical polymerization of aniline over PLA, is now achievable by cyclic voltammetry. The electrolyte accesses the template surface covered with C₆H₅NH₃⁺Cl⁻ as it goes through the GR flakes, while GR makes the connection between the electric connector (alligator connected to the potential source) and the electrolyte, allowing for electrochemical polymerization, as illustrated in Fig. 1.The GR-assisted electrochemical polymerization of the GR/C₆H₅NH₃⁺Cl⁻/PLA template is shown in Fig. 2. The growth of PAni is evidenced through the increase of two redox coupled peaks at low (I/I′) and high (II/II′) potentials. The oxidation of leucoemeraldine to emeraldine starts at 0.35 V, displaying an evident peak centered at 0.48 V (peak I in Fig. 2). Emeraldine is further oxidized to pernigraniline, displaying a peak at 0.73 V (peak II in Fig. 2). During the negative potential scan, pernigraniline is reduced to emeraldine at ∼0.7 V and further reduced to leucoemeraldine with a cathodic transient centered at 0.25 V, as shown by peaks II′ and I′ respectively in Fig. 2. The potential cycles were performed until the redox couple I/I′ is evident, which takes place after ∼30 cycles. At this point, the PLA is mainly covered by emeraldine, making it a conducting electrode.Open in a separate windowFig. 2Graphite-assisted electrochemical polymerization of GR/C₆H₅NH₃⁺Cl⁻/PLA template in 1 mol L⁻¹ HCl at 0.05 V s⁻¹.The electrode resistance of the polymerized material is 45 Ω, which is acquired by measuring the high frequency intercept of the Nyquist plot in an electrochemical impedance spectroscopy analysis performed at 106 Hz. The electrochemical profile shown in Fig. 2 associated to the low resistance demonstrates the success of the GR-assisted polymerization. Furthermore, the resistance measured with a multimeter is ∼36 Ω cm⁻¹, which reveals a remarkable conducting material. GR has a double contribution on the electrode construction; one is assisting polymerization, working as an electron collector (previously discussed) while also contributing to the increase in conductivity after polymerization. It is worth noting that GR used with other organochlorides does not show similar performance. As an example, we smoothly melted the PLA template in dichloromethane and dispersed GR on it. In such case, the resulting composite showed high resistance, in the MΩ scale. Therefore, PAni coverage is imperative to the increase in electrical conductivity of the template.After turning the PLA into an electrical conductor to form a GR/PAni/PLA template, the printed piece was immersed in a 6 mmol L⁻¹ PdCl₂ in 0.5 mol L⁻¹ H₂SO₄ for Pd electrodeposition. The metallic deposition was achieved by applying successive potential cycles between 0.019 and 1.319 V. Potentiodynamic deposition allows us to observe the formation of a Pd electrode through the appearance of a Pd profile, i.e. we can in situ evaluate the growth of Pd on the template. In other words, once a characteristic profile of Pd in acidic media is identified, we are certain of a success deposition. The experimental details of the electrochemical measurements are found in the ESI Section II.^† Fig. 3A shows an increase in anodic and cathodic currents wherein successive potential cycles are applied. Pd²⁺ is reduced to Pd on the template during the cycles and the currents related to the Pd surface reaction in aqueous acid solution become more evident. After identifying an obvious profile of Pd, the electrode was transferred to an electrochemical cell containing 0.1 mol L⁻¹ KOH in order to measure a profile in a controlled system.Open in a separate windowFig. 3(A) Potentiodynamic electrodeposition process achieved by successive potential cycles of GR/PAni/PLA template in 6 mmol L⁻¹ PdCl₂ in 0.5 mol L⁻¹ H₂SO₄, measured between 0.019 and 1.31 V. (B) Cyclic voltammogram of the indirect-3D-printed Pd electrode in 0.1 mol L⁻¹ KOH, measured between 0.15 and 1.27 V and (C) in the presence of 0.2 mol L⁻¹ glycerol between 0.35 and 1.29 V. All measurements are performed at 0.05 V s⁻¹. (D) and (E) show representative SEM images, while (F) and (G) show EDS Kα elemental composition maps (indicated in the figure). (H) EDS spectrum of the indirect-3D-printed Pd electrode. Fig. 3B undoubtedly evidences a characteristic profile of Pd through the cyclic voltammogram in the range of 0.15–1.27 V. Surface oxide formation starts at ∼0.7 V during the positive potential sweep, which is reduced producing a cathodic current, forming a peak centered at ∼0.7 V. Another pivotal characteristic is the hydrogen under potential deposition (H_UPD) region at around 0.15–0.5 V. Overall, this electrochemical profile shows that a Pd electrode was indeed manufactured.²⁰The indirect-3D-printed Pd electrode was then used as a working electrode in the presence of 0.2 mol L⁻¹ glycerol in 0.1 mol L⁻¹ KOH, as shown in Fig. 3C. The cyclic voltammograms display anodic currents during the positive and negative potential sweeps. The current density increases with successive cycles due to the initial cleaning of the surface, until a stable profile is reached at the fifth cycle. Such a phenomenon was previously reported for Pd catalyst.²¹ During the positive scan, electrooxidation starts at ∼0.6 V, reaching a maximum at ∼1.08 V. During the reverse scan, the surface reactivation after the reduction of Pd surface oxides promotes an accentuated oxidation peak centered at 0.76 V; these potentials match the electrocatalytic parameters reported for commercial Pd/C.¹³The Pd electrode was also investigated by EDS mapping of the Pd electrode in order to qualitative characterize the chemical composition (experimental details in ESI Section III^†). Firstly, the morphology of the electrode investigated via SEM reveals nanostructures built onto the GR (Fig. 3D). The Pd particles do not have well-defined shapes and are slightly agglomerated. The size of the particles ranges between approximately 100–450 nm. A section of an SEM image (Fig. 3E) was then used to qualitatively investigate the chemical composition of a Pd electrode using EDS. Fig. 3F shows an image highlighting the presence of carbon on the electrode surface, whereas Fig. 3G highlights the presence of Pd. The large region with high presence of Pd (Fig. 3G) overlaps with the region of Pd particles from the SEM image (Fig. 3E). Moreover, the wide carbon region (Fig. 3E) is indicated by the high-intensity green colored region on the EDS mapping (Fig. 3F). These findings assure the presence of the metal and carbon on the printed template. Finally, an EDS spectrum shows the presence of Pd through a peak at ∼2.84 keV (Fig. 3H). Aiming the manufacturing of noble metal-based indirect-3D-printed, we built Pt and Au electrodes, which were both applied for glycerol electrooxidation, as described in ESI Section IV.^†In summary, we successfully turned an insulating 3D-printed piece into a mostly metallic piece with low resistivity. The present research shows a low-cost alternative to build metallic electrodes or metallic pieces suitable for different sizes and shapes of a 3D-printed template with multipurpose. This sustainable protocol allows for the modification of any existing printed material, so it is an alternative for recycle and/or reuse of existing materials, showing that low-cost and widely available thermoplastic filaments of FDM 3D-printers can be used as source of template. This work opens up the possibility of indirect-3D-printing any metallic piece in any shape.     

Correction: A sensitive OFF–ON–OFF fluorescent probe for the cascade sensing of Al3+ and F− ions in aqueous media and living cells

Correction for ‘A sensitive OFF–ON–OFF fluorescent probe for the cascade sensing of Al³⁺ and F⁻ ions in aqueous media and living cells’ by Lingjie Hou et al., RSC Adv., 2020, 10, 21629–21635, DOI: 10.1039/D0RA02848G.

The authors regret that an incorrect version of Fig. 4 was included in the original article. The correct version of Fig. 4 is presented below.Open in a separate windowFig. 4The ESI-MS spectrum of Al³⁺–HNS complex.The Royal Society of Chemistry apologises for these errors and any consequent inconvenience to authors and readers.     

Hydrodeoxygenation of 2,5-dimethyltetrahydrofuran over bifunctional Pt–Cs2.5H0.5PW12O40 catalyst in the gas phase: enhancing effect of gold

2,5-Dimethyltetrahydrofuran (DMTHF) is deoxygenated to n-hexane with >99% selectivity at mild conditions (90 °C, 1 bar H₂ pressure, fixed-bed reactor) in the presence of the bifunctional metal-acid catalyst Pt–CsPW comprising Pt and Cs_2.5H_0.5PW₁₂O₄₀ (CsPW), an acidic Cs salt of Keggin-type heteropoly acid H₃PW₁₂O₄₀. Addition of gold to the Pt–CsPW catalyst increases the turnover rate at Pt sites more than twofold, whereas the Au alone without Pt is not active. The enhancement of catalyst activity is attributed to PtAu alloying, which is supported by STEM-EDX and XRD analysis.

Addition of gold to the Pt–CsPW catalyst has an enhancing effect on the HDO of DMTHF, with a twofold increase of turnover rate at Pt sites.

Biomass-derived furanic compounds are of interest as a renewable feedstock, which can be processed into a range of value-added chemicals and green fuels via catalytic hydroconversion.^1–8 Hydrodeoxygenation (HDO) of furanic compounds using bifunctional metal–acid catalysis has been demonstrated to be an effective strategy to produce green fuels under mild conditions^3,4,6,8–12 and references therein. The HDO over bifunctional metal-acid catalysts is much more efficient compared to the reaction over monofunctional metal catalysts.^11,12 Previously, we have reported HDO of a wide range of oxygenates in the gas phase to produce alkanes in the presence of bifunctional catalysts comprising Pt, Ru, Ni and Cu as metal components and Keggin-type heteropoly acids, with their activity decreasing in the order Pt > Ru > Ni > Cu.^13,14 Pt–CsPW comprising Pt and strongly acidic heteropoly salt Cs_2.5H_0.5PW₁₂O₄₀ (CsPW) has been reported to be a highly efficient catalyst for the HDO of 2,5-dimethylfuran (DMF) and 2,5-dimethyltetrahydrofuran (DMTHF) to produce n-hexane with 100% yield at 90–120 °C and ambient pressure.^11,12 The HDO of DMTHF over Pt–CsPW occurs through a sequence of hydrogenolysis, dehydration and hydrogenation steps catalysed by Pt and proton sites of the bifunctional catalyst (Scheme 1). These include the ring opening of DMTHF to form 2-hexanol on Pt sites followed by its dehydration on proton sites of CsPW to hexene, which is finally hydrogenated to n-hexane on Pt sites.¹² It is the facile dehydration of the secondary alcohol intermediate that drives the HDO process forward.^11,12 The rate-limiting step is either the ring hydrogenolysis or 2-hexanol dehydration depending on the ratio of accessible surface metal and acid sites Pt/H⁺.¹² Other platinum group metals such as Pd, Ru and Rh, that have high selectivity to ring hydrogenation rather than ring hydrogenolysis,^2,7 have low activities in HDO of DMF and DMTHF.¹¹Open in a separate windowScheme 1Reaction pathway for hydrodeoxygenation of DMTHF over Pt–CsPW.Bimetallic PtAu and PdAu catalysts have been reported to have an enhanced performance in comparison to monometallic Pt and Pd catalysts,^15–31 for example, in hydrogenation,^16,21,29 hydrodesulphurisation,^27,28 oxidation,^22,24 isomerisation^15,19,30,31 and other reactions.^{17,18,20,25,26} The enhancement of catalyst performance by addition of gold can be attributed to geometric (ensemble) and electronic (ligand) effects of the constituent elements in PtAu and PdAu bimetallic species.^25,26Here we looked at the effect of Au on the performance of Pt–CsPW catalysts in the HDO of DMTHF in the gas phase (see the ESI for experimental details^†). The CsPW heteropoly salt is a well-known solid acid catalyst; it possesses strong proton sites, large surface area and high thermal stability (∼500 °C decomposition temperature).^9,32–34 Supported bimetallic catalysts PtAu/SiO₂ and PtAu/CsPW were prepared by co-impregnation of H₂PtCl₆ and HAuCl₃ onto SiO₂ and CsPW followed by reduction with H₂ at 250 °C (ESI^†). This method gives supported bimetallic PtAu nanoparticles of a random composition together with various Pt and Au nanoparticles.^15,16,31 Information about the catalysts studied is given in CatalystSurface area^a (m² g⁻¹)Pore volume^b (cm³ g⁻¹)Pore diameter^c (Å) D ^d d ^e (nm)Cs_2.5H_0.5PW₁₂O₄₀ (CsPW)1350.089276.5% Au/SiO₂2571.011570.019^f46^g, 38ⁱ4.7% Au/CsPW1030.048330.016^f60^g6.4% Pt/SiO₂2661.061590.28 ± 0.04^h3.2^f, 8.0^g, 5ⁱ6.0% Pt/CsPW840.052250.17 ± 0.03^h5.3^f6.6% Pt/5.9% Au/SiO₂2401.081790.29 ± 0.05^h3.1^f5.9% Pt/4.4% Au/CsPW910.082360.17 ± 0.04^h5.3^fOpen in a separate window^aBET surface area.^bSingle point total pore volume.^cAverage BET pore diameter.^dMetal dispersion.^eMetal particle size.^fCalculated from the equation d (nm) = 0.9/D.^gMetal particle diameter from powder XRD (Scherrer equation).^hPt dispersion determined by H₂/O₂ titration (average from three measurements); for PtAu catalysts, assumed negligible H₂ adsorption on gold (see the ESI).ⁱFrom STEM.Powder X-ray diffraction (XRD) has been widely used for the characterization of supported Au alloy catalysts.²⁶ The XRD patterns for the silica-supported catalysts 6.4% Pt/SiO₂, 6.5% Au/SiO₂ and 6.6% Pt/5.9% Au/SiO₂ are shown in Fig. 1. As expected, the 6.4% Pt/SiO₂ and 6.5% Au/SiO₂ catalysts display the fcc pattern of Pt and Au metal nanoparticles. The Pt peaks are broader than the Au peaks, which indicates a higher dispersion of Pt particles, with an average particle size of 8.0 nm for Pt and 46 nm for Au, which is in agreement with the STEM values (Open in a separate windowFig. 1Powder XRD patterns of 6.4% Pt/SiO₂, 6.5% Au/SiO₂ and 6.6% Pt/5.9% Au/SiO₂; the pattern for 6.6% Pt/5.9% Au/SiO₂ shows broad [111], [200], [220] and [311] fcc PtAu alloy peaks in the range 38–40°, 44–48°, 65–68° and 78–81°, respectively.The pattern for the 6.6% Pt/5.9% Au/SiO₂ catalyst clearly shows the presence of PtAu bimetallic particles with broad [111], [200], [220] and [311] diffraction peaks of the fcc PtAu alloy between the corresponding diffractions of the pure metals in the range 38–40°, 44–48°, 65–68° and 78–81°, respectively. Fig. 2 shows the high-angle annular dark field (HAADF) STEM images of the three silica-supported catalysts 6.4% Pt/SiO₂, 6.5% Au/SiO₂ and 6.6% Pt/5.9% Au/SiO₂ with metal nanoparticles indicated as bright spots on the darker background. In the Pt/SiO₂ catalyst, there are two populations: small Pt particles of 5 nm size and coalesced Pt particles of a larger size. The Au/SiO₂ catalyst displays Au particles of spherical, rectangular and triangular morphology, with an average size of 38 nm. The bimetallic PtAu/SiO₂ catalyst shows a high agglomeration and different kinds of morphology of metal particles.Open in a separate windowFig. 2HAADF-STEM images of (a) 6.4% Pt/SiO₂, (b) 6.5% Au/SiO₂ and (c) 6.6% Pt/5.9% Au/SiO₂ catalysts, showing noble metal nanoparticles as bright spots.The energy dispersive X-ray spectroscopic analysis (EDX) of metal particles in the PtAu/SiO₂ catalyst shows that these particles contain both platinum and gold. EDX elemental mapping clearly demonstrates that Pt and Au maps cover the same areas of PtAu/SiO₂ catalyst (Fig. 3), indicating PtAu alloying with formation of a non-uniform bimetallic PtAu particles. More EDX mapping is presented in the ESI (Fig. S1^†).Open in a separate windowFig. 3HAADF-STEM image of 6.6% Pt/5.9% Au/SiO₂ catalyst and the corresponding STEM-EDX elemental maps showing the spatial distribution of Au (red) and Pt (green) in the sample.STEM–EDX for CsPW-supported Pt, Au and PtAu catalysts has been reported elsewhere.¹⁶ These STEM images are difficult to analyse due to W, Pt and Au having similar large atomic numbers Z (74, 78, and 79, respectively). Crystalline CsPW containing 70 wt% of W displays a strong background which makes it difficult to discern smaller Pt and Au particles from the Z-contrast HAADF images and determine accurately metal particle size. Nevertheless, the STEM-EDX analysis indicates the presence of bimetallic PtAu particles in the PtAu/CsPW catalyst with a wide range of Pt/Au atomic ratios.¹⁶Representative results for HDO of DMTHF in the presence of bifunctional metal-acid catalysts Pt–CsPW and PtAu–CsPW, which were used as physical mixtures of metal and acid components at similar Pt loadings, are shown in †). The molar ratio of surface metal and proton sites in the catalysts was chosen low enough (Pt/H⁺ = 0.03–0.1) to ensure the reaction being limited by the DMTHF ring opening step.¹² The density of surface Pt sites was estimated from the Pt dispersion (32,33 and the CsPW surface area of 135 cm² g⁻¹ (EntryCatalystConversion (%)TOF^b (h⁻¹)Product selectivity (% mol) n-Hexane2-Hexanol1CsPW2.124.7% Au/CsPW + CsPW2.236.0% Pt/CsPW + CsPW8.67098.60.745.9% Pt/4.4% Au/CsPW + CsPW1717098.60.856.5% Au/SiO₂ + CsPW1.966.4% Pt/SiO₂ + CsPW6439099.40.576.6% Pt/5.9% Au/SiO₂ + CsPW8549099.60.386.4% Pt/SiO₂ + CsPW^c8.015098.61.296.6% Pt/5.9% Au/SiO₂ + CsPW^c1326098.50.7Open in a separate window^a0.20 g total catalyst weight (physical mixture of 0.020 g metal catalyst + 0.18 g CsPW), 0.6% Pt, 90 °C, 2.3 kPa DMTHF, 20 ml min⁻¹ H₂ flow rate, catalyst pre-treatment at 90 °C for 1 h in H₂ flow, 1 h TOS.^bTOF values per Pt surface site, the contribution of Au and CsPW subtracted.^cCatalyst bed contained 0.005 g metal catalyst + 0.18 g CsPW; catalyst pre-treatment at 250 °C for 1 h in H₂ flow.In the absence of Pt, the CsPW alone (entry 1) and Au–CsPW (entries 2 and 5) showed a negligible activity (1.9–2.2% DMTHF conversion with practically no 2-hexanol and n-hexane formed). Physically mixed Pt–CsPW catalysts, Pt/CsPW + CsPW and Pt/SiO₂ + CsPW (1 : 9 w/w), exhibited a high activity giving >99% n-hexane selectivity at 8.0 to 85% DMTHF conversion depending on the catalyst and reaction conditions, in agreement with the previous report.¹² It should be noted that the catalyst based on Pt/SiO₂ had almost 6-fold greater activity than the one based on Pt/CsPW in terms of turnover frequency (TOF) per surface Pt site (cf. entries 3 and 6), thus demonstrating a strong effect of Pt support.As can be seen from 15,31The enhancement of catalyst activity by addition of gold has been attributed to geometric (ensemble) and electronic (ligand) effects of the constituent metals in PtAu bimetallic nanoparticles.²⁶ The XRD and STEM-EDX data shown above clearly demonstrate PtAu alloying in the PtAu/SiO₂ catalyst leading to the formation of bimetallic PtAu species. The same has also been reported for the PtAu/CsPW catalyst.¹⁶ Previously, it has been shown that the HDO of DMTHF on Pt–CsPW is a structure-sensitive reaction,¹² hence the geometric effects may be expected to contribute to the gold enhancement. However, in order to prove the role of geometric and electronic effects as the cause of the gold enhancement, more accurate metal dispersion measurements complemented by spectroscopic characterisation will be required.We also tested the performance of bifunctional PdAu/SiO₂ + CsPW and PtPd/SiO₂ + CsPW bimetallic catalysts under similar conditions in comparison to the corresponding monometallic Pd and Pt catalysts. However, no enhancement of activity was observed. This is in agreement with XRD analysis, which showed no distinct PdAu alloying in PdAu/SiO₂ (Fig. S2 in the ESI^†).In conclusion, we have demonstrated that the addition of gold to the Pt–CsPW catalyst has an enhancing effect on the HDO of DMTHF, increasing the turnover rate at Pt sites more than twofold. The enhancing effect is attributed to PtAu alloying. The formation of bimetallic PtAu nanoparticles in the PtAu–CsPW catalyst is confirmed by STEM-EDX and XRD.