View PDF VersionPrevious ArticleNext Article
 
DOI: 10.1039/C6CP00375C (Paper) Phys. Chem. Chem. Phys., 2016, 18, 13092-13107

Enhanced photophysical properties of plasmonic magnetic metal-alloyed semiconductor heterostructure nanocrystals: a case study for the Ag@Ni/Zn1−xMgxO system†

Sumana Paul , Sirshendu Ghosh *, Manas Saha and S. K. De *
Department of Materials Science, Indian Association for the Cultivation of Science, Kolkata-700032, India. E-mail: mssg2@iacs.res.in; msskd@iacs.res.in

Received 18th January 2016 , Accepted 11th April 2016

First published on 11th April 2016

Understanding the effect of homovalent cation alloying in wide band gap ZnO and the formation of metal–semiconductor heterostructures is very important for maximisation of the photophysical properties of ZnO. Nearly monodisperse ZnO nanopyramid and Mg alloyed ZnO nanostructures have been successfully synthesized by one pot decomposition of metal stearate by using oleylamine both as activating and capping agent. The solid solubility of Mg(II) ions in ZnO is limited to ∼30% without phase segregation. An interesting morphology change is found on increasing Mg alloying: from nanopyramids to self-assembled nanoflowers. The morphology change is explained by the oriented attachment process. The introduction of Mg into the ZnO matrix increases the band gap of the materials and also generates new zinc interstitial (Zni) and oxygen vacancy related defects. Plasmonic magnetic Ag@Ni core–shell (Ag as core and Ni as shell) nanocrystals are used as a seed material to synthesize Ag@Ni/Zn1−xMgxO complex heterostructures. Epitaxial growth is established between Ag(111) and ZnO(110) planes in the heterostructure. An epitaxial metal–semiconductor interface is very crucial for complete electron–hole (e–h) separation and enhancement of the exciton lifetime. The alloyed semiconductor–metal heterostructure is observed to be highly photocatalytically active for dye degradation as well as photodetection. Incorporation of magnetic Ni(0) makes the photocatalyst superparamagnetic at room temperature which is found to be helpful for catalyst regeneration.

Introduction

Nanoscale heterostructured materials are currently of great interest from the application point of view in the field of interdisciplinary nanoscience. Among the various kinds of heterointerface, semiconductor–semiconductor, metal–semiconductor and core–shell heterostructures are found to be very important to engineer physicochemical properties.1 Metal–semiconductor heterostructures are most intriguing for their multimodal applications in a wide range of areas like sensing,2 magnetism, catalysis,3 biology,4 photovoltaics,5 photocatalysis6 and more. In metal–semiconductor heterostructures, a coupling between the quantum confined excitonic states of the semiconductor and dielectrically confined electronic states in the metal part is expected for efficient interaction between metal and semiconductor. In particular, noble metal–semiconductor hybrid structured NCs have become a most attractive topic of research due to tunability of the optical properties over a wide spectral range. The excitation of localized surface plasmon resonance (LSPR) of metal nanocrystals (NCs) often enhances the photophysical properties of closely situated semiconductor NCs and even generates new functionality. Au, Cu and Ag metal NCs are found to be highly efficient light absorbers in the UV-Vis and also the NIR region depending upon the size, shape and degree of alloying for their strong and stable LSPR phenomenon. For example improved photocatalytic properties have been reported in Au–ZnO,7 Au–CdSe,8 Au–TiO2,9 Au–SnS10 and Au–Cu2ZnSnS411 hybrid nanostructures.

The wide band gap metal oxides such as TiO2 and ZnO are the most attractive class of semiconductors because these materials have high catalytic efficiency, low cost and are environmentally sustainable. ZnO is a common and inexpensive green material with high electron mobility and large carrier concentration due to intrinsic defects and it proved to be a promising candidate for transparent conducting oxide (TCO) coatings,12,13 photocatalysts,14 gas sensors15 and photodetectors.16 The combination of plasmonic metal NCs (Au, Ag, Cu) with the ZnO nanostructure improves the photocatalytic7 and gas sensing performances17 of ZnO. To maximize the photocatalytic process of semiconductors, it is important to achieve photo-induced charge separation by band gap excitation of the semiconductors. Also the increase in exciton lifetime enhances the photoresponse properties of a material. Fast recombination of electron–hole pairs is undesirable and always leads to a reduced photocatalytic efficiency and poor photoresponse. One of the promising ways to prevent the fast recombination of photogenerated electrons and holes is to combine the ZnO with metal NCs to form a hybrid heterostructure. Metal NCs can act as an electron sink (for high electron affinity and lower Fermi energy level) which can accept the photogenerated electrons from the semiconductor and increase the exciton lifetime. Efficient electron transfer is only possible if the metal and semiconductor moieties are connected by an epitaxial crystallographic relation. Direct contact of metal through epitaxy can lead to effective charge separation and enhance the interfacial charge transfer which can promote photocatalytic efficiency and photoresponse.

Photochemically charged ZnO colloidal NCs have been established as powerful reductants.18–20 Detailed experimental studies have proved that Mg2+ alloying in ZnO not only widens the band gap but also increases the conduction band (CB) potential and lowers the valence band (VB) potential in a ratio of ΔECB[thin space (1/6-em)]:[thin space (1/6-em)]ΔEVB = 0.68[thin space (1/6-em)]:[thin space (1/6-em)]0.32, which makes Mg alloyed ZnO more reducing than pure ZnO.21 Theoretical calculations also predict a more reducing capability of Mg alloyed ZnO than pure ZnO.22,23 So from a literature survey, we found two ways to make ZnO more photocatalytic, either by making proper epitaxy between the metal and semiconductor or by Mg alloying in ZnO. The similar tetrahedral ionic radii of Mg2+ (0.57 Å) and Zn2+ (0.60 Å) give rise to the high solid solubility of Mg2+ in wurtzite ZnO.24 Zn1−xMgxO thin films have been prepared by different solid state techniques like pulsed laser deposition (PLD),25 molecular beam epitaxy,26 magnetron sputtering, chemical vapor deposition27 and chemical methods like sol–gel deposition. Synthesis of colloidal Zn1−xMgxO NCs with well defined shapes is very limited.28,29

Here we report the synthesis of nearly monodisperse ZnO and Zn1−xMgxO NCs by a thermal decomposition process via a one pot reaction taking metal stearate as the metal ion source and oleylamine as the activating and capping agent. The Mg2+ ion incorporation into the wurtzite matrix has an unusual effect on the shape evolution of the Zn1−xMgxO NCs. On Mg2+ alloying, the perfect pyramidal structure of pure ZnO changes to a flower-like morphology. The present observation is quite different from previously reported Zn1−xMgxO systems where Mg2+ incorporation changes the shape from small nanopyramids to nanorods for the preferential growth of the material along the c-axis.28,29

A series of Au–ZnO hybrid nanostructures including pyramids,30 nanorods,7 and nanourchins31 of ZnO NCs have been studied. Some reported methods show uncontrolled deposition of Au on ZnO without any proper epitaxy. Excellent epitaxial interface between Au and ZnO and enhanced photocatalytic activity have been reported by Tahir et al.7 As Au is a precious metal, large scale commercial applications of such nanoheterostructures (NHS) would be limited. Here we use Ag for the fabrication of a metal-alloyed semiconductor heterostructure for two reasons: (i) it is more cost-effective, and (ii) the Fermi level position of Ag metal (−4.7 eV) is higher than that of Au (−5.1 eV).32 The appearance of the Fermi level of Ag closer to the conduction band edge of ZnO makes the transfer of photoexcited electrons from ZnO to Ag metal energetically more favorable. However, the density of states of unoccupied energy levels in the metal at the conduction band edge of a semiconductor determines the rate of electron transfer between the photoexcited semiconductor and metal. Besides, the ZnO NCs contain a large amount of defect states in the band gap region (regardless of the synthesis procedure). So one might prefer a metal whose Fermi energy level lies in the close vicinity of the CB edge to quench the radiative transitions to defect states and consequently to gain higher photoefficiency.

Due to the high cost of noble metals like Ag or Au, large-scale commercial application of these metal–semiconductor NHS would be restricted. One of the solutions lies in the isolation of the catalyst through physical methods after each cycle of catalysis and reusing it. The easiest method is the introduction of magnetic properties in the heterostructure so it can be magnetically separable. From literature study33,34 we found that Ag@Ni and Ag@Pd core–shell structures exhibit room temperature ferromagnetism and can be separated from solution by a small magnet. We have chosen Ag@Ni core–shell structure (because of the lower cost of Ni) to form photo-catalyst metal–semiconductor NHS for easy physical separation of the catalyst after each cycle of catalysis, making the overall process a cost-effective one.

Here we report an easy one pot green synthesis procedure of Ag@Ni/Zn1−xMgxO ternary NHS. The synthesis procedure involves first the synthesis of Ag@Ni core–shell structures as seed material from non-toxic silver and nickel salt, and in situ growth of Zn1−xMgxO NCs over the metal NCs. Both the pure and alloyed systems are found to be epitaxially related to the Ag(111) plane of the Ag@Ni moiety. An interesting shape change occurs for the heterostructure and the pure alloyed system on Mg2+ alloying. Here we investigate both the effects of Mg alloying and plasmonic magnetic metal–semiconductor formation on the photophysical properties of ZnO.

Experimental section

Materials

Zinc stearate (Zn(st)2), magnesium stearate (Mg(st)2), trioctylphosphine (TOP) and 1-octadecene [ODE, 90%] were purchased from Alfa Aesar. Oleylamine [(OLAM), 70%, tech], nickel acetylacetonate [Ni(acac)2] and tetrachloroethylene (TCE) were purchased from Aldrich. N-Butylamine (GR), acetonitrile (GR), n-octane (GR), silver nitrate (GR), ethyl alcohol (GR), acetone (GR), n-hexane (GR) and toluene (GR) were purchased from MERCK. All chemicals were used as received without any purification.

Synthesis

Synthesis of ZnO and Mg doped ZnO (Zn1−xMgxO) NCs. Pure ZnO and Zn1−xMgxO NCs were synthesized with x = 0.05, 0.10, 0.20, 0.30, 0.40 and 0.50. In order to prepare ZnO NCs, 1 mmol of Zn(st)2 was mixed with 5 ml of ODE in a 25 ml round bottom three neck flask fitted with a reflux condenser. Initially the mixture was evacuated for 30 min at room temperature. Then the temperature was gradually increased to 90 °C under constant evacuation conditions. N2 gas was purged into the three neck flask and 5 ml of degassed OLAM at 120 °C was injected into that solution and was kept for 1 h. The solution turned transparent yellow. The transparent solution was again heated to 280 °C and the reaction was continued for 60 min. After completion of reaction, the solution was cooled to room temperature and 5 ml of n-hexane was added to this solution. The product was obtained by precipitation of the solution by adding excess ethanol as a polar solvent. The product was collected by centrifugation at 10[thin space (1/6-em)]000 rpm for 3 min. The centrifugation was repeated 3 to 4 times to remove impurities. The NCs were well soluble in non-polar solvents like hexane, toluene, TCE, etc.

The Zn1−xMgxO NCs were prepared under the same reaction conditions, by adding Mg(st)2 in the required stoichiometric amount in ODE.

Synthesis of Ag@Ni nanoparticles. Ag@Ni core–shell nanoparticles were synthesized by a one-pot seed growth method. Here Ni(acac)2 and AgNO3 were thermally decomposed in oleylamine. First of all Ni(acac)2 and AgNO3 were mixed with 6 ml of OLAM in 4[thin space (1/6-em)]:[thin space (1/6-em)]1 ratio in a three neck round bottom flask. Then 0.05 ml of trioctylphosphine (TOP) was added. The mixture was kept under vacuum for 20 min and the temperature was increased to 80 °C. The reaction was continued for 15 min under these conditions, after which N2 gas was purged into the solution. Then the reaction temperature was increased to 170 °C slowly and reaction was continued for 60 min. After completion of reaction, the solution was cooled to room temperature, hexane was added to the reaction solution and the product was collected by centrifugation of the solution by adding excess ethanol as a polar solvent. The centrifugation was done at 10[thin space (1/6-em)]000 rpm for 3 min.

Preparation of Ag@Ni/Zn1−xMgxO NHSs

Metal–semiconductor NHSs can be synthesized by two procedures: the first is by growing the metal core followed by deposition of the semiconductor on it, and the second is vice versa. We followed the first procedure to prepare Ag@Ni/Zn1−xMgxO NHSs. Ni(acac)2 and AgNO3 in 4[thin space (1/6-em)]:[thin space (1/6-em)]1 molar ratio were taken in 6 ml OLAM and 0.05 ml TOP in a three neck flask. After the complete formation of Ag@Ni at 170 °C, the reaction system was cooled. At 80 °C under N2 gas atmosphere, Zn(st)2 and Mg(st)2 in the required stoichiometry were added to the solution and reaction was conducted at 280 °C for 1 h. The washing procedure was the same as that adopted for the pure system.

Characterization

The crystalline phases of the products were determined by X-ray powder diffraction (XRD) by using a Bruker AXS D8SWAX diffractometer, with Cu Kα radiation (λ = 1.54 Å), employing a scanning rate of 0.5° S−1 in the 2θ range from 30°–80°. For XRD measurement the TCE solution of the NCs was drop cast over an amorphous silicon sample holder till a thin layer visible to the naked eye was formed. Transmission electron microscopy (TEM) images, high angle annular dark field scanning TEM (HAADF-STEM) images and energy dispersive spectra (EDS) were taken using an ultra-high resolution field emission gun transmission electron microscope (UHR-FEG TEM, JEM-2100F, Jeol, Japan) operating at 200 kV. For the TEM observations, the sample was dissolved in TCE and was drop cast on a carbon coated copper grid. The room temperature optical absorbance of the samples was measured by a Varian Cary 5000 UV-VIS-NIR spectrometer. Room temperature photoluminescence (PL) measurements were carried out with a fluorescence spectrometer (Hitachi, F-2500). Nitrogen sorption isotherms were obtained using a Quantachrome Autosorb 1C surface area analyzer at 77 K.

Thin film preparation and measurements

Long chain OLAM was first replaced by n-butylamine by standard ligand exchange procedures. The n-butylamine capped nanomaterials were found to be well soluble in octane. A 2 × 2 mm2 ITO substrate was first etched in the middle by using zinc dust and 1 N HCl solution. The substrate was further cleaned with de-ionised (DI) water and ethanol thoroughly and dried at 120 °C in N2 atmosphere for 4 h. Then a total 0.2 ml of NHS/NC solution in octane (30 mg ml−1) was spin cast onto a clean etched area of the ITO substrate at a speed of 1000 rpm for 60 s, which was then dried for 15 min in air. After each spin coating the ITO part was cleaned with octane. The film was annealed in Ar atmosphere for 3 h at 250 °C. Electrical contact was made by Ag paste with ITO parts. The photoresponse of the sample was measured using a Keithley Electrometer 6517A and a Keithley multimeter 2000 using a computerized program.

Photocatalytic activity study

Photocatalytic activities of Mg doped ZnO and Ag@Ni/Zn1−xMgxO samples were analyzed through the Rhodamine B dye degradation technique in the presence of a light source (KRATOS, Analytical instruments, universal arc lamp supply: 250 Watt, 150XE, model no. 1152). Rhodamine B (RhB) and the catalysts (0.02 g, n-butylamine capped ZnO and Ag@Ni/Zn1−xMgxO NCs calcined at 250 °C) were added to a 100 ml beaker containing 67 ml of de-ionized water. After ultra-sonication, RhB dye (0.5 ml, 1 mmol l−1) was injected into the above solution and was stirred for 2 h in the dark. Photocatalytic tests were conducted and 3 ml samples at different reaction times were centrifuged for the absorbance measurement using a Varian Cary 5000 UV-VIS-NIR spectrophotometer.

Results and discussion

Fig. 1 shows the XRD pattern of the pure ZnO and Zn1−xMgxO NCs where x varies from 0.05 to 0.5. The characteristic peaks are found to be well matched with the hexagonal wurtzite structure of ZnO (JCPDS no. 36-1451, space group P63mc). No impurity phase related to the formation of MgO phase was observed up to 30% doping. At higher doping concentration (40%), a peak indexed as (200) associated with cubic MgO appears. The smaller ionic radius of Mg2+ (0.057 nm) compared to Zn2+ (0.06 nm) generates lattice strain in the Zn1−xMgxO alloy system. Size (D) and lattice strain (η) influence the shape parameters of XRD peaks and can be simultaneously evaluated using the Williamson–Hall equation:35,36
 
image file: c6cp00375c-t1.tif(1)
We have plotted image file: c6cp00375c-t2.tif against image file: c6cp00375c-t3.tif to obtain the crystal strain from the slope, and the particle size from the intercept of the straight line. Table S1 (ESI†) also shows the values of particle size and crystal strain of the NCs. Two types of strain can be generated in a doped crystal, viz. tensile strain and compressive strain depending on the relative ionic radius of host and dopant ions. ZnO and Zn1−xMgxO samples reveal compressive strain as summarized in Table S1 (ESI†). This suggests that the strain plays an important role during the growth process.
image file: c6cp00375c-f1.tif
Fig. 1 XRD pattern of pure and Mg(II) alloyed ZnO (Zn1−xMgxO) nanocrystals up to x = 0.5. Impurity peaks related to cubic MgO appear at x = 0.4.

The XRD pattern of the as-prepared Ag@Ni core–shell nanoparticles is shown in Fig. S1a (ESI†). The diffraction peaks are assigned to fcc Ag and Ni. Fig. S2 (ESI†) shows XRD peaks of the Ag@Ni/Zn1−xMgxO heterostructures. The XRD patterns reveal three crystalline phases of the samples which are fcc Ag, fcc Ni, and hexagonal wurtzite ZnO. The heterostructure formation does not affect the solubility of Mg doping in the ZnO lattice.

TEM analysis

Fig. 2a shows the large area TEM image of nearly monodispersed ZnO NCs with a very low degree of agglomeration. The inset shows a tilted view of single NC which depicts the pyramid shape of the particle. The average dimensions of the pyramid are found to be 30 ± 5 nm along the height with a basal length of nearly 37 ± 5 nm. Fig. 2b shows the HRTEM image of a single nanopyramid along the tip direction, i.e. c-axis. The inner part shows higher contrast than the outer region of the pyramid. The image shows a perfect hexagonal base of the pyramid. The corresponding FFT pattern in Fig. 2c indicates that the NCs are single crystalline. The assigned planes from the FFT pattern are (100), (010), (1[1 with combining macron]0) and (110). The nanopyramid grows along the (110) direction (with d = 0.16 nm). The surface is found to be terminated by the (100) group of (d = 0.28 nm) planes which are parallel to basal edges. The side surfaces of nanopyramids consist of polar {101} facets which are the most exposed surfaces in the NCs. Oleylamine, which was used as activating and capping agent in our synthesis protocol, has a strong effect on stabilizing polar facets like {001}, {101} etc. The polar head –NH2 of the oleylamine molecule can bind to the thermodynamically unstable Zn-rich facets and can stabilize a near perfect pyramid shape.
image file: c6cp00375c-f2.tif
Fig. 2 (a–c) Large area TEM, HRTEM and SAED pattern of a single nanocrystal along the c-axis. Inset of (a) shows the tilted view of a nanocrystal. (d–f) Large area TEM, HRTEM and SAED of 10 mol% Mg alloyed ZnO self-assembled nanocrystal. Inset of (d) shows the tilted view of self-assembly. (g and h) TEM and HRTEM of 20 mol% Mg alloyed ZnO multipetal nanoflowers. (i) FFT pattern of yellow squared area in (h) showing the presence of twin (010) and (020) planes. (j and k) Large area and closer view of 30 mol% Mg alloyed ZnO showing the decrease of assembly size. (l) HRTEM showing the presence of (002) planes and the inset shows its FFT pattern. (m–o) TEM, HRTEM and FFT of 40 mol% Mg alloyed ZnO showing a decrease of particle size.

Fig. 2d and g show large area TEM images of 10% and 20% Mg alloyed assembled ZnO NCs. Although a morphological change can be observed, the colloidal stability and well dispersed nature of the product remains unaltered. The general image and the tilted view in Fig. 2d inset predicts that the NCs are formed by in situ self-assembly but maintained the hexagonal base nature. The morphology of 20% Mg doped ZnO NCs is flower like without a perfect hexagonal base. Normally homovalent dopant like Cd2+ or Mg2+ changes the morphology from pyramid to nanowire/nanorod by accelerating the growth along the c-axis of ZnO.28,29 The present observation of self-assembly of NCs is quite different from the previous results. The degree of self-assembly increases with Mg concentration. The HRTEM and the corresponding FFT patterns in Fig. 2e and f show a single crystalline nature of the NCs, although a large amount of void spaces exist in the nanoflower. Here we believe that an oriented attachment (OA) process drives the formation of the self-assembled structure. The OA process in colloidal synthesis generally occurs in the presence of inorganic anions which can be absorbed onto a particular facet of NC. Consequently it decreases the solubility of NCs and accelerates the OA of facets.37 Such a possibility can be ruled out in our synthesis protocol as the reaction mixture does not contain any inorganic anions since both the Zn and Mg sources were supplied from stearate salts. The dopant ions can generate or increase the surface charge/local dipole moment by introducing surface defect states (vide infra in Photoluminescence properties section) and increase the chemical attraction forces between the NCs. This favors the attachment of similar crystallographic planes or facets of two or more NCs and leads to nearly self-assembled single NC with higher volume through the OA process. The sizes of the assembled alloyed NCs (10% Mg doped ∼45 nm, 20% Mg doped ∼60 nm) are found to be higher than the pure ones. The Mg alloyed ZnO nanoflowers can be considered as the self-assembly of some small sized nano-‘petals’ which are formed at the early stage of the growth process and undergo coalescence during the latter stage to generate nearly single crystalline nanoflowers. Homovalent (Co2+, Mn2+) or aliovalent ion (In3+, Ga3+) doping in ZnO NCs decreases the crystallite size.12,38 The decrease of crystallite size upon doping can be explained by the Gibbs–Thomson relationship:38

 
image file: c6cp00375c-t4.tif(2)
where Sr is the solubility of the crystallite of radius r, SB is the solubility of the corresponding bulk material, Vm is the molar volume, and σ is the specific surface energy of NCs. The dopant residing at or near the surface causes lattice strain (also observed from XRD analysis) that elevates the specific surface energy, σ, and increases the crystallite solubility. So initially, the Mg alloying decreases the crystallite size of ZnO and decreased crystallites undergo coalescence to give self-assembled morphology/nanoflowers.

As there was no change in surface protecting ligand on Mg alloying, we can speculate that Mg alloying not only decreases the particle size but also introduces some defects (missing one or more atoms or interstitial ions) in the non-polar facets of wurtzite ZnO (which will be discussed in the Photoluminescence properties section, see later). Such defects can generate a local dipole moment at the facet and can accelerate the coalescence along that facet. Although most of the particles in 10% Mg alloyed ZnO were found to be single crystalline, for 20% Mg doped ZnO NCs the presence of twin planes is often observed. The imperfect OA leads to the formation of twin boundaries. A representative HRTEM image is shown in Fig. 2h. The FFT pattern of the image (Fig. 2i) shows the twin planes of [010], (020) and [1[1 with combining macron]0]. The perpendicular facets of the [010] direction are {001} and {002}. The würtzite ZnO crystal consists of alternating planes composed of fourfold tetrahedrally coordinated Zn2+ and O2− ions, stacked alternately along the c-axis. So the Zn2+-rich positively charged {001}/{002} facets of one NC and O2− rich negatively charged {001}/{002} polar facets of another might come closer to undergoing OA during the growth process. Fig. S3 in ESI† shows an atomic model of OA along the [010] direction. Also the {010} and {110} facets are involved in OA along the [1[1 with combining macron]0] direction. Although the {010} and {110} facets are neutral (Fig. S4, ESI†), incorporation of Mg2+ ions in interstitial lattice positions might create local charges which accelerate the OA process along the [1[1 with combining macron]0] direction.

Fig. S5–S7 in ESI† show TEM images from early reaction times to understand the OA growth of 20% Mg doped ZnO NCs. TEM images in Fig. S5 (ESI†) of reaction products collected at 15 s indicate the lowest size (20–30 nm) assembly of nanoparticles. Each assembly consists of 4–5 nanocrystals with size range of 5–10 nm. Fig. S5b (ESI†) shows the grain boundary region. The appearance of twin (010), ([1 with combining macron]00) and (1[1 with combining macron]0) planes in FFT (Fig. S5, ESI†) of the yellow square region suggests an OA process. So Mg dopants reduce the crystallite size and induce the OA process. Fig. S6 (ESI†) shows the TEM images after 1 min of reaction. The marked area in Fig. S6a (ESI†) shows assembly of smaller assemblies. The closer view shown in Fig. S6b (ESI†) shows nearly 6–7 small assemblies (20–25 nm in size) further undergo coalescence to form larger clusters. FFT from different grain boundary areas show that the OA process is operating along [011] and [010] directions. Fig. S7 (ESI†) shows TEM images of 10 min reaction product. Most of the self-assembled clusters are 35–50 nm in size (Fig. S7b, ESI†). Twin structures in FFT (Fig. S7c, ESI†) indicate clear evidence of an OA process.

30% Mg alloyed ZnO samples show a decrease in the size of nanoflowers as evidenced from Fig. 2j. The nanoflower assembly only contains 3–4 petal units nearly ∼15 nm in size (Fig. 2k). The HRTEM image in Fig. 2l shows a single petal with growth along the [002] direction. The HRTEM image also shows the presence of an amorphous layer probably coming from surface segregation of MgO moiety which impedes the OA process resulting in a decrease in nanoflower size. The 40% Mg doped ZnO sample does not show any self-assembly or oriented attachment (Fig. 2m). A large decrease in nanopyramid height (thickness ∼5–7 nm) is found from the HRTEM image in Fig. 2n. The corresponding FFT pattern in Fig. 2o shows the presence of the (110) plane. So on Mg alloying the crystallite size of ZnO NCs decreases on increasing Mg%. The reduced size Zn1−xMgxO NCs participate in oriented attachment to yield a free-standing flower-like morphology and the degree of OA is maximum for the 20% Mg alloyed system. For higher alloying % the OA process is quenched by the amorphous layer formation around each NC. The actual alloy percentage of Mg in ZnO has been estimated by EDS analysis and is given in Fig. S8 in ESI.† Fig. S9 and S10 (ESI†) show the elemental mapping and EDS line scan of the 20% Mg alloyed ZnO sample which shows a homogenous distribution of Mg in the ZnO system.

Fig. S11a (ESI†) shows a representative TEM image of the as-synthesized Ag@Ni core–shell structure. The NCs have usually spherical or prolate type morphology. The HRTEM image in Fig. S11b (ESI†) indicates that the particle has two distinct regions: a darker core region with a lighter shell. The higher Z value of Ag at the core has higher electron scattering ability than the lighter element Ni in the shell region. The shell thickness varies from 1.5 to 2.5 nm. Fig. S11c (ESI†) shows the FFT pattern of the interfacial region of Ag and Ni phases as highlighted by the squared area in Fig. S11b (ESI†). The reconstructed HRTEM image by masking the yellow circled spots shows the (200) plane of Ag (d = 0.208 nm) (Fig. S11d, ESI†), whereas the white circled spots show the (210) plane of Ni (d = 0.178 nm) (Fig. S11e, ESI†). Some dislocations are observed in reconstructed HRTEM images. This dislocation may be formed by the lattice deformation at the core–shell interfacial region, which is often observed in quasi-epitaxial growth of core–shell or heterostructures.39,40 The selected area diffraction (SAED) pattern obtained from these nanoparticles (Fig. S11f, ESI†) shows diffraction rings that are mainly composed of fcc Ag and Ni.

Fig. 3a shows the large area TEM image of as-synthesized Ag@Ni/ZnO NHSs. The image depicts that ZnO nanopyramids are connected with one or more spherical shape and darker in contrast Ag@Ni metal moieties. Interestingly the morphology of ZnO does not change in the presence of Ag@Ni seeds during the synthesis. Fig. 3b shows the HRTEM image of a single heterostructure at the site of interest where the interface between metal and semiconductor is believed to be formed. The metal area shows a distinct core–shell nature with the darker core as Ag and the lighter shell Ni. From literature7 study of Au/ZnO heterostructures, we can generalize that in most cases metal NC (Au) are preferentially situated at the tip or base of ZnO pyramids with growth of ZnO along the [001] direction over Au(111) planes. In the present case, Ag NCs are randomly situated at the tips and facets of ZnO pyramids. Fig. 3c shows the FFT pattern obtained from the yellow squared area of Fig. 3b, i.e. the pure ZnO region. The reconstructed HRTEM image (Fig. 3d) shows the presence of (102) and (110) planes of ZnO. Fig. 3e shows the FFT pattern obtained from the orange squared area of Fig. 3b. The FFT image contains a large number of spots which are the reflections of planes from both Ag@Ni and ZnO parts as the FFT was taken at the interface region. We can establish an epitaxial growth of ZnO onto Ag from FFT. The calculated d value from the white circled spots is 0.235 nm, which is the reflection of the Ag(111) plane. Another set of spots is found in nearly the same area (cyan spots). The plane from these spots is found to be (110) of ZnO with a d-spacing value of 0.161 nm. The simulated HRTEM image by masking these two colored spots is presented in Fig. 3f, which shows a coincidence of Ag(111) and ZnO(110) planes. We can generalize the epitaxial relation by the periodic arrangement of 2Ag(111) planes (2 × 0.235 nm = 0.47 nm) with 3ZnO(110) planes (3 × 0.16 nm = 0.48 nm). The coincidence of 2d (111) of Ag with 3d (110) of ZnO is highlighted by the straight lines (drawn for visibility) in Fig. 3f. The atomic arrangement of (111) planes of Ag and (110) of ZnO and matching between 2d (111) with 3d (110) planes is shown in Fig. S12 (ESI†).


image file: c6cp00375c-f3.tif
Fig. 3 (a) TEM image of Ag@Ni/ZnO NHS. (b) HRTEM image of NHS at the site of interest. (c) FFT pattern of yellow squared area, the pure ZnO region. (d) Reconstructed HRTEM image showing the presence of (110) and (102) planes of ZnO. (e) FFT pattern of orange squared area: interface region. The highlighted spots show the same direction of growth of Ag(111) and (110) of ZnO. (f) Reconstructed HRTEM image by masking the white and cyan circled spots in the FFT pattern, showing the minimal lattice mismatch between Ag(111) and (110) of ZnO planes.

Fig. 4a shows a typical large area TEM image of Ag@Ni/Zn0.8Mg0.2O metal alloyed semiconductor NHS. A striking morphological change of the semiconductor part is found similar to non-metal decorated systems. The inverted bright field TEM image (Fig. 4b) shows a flower-like morphology decorated with spherical metal NCs. The higher magnification view in Fig. 4c shows that each nanoflower is connected with multiple metal NCs. Such hybrid heterostructures are colloidally stable in ambient conditions over several months. Fig. 4d shows a typical HRTEM image of a heterostructured nanoflower where both the Ag and alloyed ZnO are present. The image clearly shows that the Ag NC is attached to three nanopetals. To verify the epitaxy formation the FFT pattern from the interface region is depicted in Fig. 4e. The (111) plane for Ag corresponding to a d1 value of 0.23 nm has been identified and is marked with a white circle. Along the same direction of the Ag(111) plane we also identified the (012) plane of ZnO with a d-spacing value of 0.191 (d2) nm, highlighted with a cyan circle. The simulated HRTEM image in Fig. 4f formed by masking these two set of spots shows the interface region of Ag and ZnO. The lattice mismatch between Ag(111) and ZnO(012) planes was calculated using the following formula:

 
image file: c6cp00375c-t5.tif(3)
The value of η is found to be ∼18% based on the estimated d values from the HRTEM image. The prediction of epitaxial growth is consistent with earlier reports for wide lattice mismatch between metal and semiconductor.41


image file: c6cp00375c-f4.tif
Fig. 4 (a) Large area TEM image of Ag@Ni/Zn0.8Mg0.2O NHS. (b) Inverted bright field TEM image of the same NHS showing the nanoflower morphology of alloyed semiconductor connected with multiple metal core–shell dots. (c) A closer view of the sample. (d) HRTEM image of a part of NHS showing the connectivity of the metal part with alloyed semiconductor. (e) FFT pattern of orange squared area, the interface of Ag and ZnO in (d) shows the same direction of Ag(111) and ZnO(012) planes. (f) Reconstructed HRTEM image by masking the white and cyan spots in the FFT pattern. (g) FFT pattern of yellow squared area of (d) showing the presence of twin (012) and (020) planes of ZnO. (h) Reconstructed HRTEM image from FFT showing the presence of dislocations.

The FFT pattern obtained from the yellow square area (pure alloyed ZnO region) in Fig. 4d is presented in Fig. 4g. We have identified mainly (020) and (012) planes of ZnO viewed along the a-axis. A closer view of the (012) spots shows that each spot is the overlap of two distinct spots highlighted in yellow circles. The reconstructed HRTEM image from the FFT (in Fig. 4h) shows the presence of twin (012) and (020) planes of ZnO. The yellow square area in Fig. 4d, which is the joining region of two ZnO petals, is mainly single crystalline in nature (as observed from the reconstructed HRTEM image in Fig. 4h) and the two petals are attached by the twin (012) plane of ZnO. The reconstructed HRTEM image also shows a considerable amount of dislocation (black arrows) at the joining region. This kind of defect is normally observed for imperfect OA involving multiple smaller NCs. To further understand the growth process of this metal alloyed semiconductor heterostructure, we quenched the reaction at 10 min and carried out TEM analysis for the intermediate product. Fig. S13 in ESI† shows the TEM images of Ag@Ni/Zn0.8Mg0.2O sample collected at 10 min of reaction. The images indicate mainly metal (Ag)–semiconductor (ZnO) dimers where the ZnO part has 10–15 nm in size (much smaller than the final nanoflower dimensions). The entire metal domain is covered with smaller ZnO nanoparticles. The image also shows the presence of some free-standing ZnO (not connected with the metal domain). So these free-standing ZnO NCs and Ag–ZnO dimer might undergo OA to give beautifully grown metal-decorated alloyed nanoflowers upon complete reaction for 60 min. The metal domain size remains unaltered and does not undergo coalescence (metal–metal attachment). The coalescence occurs between Ag@Ni/Zn1−xMgxO and pure Zn1−xMgxO, one Ag@Ni/Zn1−xMgxO with another metal-decorated Zn1−xMgxO. The final product appears to be a nanoflower decorated with multiple metal domains accompanied with multiple crystal plane dislocations.

Optical properties study

The effective band gap was investigated from the UV-Vis absorption spectra of the as-prepared samples at room temperature (Fig. 5a). An abrupt absorption edge is observed at about 362 nm corresponding to the characteristic absorption of pure ZnO NCs. The absorption band edge shifts to 333 nm with increasing Mg content. Here ZnO is a direct band gap semiconductor. The optical band gaps of the samples were calculated using the following classical Tauc relation of optical absorption near the band edge of the semiconductor:
 
(αhν)2 = C(hν − Eg)(4)
where α is the optical absorption coefficient, hν is the incident photon energy, and C is a constant. The band gap Eg was determined by plotting (αhν)2 as a function of hν and extrapolating the linear absorption edge of the curve to intersect the energy axis as shown in Fig. 5b. The variation of band gap with dopant (Mg) concentration (x) is shown in Fig. 5c, indicating linear behavior.42 Therefore blue shifting of the band gap in Mg alloyed ZnO NCs satisfies Vegard's law corresponding to band gap variation in semiconducting alloys. It can be inferred from the above discussion that band gap tuning of the Zn1−xMgxO NCs towards higher energy is a consequence of Mg alloying.
image file: c6cp00375c-f5.tif
Fig. 5 (a) Absorbance spectra of pure ZnO and Zn1−xMgxO nanocrystals. The absorbance edge shifts toward lower wavelength on Mg alloying. (b) Tauc plot of different samples showing the change of band gap. (c) Change of band gap with x. (d) Absorbance spectra of different Ag@Ni/Zn1−xMgxO NHS. The green shaded area depicts the plasmonic absorbance of the metal part, and the light blue region depicts the band gap absorbance of Zn1−xMgxO.

Fig. S1b in ESI† shows the absorbance spectrum of Ag@Ni core–shell nanoparticles. Pure Ag NCs which were used as a core material show an LSPR band centred at 410 nm (Fig. S1c, ESI†) and pure Ni particle shows a plasmon band near 330 nm according to the Mie theoretical calculation.43 The LSPR band for Ag@Ni core–shell nanoparticles is blue shifted by 19 nm (absorption band centered at 391 nm) compared to pure Ag NCs and also broadened. When two heterometals have an interparticle separation less than 2 nm there might be a plasmonic coupling between them. This can be a consequence of the overlapping of different plasmon modes, such as capacitive plasmon mode and charge transfer plasmon mode.44,45 In the reaction vessel Ni2+ ions were reduced by TOP and started to grow upon the surface of the Ag seeds and form the Ag@Ni heterometallic system. Therefore, due to the coupling of two plasmonic bands, the LSPR band of the core–shell structure is broadened and damped. In some research works such as Ag core/Au shell nanoparticles,46–48 the plasmon band is shifted towards the plasmon band of the shell material following Mie's theory.49 The presence of the Ni shell shifts the plasmon band of Ag towards the Ni plasmon band (lower wavelength).50–52 Therefore we can infer that the blue shifting of the plasmon band in the Ag@Ni system is due to the Ni shell formation upon Ag.

The exciton absorption of Ag@Ni/Zn1−xMgxO heterostructure as demonstrated in Fig. 5d is similar to that of Zn1−xMgxO. So the Ag@Ni/Zn1−xMgxO heterostructure does not affect the band gap widening of the semiconductor. The most interesting observation is that the plasmon absorption wavelength of the Ag@Ni core–shell structure (at 391 nm) is found to be red shifted by 50–90 nm. There might be two possible reasons for such variation. Firstly it might be caused by the larger refractive index of the semiconductor surrounding the heterometal core–shell structure. This behavior is consistent with previous observations of plasmon shift to higher wavelengths, i.e. lower energy, in the presence of a high refractive index environment.53,54 The LSPR frequency of metal NCs can be expressed as:

 
image file: c6cp00375c-t6.tif(5)
where ωsp is the LSPR frequency, ωp is the bulk plasmon frequency, εm is the dielectric constant of the surrounding environment and γ is the damping factor. So an increase of dielectric constant of the environment of metal NCs results in red shifting of the LSPR band position. The dielectric constant of ZnO is much higher than that of the TCE solvent.55 Secondly the plasmon field of the heterometallic core–shell provides favorable conditions for plasmon–exciton coupling and shifts the plasmon peak towards higher wavelengths. Since the Ag@Ni core–shell structure experiences two different dielectric media (one due to the semiconductor and the other due to the solvent) the LSPR peak is damped and broadened.

Photoluminescence properties

Photoluminescence (PL) properties of ZnO NC upon Mg doping and Ag@Ni/Zn1−xMgxO heterostructure were examined at room temperature. Fig. 6 shows the PL profiles in the range 350 nm to 650 nm of pure and alloyed NCs with an excitation source of 330 nm. ZnO has a highly ionic lattice, the valence band (VB) is composed of oxygen 2p orbital hybridized with Zn 3d states and the conduction band (CB) is composed of mainly Zn2+ excited states. So alloying with homovalent Mg2+ ion not only changes the valence band and conduction band structure as we already found a widening of the band gap, but also changes the defect chemistry of ZnO. Normally the PL structure of ZnO NC (of any size) shows two types of emission: (i) near band edge emission (NBE) and (ii) deep level emission. The PL profiles of as-synthesized NCs show broad emission featuring multiple humps in the UV-Vis region. The emission profiles were deconvoluted based on a Gaussian distribution to achieve the multiple emission bands. The colors of the deconvoluted curves depict the color of emission in the particular region. Three sets of deconvoluted peaks were obtained for all the samples. The peaks in the UV region (370–410 nm) for pure ZnO are composed of excitonic recombination. The excitonic recombination band is found to be gradually blue shifted with Mg2+ alloying from 384 nm to 353 nm, similar to the absorbance band due to the increase of band gap with Mg2+ alloying. Several kinds of defects and vacancies are found to be responsible for the origin of the broad band emission.56 The defect states in ZnO can be generalised into two categories: (i) shallow trapped states which are closely situated to CB and are mainly composed of optically active zinc-related intrinsic defect states like zinc interstitial (Zni)57 and zinc vacancy (VZn), (ii) deep trapped defect states like oxygen vacancy (VO), oxygen interstitial (Oi), zinc vacancy, oxygen at zinc lattice site (OZn). Previous experimental results proved that Zni states are 0.22 eV below the CB.58 So the violet and blue emissions at 395 nm and 416 nm for pure ZnO can be assigned as NBE and Zni level emissions. With increase in Mg2+ alloying the number of deconvoluted peaks and relative intensity with respect to exciton recombination increases from 2 to 3, which indicates the formation of more Zni defects on Mg alloying along with band gap widening. The blue PL band may originate from the recombination of photogenerated electrons in extended Zni states, which reside slightly below the normal Zni states59 or the recombination of electrons from CB to oxygen interstitial site (Oi) which reside at a much deeper level in the band gap.60 Although the formation energy of Zni is high, a recent study suggests that a non-equilibrium synthesis will accelerate the formation of Zni.59 We found a non-equilibrium growth process during the synthesis of NCs as we described in the TEM section. So the blue emission centred at 450 nm for pure ZnO NCs can be described as Zn interstitial or Zn interstitial-related complex defects.61 All the alloyed NCs show a prominent blue emission band at 430–450 nm. Although the blue emission band has higher intensity than the excitonic recombination band for pure ZnO NC, both 5% and 10% Mg doped ZnO NCs show a relatively lower intensity of blue emission than the excitonic recombination: this might be a result of a lower concentration of Zni defects states. The 20% and 30% Mg alloyed ZnO NCs show higher probability of Zni to VB recombination than band to band transition as evidenced from the increased intensity of the former. The green emission band at 480–600 nm is deconvoluted into three peaks for pure ZnO NCs. This broad green emission band may be due to (i) recombination of photogenerated electrons in the conduction band with the deep state defect levels (VO, VZn, Oi, Zni or OZn situated very close to the valence band) or (ii) recombination of electrons in the shallow trap states (mainly Zni states) with deep defect states. XPS analysis for ZnO and alloyed ZnO NCs (Fig. S14, ESI†) shows the presence of oxygen vacancy and Oi which appeared at lower binding energy than pure O 1s. So the oxygen vacancy related defects states like neutral (VO), singly (V+O) or doubly ionized (V++O) states might be involved in luminescence processes to generate the broad green emission. The intensity and broadness of the green emission band centered at the 490–500 nm region is found to be increased with increasing the Mg2+ alloying percentage and reaches a maximum for 30% Mg alloyed ZnO NCs. The long tail beyond 550 nm may relate to Oi defects, as a previous study indicates an Oi related PL peak appears at the higher wavelength region than the oxygen vacancy related emission.62,63 So both the Zni and oxygen vacancy related defect states are found to be increased with Mg alloying up to 20% Mg alloyed sample, except for 5% Mg alloyed NCs where Zni related PL was quenched. Strikingly, all the PL intensity for 30% Mg alloyed NCs was found to be much lower compared to the other samples. The formation of an amorphous MgO layer (which is concluded from TEM analysis) around the crystalline NC may be responsible for quenching the visible emission.
image file: c6cp00375c-f6.tif
Fig. 6 (a–e) PL emission profiles of different Zn1−xMgxO nanocrystals, where x varies from 0 to 0.3.

Fig. S15 in ESI† shows the PL profiles of Ag@Ni/Zn1−xMgxO NHSs with different amounts of Mg2+. The excitonic recombination band is found to be blue shifted with increased Mg2+ alloying as we observed for pure Zn1−xMgxO NCs. Blue emission related to the recombination of electrons in shallow trap Zni states to CB is present for all the heterostructures. The visible emission beyond 450 nm is found to be quenched for all the heterostructured NCs. This indicates a decrease in electron–hole recombination through defect derived states. Metal NCs are epitaxially connected with Zn1−xMgxO NCs/nanoflowers as confirmed from TEM analysis and may act as an electron sink. So the photoexcited electrons may prefer to go towards metal sites. Scheme 1 shows the relative band alignment of Ag and ZnO. From a literature study,64 the Fermi energy (EF) level of Ag (EF = −4.7 eV) is found to be 0.9 eV below the conduction band edge of ZnO. We found Zni related PL for all heterostructures as it resides 0.22 eV below the CB and above the EF of Ag. So upon excitation with 330 nm radiation, the photoexcited electrons in CB can decay through Zni defect states to VB giving the blue emission, or may transfer to Ag metal. The oxygen vacancy related defect states are the deep trap states lying far below the EF of Ag. So there is a lower probability of recombination of electrons with VO, V+O or V++O states which results in diminution or quenching of green emission. The probable decay paths for heterostructures are presented in Scheme 1. Therefore, metal nanoparticle attachment is very beneficial for the transfer of photoexcited electrons from Zn1−xMgxO to Ag.


image file: c6cp00375c-s1.tif
Scheme 1 Relative band alignment of Ag and ZnO/Zn1−xMgxO and probable transitions in the heterostructure.

Photocatalytic activity

Prior to photocatalytic activity testing the ligands were removed by a ligand exchange process followed by thermal annealing. Oleylamine was first replaced by n-butyl amine then the washed NCs were annealed at 250 °C to remove the organic part completely. The change of RhB dye concentration was monitored by measuring the optical absorbance at 553 nm of the suspension at 30 min time intervals for only alloyed ZnO NCs and 5 min time intervals for Ag@Ni/Zn1−xMgxO heterostructures (as this material shows high photocatalytic activity). To compare the photoreduction rate of RhB in the presence of different catalysts, we plotted C/C0vs. t as depicted in Fig. 7, where C0 is the concentration of dye at equilibrium established in dark conditions and C is the concentration of nondegraded dye after different irradiation time intervals. The photodegradation of RhB in the presence of pure ZnO was relatively slow, with 60% of RhB degraded after an illumination time of 90 min (Fig. 7a). Whereas nearly 77%, 88% and 98% RhB was found to be photodegraded in the presence of 5%, 10% and 20% Mg alloyed ZnO samples respectively in 90 min. We found a decrease in degradation rate for the 30% Mg alloyed ZnO sample in similar conditions where only 90% RhB was found to be degraded. The photoreduction kinetics of RhB follows a pseudo-first order, expressed by ln(C0/C) = Kt, where K is the apparent rate constant. Fig. 7b shows plots of ln(C0/C) vs. t for ZnO and Zn1−xMgxO samples. Zn0.8Mg0.2O has the highest rate constant value of 0.062 min−1 which is nearly ∼3.5 times higher than that of pure ZnO (k1 = 0.018 min−1). The increase of photocatalytic activity can be attributed to two reasons: (i) Mg alloying which increases the conduction band potential value making the system more reducing, (ii) change of morphology on Mg alloying. On increasing the amount of Mg the shape of NCs was found to be changed from pyramid to rough surface hexagonal base self-assembled crown-shaped and finally to hierarchical shaped nanoflowers. So on Mg alloying the surface area of NCs increases which results in an increase of dye adsorption on the NC surface. The flower-like morphology was found to be lost (from TEM analysis) for 30% Mg alloyed ZnO which gives rise to a decrease in catalytic activity compared with the 20% alloyed sample.
image file: c6cp00375c-f7.tif
Fig. 7 Light driven photodegradation of RhB dye in the presence of (a and b) alloyed ZnO nanocrystals, and (c and d) different Ag@Ni/Zn1−xMgxO NHS.

The effect of metal–semiconductor heterostructure on photocatalytic activity is shown in Fig. 7c and d. At only 60 min irradiance, 80% of RhB was found to be degraded in the presence of Ag@Ni/ZnO catalyst. Nearly 92%, 95% and 99% of RhB was degraded in the presence of 5%, 10% and 20% Mg alloyed metal–semiconductor heterostructures respectively. The maximum photodegradation rate constant value was found for Ag@Ni/Zn0.8Mg0.2O catalyst with 0.07 min−1, which is higher than the non-metal decorated counterpart. Metal decorated Zn0.7Mg0.3O does not show a sharp decrease in rate constant like the pure Zn0.7Mg0.3O sample. The rate constant value is found to be 0.068 min−1, very similar to Ag@Ni/Zn0.8Mg0.2O. The sheet-like morphology of Ag@Ni/Zn0.7Mg0.3O [Fig. S9, ESI†] and the decoration with multiple metal dots with large exposed surface area result in the adsorption of dye compounds in larger amounts compared to Zn0.7Mg0.3O. All the metal decorated catalyst was found to be magnetically separable by small bar magnets as depicted in Fig. S17 in ESI.† Fig. S20 in ESI† depicts the TEM image of recovered photocatalyst after one photocatalytic cycle which shows the coexistence of metal and semiconductor in a single nanoflower. The change of absorbance of the dye in the presence of catalyst is shown in ESI† (Fig. S18).

To access better understanding of the photocatalytic property enhancement with morphological change on Mg alloying, we performed Brunauer–Emmett–Teller (BET) gas sorption measurements. The nitrogen adsorption/desorption isotherms and the pore size distribution plots of different Ag@Ni decorated Zn1−xMgxO systems are shown in Fig. S19 in ESI.† The estimated BET specific surface areas and pore sizes of samples are listed in Table 1. The typical type IV nature of the curves and hysteresis loop indicates the mesoporous nature of the as-synthesis heterostructured materials according to the IUPAC classification. The BET surface area for Ag@Ni/ZnO is found to be 18.21 m2 g−1. The Barrett–Joyner–Halenda (BJH) pore size distribution curve indicates non-uniformity of pores in the range 1 to 15 nm with high population density at 1.5 nm and 3 nm. This distribution of pore size is for non-uniform size of ZnO nanopyramids in heterostructures and the presence of rough surfaces in ZnO as also observed from TEM analysis (Fig. 3a). The BET surface area increases with Mg alloying in the heterostructures and the values obtained for Ag@Ni/Zn0.9Mg0.1O and Zn0.8Mg0.2O heterostructures are 31.26 m2 g−1 and 38.32 m2 g−1. This increase of surface area can be attributed to an increase in oriented attachment which leads to self-assembled nanoflower formation on Mg alloying. The greater pore size contribution in the pore size distribution is found to be increased with Mg alloying, particularly for Ag@Ni/Zn0.8Mg0.2O where the contribution of pore sizes of ∼5 nm to 15 nm is higher than for the other samples. The formation of multipetal multilayer nanoflowers (Fig. 4b) for Ag@Ni/Zn0.8Mg0.2O is the main reason for the increase in pore size. Nanoflower morphology, high surface area and pore size assist the faster dye degradation.

Table 1 BET surface areas and total pore volumes of samples
Sample S BET (m2 g−1) Pore size distribution (nm)
Ag@Ni/ZnO 18.21 1 to 15
Ag@Ni/Zn0.9Mg0.1O 31.26 2.5 to 20
Ag@Ni/Zn0.8Mg0.2O 38.32 2.5 to 25

Photoresponse properties

To utilize the enhanced photoefficiency of the as-synthesized metal–semiconductor heterostructure, we fabricated a photodetector device using these materials. Fig. 8 shows the photoresponse curves for all the samples at a bias voltage of 5 V in the presence of light irradiance of 0.65 W cm−2 power density with 40 s time lapse. The yellow shaded areas show the current in the presence of light and the other areas are in the absence of light. The photocurrent gain, which is the ratio of current in the presence and absence of light (IPhoton[thin space (1/6-em)]:[thin space (1/6-em)]IDark), is found to be low for the Ag@Ni/ZnO system with a value of 11, and maximum for the Ag@Ni/Zn0.7Mg0.3O system with a value of 230. Mg alloying in Ag@Ni/ZnO increases the photocurrent value similar to the photocatalysis phenomenon, except for the Ag@Ni/Zn0.7Mg0.3O composition which reveals the most responsive device in spite of the poor photocatalytic activity compared with Ag@Ni/Zn0.8Mg0.2O. This suggests that probably the morphology of NCs does not affect the photoresponse behavior of the material. Mg alloying also has a striking effect on the current gain and decay nature during the repetitive photoresponse measurements. The pure Ag@Ni/ZnO system shows a slow gain and slow decay of photocurrent. Incomplete separation of electron–hole pairs upon exposure to light, and decay of photoexcited electrons through trap/defects states during dark conditions may result in the slow response nature. For the fabrication of a fast and stable photodetector, two major criteria must be fulfilled: (i) increase of lifetime of the exciton pair, i.e. complete separation of e− and h+, and (ii) fast recombination of the e− and h+ pair in the absence of light. These two criteria have been achieved for the Ag@Ni/Zn0.7Mg0.3O system where the maximum photocurrent gain value (∼230) is found. Moreover, the photocurrent gain and decay patterns appear to be relatively faster compared to other devices. The number of metal NCs on each semiconductor NC/nanoflower is found to be crucial in order to achieve a fast responsive device. Pure ZnO NCs are arranged with one or two metal NCs whereas the nanoflowers (Zn1−xMgxO, x = 0.05, 0.1, 0.2) and especially the nanosheets (x = 0.3) are found to be situated with three or more metal NCs, which increases the photon absorbance efficiency of each nanoflower through LSPR and gives rise to efficient charge separation.
image file: c6cp00375c-f8.tif
Fig. 8 Temporal photoresponses of different Ag@Ni/Zn1−xMgxO NHS. Photocurrent gain values are shown on the right.

For better understanding of the individual contributions of band gap exciton of Zn1−xMgxO and plasmon absorbance of Ag@Ni towards the photochemical and photophysical properties of Ag@Ni/Zn0.8Mg0.2O NHS, we investigated the photocatalytic and photoresponse properties using a cut-off filter at 390 nm. From both the experiments (Fig. S21, ESI†) we have found that the photoactivity is maximum under xenon light irradiance and minimum in the presence of visible light (>390 nm). UV excitation gives the intermediate value of rate constant of dye degradation and photocurrent gain. When the NHSs were excited only by visible light, NHSs absorb light by plasmonic excitation of Ag@Ni metal particles. Plasmon induced excited electrons of Ag@Ni do not have sufficient energy to overcome the Schottky barrier between metal and semiconductor components. This results in poor photoactivity of NHS only under visible light. UV band gap excitation of Zn1−xMgxO generates electrons which can easily transfer to Ag@Ni, preventing electron–hole recombination. Thus the separation of photogenerated charge carriers can promote photoactivity. The synergetic effect of the excitation of plasmon of metal and band gap exciton of semiconductor under xenon light illumination gives rise to the highest photoactivity. Plasmonic nanostructures can induce hot electron transfer, increase absorption coefficient and enhance the local electric field.32,41,65–67 Electromagnetic field enhancement predominates when plasmonic resonance and band gap exciton wavelengths overlap.68,69 The plasmonic absorption wavelength (391 nm) of Ag@Ni is very close to the band gap exciton (333–362 nm) of Zn1−xMgxO. This suggests that an increase of the local electric field is the major contributor to the significant increase of photocatalytic activity and photoresponse properties.

Magnetic properties

The magnetization as a function of the temperature and magnetic field of Ag@Ni/ZnO NHSs was studied using a superconducting quantum interference device (SQUID) magnetometer (Quantum Design MPMS, XL Evercool model). Zero-field cooled (ZFC) and field cooled (FC) magnetizations with temperature are displayed in Fig. 9a. A bifurcation between ZFC and FC curves occurs at about 50 K. The ZFC magnetization curve reveals a prominent peak around 27 K, known as the blocking temperature (TB). The value of TB depends on the volume of the magnetic phase. The much smaller value of TB compared to the Curie temperature (627 K) of bulk Ni suggests that the thickness of the Ni shell is very small. Magnetization vs. magnetic field (M–H) curves at 300 K and 5 K are depicted in Fig. 9b. At low temperature (5 K) the M–H curve clearly exhibits a hysteresis loop. The estimated coercivity at 5 K is 101 Oe. The absence of a hysteresis loop at 300 K indicates zero coercivity. The characteristic features of temperature dependent magnetization and M–H curves imply that the magnetic heterostructure is ferromagnetic below TB and is superparamagnetic above TB. Such magnetic behavior is very useful to isolate the magnetic heterostructures from solution using a small magnet at room temperature.
image file: c6cp00375c-f9.tif
Fig. 9 (a) Magnetization vs. temperature (M–T) plot of Ag@Ni/ZnO NHS at H = 100 Oe. (b) Magnetization vs. magnetic field (M–H) plots at 5 K and 300 K. The inset shows the presence of a distinct hysteresis loop at 5 K.

Conclusion

In summary we successfully synthesized monodispersed ZnO and Mg alloyed ZnO nanostructures with a solid solubility limit ∼30%. The morphological change of ZnO from nanopyramid to hierarchical shape with intact colloidal stability and well dispersed nature on Mg alloying has been explained based on an oriented attachment process. Mg alloying increases both the shallow trap and deep trap defect concentrations in the ZnO system. Magnetic plasmonic metal–semiconductor NHSs have been successfully synthesized using Ag@Ni as metal seed particles. Minimum lattice mismatch between the periodic arrangement of two (111) planes of Ag with three (110) planes of ZnO leads to epitaxial growth of Ag@Ni/Zn1−xMgxO heterostructures. Efficient epitaxy formation and Mg alloying increase the photoactivity of ZnO. A potent photocatalyst and fast response photodetector have been fabricated with Ag@Ni/Zn0.8Mg0.2O and Ag@Ni/Zn0.7Mg0.3O NHSs. Superparamagnetism has been found for Ag@Ni seeds which makes the photocatalyst magnetically separable.

Acknowledgements

The authors S. Paul and S. Ghosh sincerely acknowledge DST INSPIRE fellowship and CSIR, India respectively for providing fellowships during the tenure of the work.

Notes and references

  1. R. S. Selinsky, Q. Ding, M. S. Faber, J. C. Wright and S. Jin, Chem. Soc. Rev., 2013, 42, 2963–2985 RSC .
  2. A. Vaneski, A. S. Susha, J. Rodríguez-Fernández, M. Berr, F. Jäckel, J. Feldmann and A. L. Rogach, Adv. Funct. Mater., 2011, 21, 1547–1556 CrossRef CAS .
  3. R. S. Selinsky, Q. Ding, M. S. Faber, J. C. Wright and S. Jin, Chem. Soc. Rev., 2013, 42, 2963–2985 RSC .
  4. S. J. Pearton, D. P. Norton and F. Ren, Small, 2007, 3, 1144–1150 CrossRef CAS PubMed .
  5. Q. Lu, Z. Lu, Y. Lu, L. Lv, Y. Ning, H. Yu, Y. Hou and Y. Yin, Nano Lett., 2013, 13, 5698–5702 CrossRef CAS PubMed .
  6. S. Ghosh, M. Saha, S. Paul and S. K. De, Nanoscale, 2015, 7, 18284–18298 RSC .
  7. M. N. Tahir, F. Natalio, M. A. Cambaz, M. Panthöfer, R. Branscheid, U. Kolb and W. Tremel, Nanoscale, 2013, 5, 9944–9949 RSC .
  8. V. Etacheri, R. Roshan and V. Kumar, ACS Appl. Mater. Interfaces, 2012, 4, 2717–2725 CAS .
  9. N. Zhou, L. Polavarapu, N. Gao, Y. Pan, P. Yuan, Q. Wang and Q. Xu, Nanoscale, 2013, 5, 4236–4241 RSC .
  10. B. K. Patra, A. K. Guria, A. Dutta, A. Shit and N. Pradhan, Chem. Mater., 2014, 26, 7194–7200 CrossRef CAS .
  11. X. Yu, A. Shavel, X. An, Z. Luo, M. Ibáñez and A. Cabot, J. Am. Chem. Soc., 2014, 136, 9236–9239 CrossRef CAS PubMed .
  12. S. Ghosh, M. Saha and S. K. De, Nanoscale, 2014, 6, 7039–7051 RSC .
  13. M. Saha, S. Ghosh, V. D. Ashok and S. K. De, Phys. Chem. Chem. Phys., 2015, 17, 16067–16079 RSC .
  14. J. Chang and E. R. Waclawik, CrystEngComm, 2012, 14, 4041–4048 RSC .
  15. J. Li, H. Fan and X. Jia, J. Phys. Chem. C, 2010, 114, 14684–14691 CAS .
  16. Y. Jin, J. Wang, B. Sun, J. C. Blakesley and N. C. Greenham, Nano Lett., 2008, 8, 1649–1653 CrossRef CAS PubMed .
  17. N. Gogurla, A. K. Sinha, S. Santra, S. Manna and S. K. Ray, Sci. Rep., 2014, 4, 6483–6492 CrossRef CAS PubMed .
  18. M. Shim, A. Javey, N. W. S. Kam and H. Dai, J. Am. Chem. Soc., 2001, 123, 11512–11513 CrossRef CAS PubMed .
  19. W. K. Liu, K. M. Whitaker, K. R. Kittilstved and D. R. Gamelin, J. Am. Chem. Soc., 2006, 128, 3910–3911 CrossRef CAS PubMed .
  20. W. K. Liu, K. M. Whitaker, A. L. Smith, K. R. Kittilstved, B. H. Robinson and D. R. Gamelin, Phys. Rev. Lett., 2007, 98, 186804 CrossRef PubMed .
  21. A. W. Cohn, K. R. Kittilstved and D. R. Gamelin, J. Am. Chem. Soc., 2012, 134, 7937–7943 CrossRef CAS PubMed .
  22. X. Qiu, L. Li, J. Zheng, J. Liu, X. Sun and G. Li, J. Phys. Chem. C, 2008, 112, 12242 CAS .
  23. X. D. Zhang, M. L. Guo, C. L. Liu, L. A. Zhang, W. Y. Zhang, Y. Q. Ding, Q. Wu and X. Feng, Eur. Phys. J. B, 2008, 62, 417–421 CrossRef CAS .
  24. R. D. Shannon, Acta Crystallogr., Sect. A: Found. Crystallogr., 1976, 32, 751–767 CrossRef .
  25. A. Ohtomo, M. Kawasaki, T. Koida, K. Masubuchi, H. Koinuma, Y. Sakurai, Y. Yoshida, T. Yasuda and Y. Segawa, Appl. Phys. Lett., 1998, 72, 2466–2468 CrossRef CAS .
  26. S. C. Su, Y. M. Lu, Z. Z. Zhang, C. X. Shan, B. H. Li, D. Z. Shen, B. Yao, J. Y. Zhang, D. X. Zhao and X. W. Fan, Appl. Phys. Lett., 2008, 93, 082108 CrossRef .
  27. J. W. Chiou, H. M. Tsai, C. W. Pao, K. P. Krishna Kumar, S. C. Ray, F. Z. Chien, W. F. Pong, M. H. Tsai, C. H. Chen, H. J. Lin, J. J. Wu, M.-H. Yang, S. C. Liu, H. H. Chiang and C. W. Chen, Appl. Phys. Lett., 2006, 89, 043121 CrossRef .
  28. N. Zhang, X. Wang, Z. Ye and Y. Jin, Sci. Rep., 2014, 4, 4353–4360 Search PubMed .
  29. Y. Yang, Y. Jin, H. He, Q. Wang, Y. Tu, H. Lu and Z. Ye, J. Am. Chem. Soc., 2010, 132, 13381–13394 CrossRef CAS PubMed .
  30. P. Li, Z. Wei, T. Wu, Q. Peng and Y. Li, J. Am. Chem. Soc., 2011, 133, 5660–5663 CrossRef CAS PubMed .
  31. Y. Chen, D. Zeng, K. Zhang, A. Lu, L. Wang and D. L. Peng, Nanoscale, 2014, 6, 874–881 RSC .
  32. X. Zhang, Y. L. Chen, R. S. Liu and D. P. Tsai, Rep. Prog. Phys., 2013, 76, 046401 CrossRef PubMed .
  33. H. Guo, Y. Chen, X. Chen, R. Wen, G. H. Yue and D. L. Peng, Nanotechnology, 2011, 22, 195604 CrossRef PubMed .
  34. S. Sharma, B. Kim and D. Lee, Langmuir, 2012, 28, 15958 CrossRef CAS PubMed .
  35. B. D. Cullity, Elements of X-ray Diffraction, Addison-Wesley Pub. Comp. Inc., 1956 Search PubMed .
  36. Powder Diffraction, ed. R. E. Dinnebier and S. J. L. Billinge, RSC Publishing, 2008 Search PubMed .
  37. S. Ghosh, K. Das, G. Sinha, J. Lahtinen and S. K. De, J. Mater. Chem. C, 2013, 1, 5557–5566 RSC .
  38. D. A. Schwartz, N. S. Norberg, Q. P. Nguyen, J. M. Parker and D. R. Gamelin, J. Am. Chem. Soc., 2003, 125, 13205–13218 CrossRef CAS PubMed .
  39. O. K. Ranasingha, C. Wang, P. R. Ohodnicki, Jr., J. W. Lekse, J. P. Lewisab and C. Matranga, J. Mater. Chem. A, 2015, 3, 15141–15147 CAS .
  40. M. Grzelczak, B. Rodríguez-González, J. Pérez-Juste and L. M. Liz-Marzán, Adv. Mater., 2007, 19, 2262–2266 CrossRef CAS .
  41. S. K. Dutta, S. K. Mehetor and N. Pradhan, J. Phys. Chem. Lett., 2015, 6, 936–944 CrossRef CAS PubMed .
  42. J. Singh, P. Kumar, K. S. Hui, K. N. Hui, K. Ramam, R. S. Tiwaria and O. N. Srivastava, CrystEngComm, 2012, 14, 5898–5904 RSC .
  43. J. A. Creighton and D. G. Eadont, J. Chem. Soc., Faraday Trans., 1991, 87, 3881–3891 RSC .
  44. J. Lee, M. You, G. Kim and J. Nam, Nano Lett., 2014, 14, 6217–6225 CrossRef CAS PubMed .
  45. S. Ghosh, S. Khamarui, M. Saha and S. K. De, RSC Adv., 2015, 5, 38971–38976 RSC .
  46. D. Chen and C. Chen, J. Mater. Chem., 2002, 12, 1557–1562 RSC .
  47. H. Jiang, T. Akita, T. Ishida, M. Haruta and Q. Xu, J. Am. Chem. Soc., 2011, 133, 1304–1306 CrossRef CAS PubMed .
  48. A. Henglein, J. Phys. Chem. B, 2000, 104, 2201–2203 CrossRef CAS .
  49. H. Xu, Phys. Rev. B: Condens. Matter Mater. Phys., 2005, 72, 073405 CrossRef .
  50. Z. Zhang, T. M. Nenoff, J. Y. Huang, D. T. Berry and P. P. Provencio, J. Phys. Chem. C, 2009, 113, 1155–1159 CAS .
  51. M. Gaudry, E. Cottancin, M. Pellarin, J. Lermé, L. Arnaud, J. R. Huntzinger, J. L. Vialle, M. Broyer, J. L. Rousset, M. Treilleux and P. Mélinon, Phys. Rev. B: Condens. Matter Mater. Phys., 2003, 67, 155409 CrossRef .
  52. E. Cottancin, M. Gaudry, M. Pellarin, J. Lermé, L. Arnaud, J. R. Huntzinger, J. L. Vialle, M. Treilleux, P. Mélinon, J. L. Rousset and M. Broyer, Eur. Phys. J. D, 2003, 24, 111–114 CrossRef CAS .
  53. J. Lee, E. V. Shevchenko and D. V. Talapin, J. Am. Chem. Soc., 2008, 130, 9673–9675 CrossRef CAS PubMed .
  54. E. Shaviv, O. Schubert, M. Alves-Santos, G. Goldoni, R. D. Felice, F. Vallee, N. D. Fatti, U. Banin and C. Sonnichsen, ACS Nano, 2011, 5, 4712–4719 CrossRef CAS PubMed .
  55. Y. Yang, W. Guo, X. Wang, Z. Wang, J. Qi and Y. Zhang, Nano Lett., 2012, 12, 1919–1922 CrossRef CAS PubMed .
  56. A. Janotti and C. G. Van de Walle, Phys. Rev. B: Condens. Matter Mater. Phys., 2007, 76, 165202 CrossRef .
  57. B. Lin, Z. Fu and Y. Jia, Appl. Phys. Lett., 2011, 79, 943–945 CrossRef .
  58. E. G. Bylander, J. Appl. Phys., 1978, 49, 1188 CrossRef CAS .
  59. H. Zeng, G. Duan, Y. Li, S. Yang, X. Xu and W. Cai, Adv. Funct. Mater., 2010, 20, 561–572 CrossRef CAS .
  60. L. Han, L. Cui, W. Wang, J. Wang and X. Du, Semicond. Sci. Technol., 2012, 27, 065020 CrossRef .
  61. D. C. Look, G. C. Farlow, P. Reunchan, S. Limpijumnong, S. B. Zhang and K. Nordlund, Phys. Rev. Lett., 2005, 95, 225502 CrossRef CAS PubMed .
  62. S. S. Kurbanov, G. N. Panin, T. W. Kim and T. W. Kang, J. Lumin., 2009, 129, 1099–1104 CrossRef CAS .
  63. J. W. P. Hsu, D. R. Tallant, R. L. Simpson, N. A. Missert and R. G. Copeland, Appl. Phys. Lett., 2006, 88, 252103 CrossRef .
  64. Y. Zheng, L. Zheng, Y. Zhan, X. Lin, Q. Zheng and K. Wei, Inorg. Chem., 2007, 46, 6980–6986 CrossRef CAS PubMed .
  65. J. Won Ha, T. Purnima, A. Ruberu, R. Han, B. Dong, J. Vela and N. Fang, J. Am. Chem. Soc., 2014, 136, 1398–1408 CrossRef PubMed .
  66. C. Clavero, Nat. Photonics, 2014, 8, 95–103 CrossRef CAS .
  67. K. Wu, W. E. Rodríguez-Córdoba, Y. Yang and T. Lian, Nano Lett., 2013, 13, 5255–5263 CrossRef CAS PubMed .
  68. Z. Chen, L. Fang, W. Dong, F. Zheng, M. Shena and J. Wang, J. Mater. Chem. A, 2014, 2, 824–832 CAS .
  69. S. C. Warren and E. Thimsen, Energy Environ. Sci., 2012, 5, 5133–5146 CAS .

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6cp00375c
This journal is © the Owner Societies 2016