Thermal relaxation in combination with fiberglass confined interpenetrating networks: a key calcination process for as-desired free standing metal oxide nanofibrous membranes

Abstract

With unprecedented solar light active ZnWO₄/mixed-phased TiO₂ nanofibers as a model nanostructure, we showed that by simply allowing as-spun organic–inorganic hybrid samples to undergo thermal relaxation in combination with fiberglass confined calcination, as-desired and free-standing metal oxide nanofibrous membranes could be achieved for the first time. Under an electron microscope, the resulting metal oxide nanofibers were revealed to be freely interwoven without fixed crosslinks between the two contacting points and the fiber’s contour between any two slightly touching points appeared either curved or moderately digressed away from their original direction. The absence of crosslinking was attributed to the porous surface of the fiberglass fabric which offered a certain degree of spatial freedom to the membranes during calcination, leading to enhanced stress relaxation within the fiber during thermal annealing above the materials’ T_g. Such unique fibrous conformations promoted stress transfer along the fibers’ contours and prevented any possible stress concentration, enabling the calcined sample to withstand certain degrees of mechanical stress and so become less likely to fracture. We further demonstrated that the specific combination of metal precursors led to a photocatalytic platform for solar light induced deposition of a single layer of fully dispersed and highly catalytic noble metal nanoparticles.

---

1. Introduction

In various circumstances, an effective material must have several supporting characteristics: (i) multiple components, (ii) thermal stability, (iii) a highly accessible active site or surface area, (iv) mechanical stability and (v) flexibility. For metal oxides, the first two are easily and inherently achievable as ionic combinations among metal ions and oxides are infinite. Recent advances in the nanofabrication of films, particles and fibers have profusely led to high surface area metal oxides that fulfil the third requirement.^1–3 In bulk, the forth feature can easily be achieved especially when the metal components come from a family of transition metals as evidenced by ZrO.⁴ As brittleness is one of their inherent characteristics, metal oxides in various forms suffer greatly from a lack of flexibility which inevitably limits their usability against mechanical failure, especially when in nanostructured forms.^5,6

In recent years, several metal oxide nanostructures have been synthesized in several forms such as nanospheres, nanorods, nanofibers or nanofilms.^2,7–9 The nanofibrous form stands out in terms of its uncompromising benefits of chemical versatility and highly accessible surface area due to its zero degree of agglomeration unlike its nanosphere counterpart.¹⁰ However, a limited amount of research has been conducted into making these nanofibers into more viable structures to push the potential of this technology beyond laboratory studies towards practicality. The main reason is because metal oxide nanofibers are so fragile that their assembly into nanofibrous membranes can be quite a challenge. Specifically, advances towards thermally and mechanically stable stand-alone metal nanofibrous membranes have been limited by both a lack of understanding of the root causes of fragility and the small number of relevant investigations.

Capable of producing long and continuous fibrous constructs, electrospinning¹¹ is a potential method for the fabrication of stand-alone nanofibrous membranes. For metal oxides, high temperature treatment and crystallization are two interrelated processes which govern the final physical structure, especially crystallinity. Nevertheless, the nanofibers constructed often suffer drastic physical transformations, extreme fragility and break down into short fibers immediately after calcination. The same is true for several high activity and low environmental and health impact photocatalysts derived from metal oxides such as tungsten oxide (WO₃),¹² titanium dioxide (TiO₂)¹³ and zinc oxide (ZnO).¹⁴ Specifically, active towards visible, UV and solar light, these materials show benefits for applications in sensing, catalysis and electrochromics.^14–17

While the broad compositional possibilities among metal oxides have been considered endless, the incorporation of metals could be a prerequisite for several advancements.^3,18,19 In particular, metal/metal oxide hybrid materials have emerged as potential hybrid materials for a broad range of applications, especially in environmental and energy applications.^8,20–23

Among the greatest global challenges, environmental issues are the subject that can be dealt with effectively by photocatalysis. Clean air and water have become both societal and economic challenges affecting the wellbeing of people from all walks of life. Considered as state-of-the-art advances for water remediation, filtration,^11,24 absorption^25,26 and sedimentation^27–29 are applicable for colloids and solid suspensions. However, their limitation leaves quite a large performance gap for chemical and biological contaminants, the two major classes of waste water from industrial production plants, hospitals and scientific laboratories, etc. These two types of injurious pollutants require specially designed catalyts.³⁰

In this present contribution, with Pd/Pt-decorated solar light active ZnWO₄/mixed-phased TiO₂ nanofibers as a model subject, we investigate three aspects as prerequisites for the construction of mechanically stable metal oxide nanofibrous membranes. These involve (i) the processing of organic–inorganic nanofiber fabrications by electrospinning, (ii) the thermal transformation of the hybrid nanofibers and most importantly (iii) probing the physical structures of the resulting all-inorganic nanofibrous membranes. During our investigations, we analyse the point at which the presence of a single-crystal induced void manifests into a determining structural factor that affects the nanofibers’ physical stability. In addition, we study the correlation between the spatial distribution among the resulting all-inorganic nanofibers and the fragility at the membrane level. Following the stability improvements, we further demonstrate that the unique combination of nanofiber metal oxides leads to a photocatalytic platform for further light-induced noble metal nanoparticle deposition. Interestingly, the deposited metal forms a single layer of fully dispersed and supported metal nanoparticles making the hybrids advantageous for catalysis and sensing for energy and environmental applications that particularly require comprehensive accessibility to the catalytic active sites.

2. Experimental section

2.1. Materials

Ammonium metatungstate hydrate (AMT, (NH₄)₆H₂W₁₂O₄₀·xH₂O, ≥99.0%, Sigma Aldrich), zinc acetate dihydrate (ZAH, C₄H₆O₄Zn·2H₂O, ≥98.0%, Sigma Aldrich), titanium dioxide nanoparticles (P25, Degussa), titanium(IV) isopropoxide solution (TIP, C₁₂H₂₈O₄Ti, ≥97.0%, Sigma Aldrich), ethanol (99.8%, Sigma Aldrich), N,N-dimethylformamide (DMF ≥ 99.0%, Sigma Aldrich), polyvinylpyrrolidone (PVP, M_w ∼ 1 300 000, Fluka), palladium(II) nitrate hydrate (Pd(NO₃)₂·xH₂O, Sigma Aldrich), chloroplatinic acid hexahydrate (H₂PtCl₆·6H₂O, Sigma Aldrich) and glacial acetic acid (Sigma Aldrich), were of analytical grade and used as received.

2.2. Instrumentations

The morphologies of the nanofibers were characterized by scanning electron microscopy (SEM, S3400N, 20 kV, working distance 5–7 mm; SEM EDX, 20 kV, working distance 10 mm) and high resolution transmission electron microscopy (HRTEM, JEOL JEM-2010).

The crystal structures of the metal oxide nanofibers were determined from X-ray diffraction (XRD) patterns in a range of 2θ = 15–80° using an X-ray diffractometer (Bruker, D8 advance) with a Cu source (λ = 1.542 Å).

Quantification of gas elimination was performed using gas chromatography/mass spectrometry (GC/MS 5975, Agilent Capillary Column DB-624 60 m, 0.25 mm 1.40 S/N, USA. Network headspace sampler (G 1888)).

2.3. Preparation of electrospinning precursor solutions

2.3.1. Preparation of AMT electrospinning precursor solution. In a beaker, PVP (1 g) was added into ethanol (10 ml) under magnetic stirring for 20 minutes. In a separate beaker, AMT (0.2 g) was added into DI water (2 ml) under magnetic stirring for 10 minutes. Finally, the two solutions were mixed under magnetic stirring for 10 minutes prior to electrospinning.

2.3.2. Preparation of AMT and ZAH electrospinning precursor solution. The double component metal precursor solution was prepared by adding PVP (1 g) into a beaker containing ethanol (10 ml) under magnetic stirring for 20 minutes. In the second beaker, AMT (0.2 g) was added into DI water (2 ml) under magnetic stirring for 10 minutes. In the third beaker, ZAH (0.2 g) was added into DI water (2 ml) under magnetic stirring for 10 minutes. Finally, all three solutions were mixed together under magnetic stirring for 10 minutes prior to electrospinning.

2.3.3. Preparation of AMT, ZAH and P25 nanoparticle electrospinning precursor solution. The first solution with 10% wt/v of PVP in ethanol, the second solution with 10% wt/v of AMT in DI water and the third solution with 10% wt/v of ZAH in DI water were prepared in three separate beakers. In another beaker, P25 nanoparticles (0.2 g) were added in ethanol (10 ml) under magnetic stirring for 10 minutes. Finally, all four solutions were mixed together under magnetic stirring for 10 minutes.

2.3.4. Preparation of AMT, ZAH and TIP electrospinning precursor solution. In the first beaker, PVP (1 g) was added into ethanol (10 ml) under magnetic stirring for 20 minutes. In the second beaker, AMT (0.2 g) was added into DMF (2 ml) and stirred magnetically for 20 minutes. In the third beaker, ZAH (0.2 g) was added into DMF (2 ml) and stirred magnetically for 20 minutes. Subsequently, the three prepared solutions were mixed under magnetic stirring for 20 minutes. Finally, TIP (2 ml) was added into the above sol–gel mixture followed by acetic acid (2 ml) to adjust the pH to 3–4.

2.4. Fabrication of nanofibers and nanofibrous membranes using a Nanospider machine

Nanospider technology was employed for a needle-free electrospinning process, producing nanofibers with high quality and throughput. For pre-pilot scale nanofiber fabrication by the Nanospider machine (NS LAB 500), the prepared sol–gel precursor solution (30 ml) was transferred into a 50 ml semi-cylindrical chamber. The distance between the rotating wire electrode and the ground electrode was set at 18 cm. During the fabrication process, the wire electrode was rotated at 8 rpm with an applied voltage of 50 kV. The as-spun nanofibrous membrane was calcined in a furnace (Nabertherm GmbH, temperature range: 30–3000 °C, model: LT 15/12/P320) under designated conditions to remove organic contents from the metal oxide nanofibers.

2.5. Thermal annealing and high temperature calcinations

An electrospun nanofibrous membrane was cut into a 3 cm × 3 cm piece. Then, the prepared membrane was placed inside a furnace on a support or in between sandwiches of glass microfibrous mats (Grade 934-AH, diameter 70 mm, Sigma Aldrich) or glass slides (25.4 × 76.2 mm, 1–1.2 mm thickness, Sigma Aldrich) depending on the designated conditions. For annealing and calcination, the treatment was set under air for an hour at the designated temperature (100 or 200 °C) before calcination also under air for 4 hours at 500, 600 or 700 °C. After the thermal treatment, the length, width and thickness of the nanofibrous membrane were measured using a ruler and a micrometer, respectively.

2.6. Noble metal photodeposition on metal oxide nanofibrous membranes

Three types of light, visible, UV and natural sunlight, were employed as activating energy sources for photodeposition of metals on the metal oxide nanofibers. For visible and UV light photodeposition processes, the nanofibrous membrane (50 mg) was added into DI water (60 ml). Then, palladium(II) nitrate hydrate (1 mg) and chloroplatinic acid hexahydrate (3 mg) were added into the above suspension and magnetically stirred for an hour under visible light (60 W light bulb) or UV light (365 nm) irradiation from the light source set 10 cm apart from the top level of the solution. Finally, the resulting yellowish suspension was filtered, washed and dried at 60 °C for an hour.

For the photodeposition under solar light, a similar suspension was prepared. The nanofibrous membrane (50 mg) was suspended in DI water (60 ml). Then, palladium(II) nitrate hydrate (1 mg) and chloroplatinic acid hexahydrate (3 mg) were added into the solution and magnetically stirred for an hour under natural sunlight. Finally, the resulting yellowish suspension was filtered, washed and dried at 60 °C for an hour.

2.7. Photocatalyst recyclability against methylene blue degradation under visible light

Recyclability of the photocatalytic nanofibers against MB degradation was performed under visible light irradiation (60 W light bulb) in a dark chamber. In each photocatalytic test with MB, the nanofibers (50 mg) were added into a 5 ppm MB solution (300 ml). Then, the suspension was placed under the light bulb with a fixed distance between the light bulb and the suspension of 10 cm. After every given interval of light irradiation, a 1 ml aliquot of the sample was collected, centrifuged (10 000 rpm for 10 minutes) and filtered to remove the photocatalyst. Subsequently, the concentration of MB was measured using a UV-Vis spectrophotometer (Perkin Elmer, Lambda 650, USA) at 663 nm, the maximum absorption of MB. After the first catalytic cycle, the remaining suspension was centrifuged at 10 000 rpm for 10 minutes. Then, the MB solution was separated from the catalyst which was subsequently rinsed and dried at room temperature for an hour. Catalytic reactions using the MB solution were performed until five cycles were complete.

2.8. Methylene blue degradation under sunlight

Photocatalytic degradation of MB was performed under natural sunlight (sunlight intensity was measured by Lux meter: Heavy duty light meter, Extech instrument, model: 407026). Firstly, the selected nanofibers (50 mg) were added into a 5 ppm MB solution (300 ml). Subsequently, the resulting suspension was placed under direct sunlight. After every given interval of light irradiation, a 1 ml aliquot of the sample was collected, centrifuged and filtered to remove the photocatalyst. The concentration of MB was measured using a UV-Vis spectrophotometer at 663 nm, the maximum absorption of MB.

2.9. Gaseous benzene and methanol decomposition under visible light photocatalysis

Photocatalytic decomposition of gaseous benzene and methanol was performed under visible light (60 W bulb) by a head-space technique. First of all, a selected nanofibrous sample was added into a vial containing 500 ppm of benzene/methanol in the vapor phase. Subsequently, the vial was sealed using aluminum foil and allowed to equilibrate at room temperature for 2 hours. The vial was then exposed to visible light for 4 hours. The concentrations of benzene/methanol were measured by a GC/MS headspace method (gas chromatography-mass spectrometer, GC/MS 5975, Agilent Capillary Column DB-624 60 m, 0.25 mm 1.40 S/N, USA, Network headspace sampler (G 1888)).

3. Results and discussion

Our ultimate goal was to fabricate thermally and mechanically stable metal oxide nanofibrous membranes for further metal decoration for highly effective chemical catalysis. As such, the investigation was divided into four consecutive parts: (i) formation of stable, multicomponent and broad spectrum photocatalytic nanofibers, (ii) fabrication of stable and flexible metal oxide nanofibrous membranes, (iii) fabrication of metal-decorated metal oxide nanofibrous membranes and (iv) evaluation of the materials’ efficiency.

3.1. Fabrication and characterization of metal oxide nanofibers

The starting multicomponent precursor solutions were designed to contain one of three metal oxide precursors—tungsten oxide, titanium dioxide and zinc oxide. It was evident that the first two had prominent visible and UV light activities while the latter was expected to prevent extensive single crystal formation in the domains of the first two. The consequential desire for mixtures of the three metal precursors in this study required the careful selection of solvent due to the different solubility of each metal precursor. Experimentally, the investigation of the metal precursor solubility in environmentally benign solvents was the first trial. As such, compatibility of one or two of the metal precursors in the selected solvents was observed. Firstly, the AMT precursor solution was homogeneous and colorless after mixing with PVP. Nanofiber fabrication using the Nanospider machine resulted in stable and bead-free nanofibers (Fig. S1a, ESI†). Secondly, the nanofibers fabricated from a precursor solution containing AMT and ZAH showed no difference from their one-component (AMT) counterpart described above (Fig. S1b, ESI†). Following this promising result, a multicomponent metal precursor solution containing AMT, ZAH and P25 nanoparticles was prepared in the next step. Unfortunately, the addition of P25 led to a cloudy solution with clearly discernable white precipitates. Furthermore, subsequent electrospinning resulted in nanofibers with heavily populated beads due to agglomeration of the nanoparticles in water and ethanol (Fig. S1c, ESI†). Consequently, thermal treatment of these triple metal precursor component nanofibers at 500 °C led to a catastrophic outcome with nanofibers of irregular shape and with rough surfaces (Fig. S1d, ESI†). The results were consistent with a few previous findings where adding nanoparticles to an electrospinning precursor solution would often lead to physical irregularity in the final metal oxide nanofibers due partly to the agglomeration of the nanomaterials.³¹

In the following experiment, P25 was replaced by TIP as an alternative precursor for the UV light active component. However, this resulted in an opaque solution. Such an observation could be explained by the condensation of TIP which was caused by the surrounding moisture favoring formation of minute solid particles suspended throughout the solution. However, subsequent nanofiber fabrication using the Nanospider machine resulted in large diameter and non-homogeneous fibers (Fig. S1e, ESI†). After calcination at 500 °C, clusters of particles appeared in the absence of fiber formation as one would expect (Fig. S1f, ESI†).

In order to avoid the self-condensation of TIP in such moisture containing environments, organic solvents became the media of interest. However, complete elimination of water would result in a suspension of ZAH due to its insolubility in the remaining ethanol. As a consequence, DMF was introduced to replace the water content of the solution mixtures. After mixing ZAH with the rest of the precursors, the resulting solution appeared clear, homogeneous and yellowish with no trace of solid particulates from TIP condensates as observed previously. The following electrospun nanofibers were homogeneous in shape and size (Fig. S2a, ESI†). After calcination at 500 °C, well-defined nanofibers were obtained with a slight reduction in the fibers’ diameter due to the elimination of the original carbon-containing constituents (Fig. S2b, ESI†).

Typically, the degree of heat treatment strongly influenced the crystal structure and final properties of the metal oxides. In order to investigate the relevance of this influence, three variations of calcination temperatures were performed at 500, 600, and 700 °C. Interesting morphological changes on the nanofibers’ surface were observed after calcination at 600 °C (Fig. 1a), with rod-like nanostructures stemming from the surface along the nanofibers. Similar surface distortions were also discerned for a sample treated at 700 °C (Fig. S2d, ESI†). Under SEM-EDX, all characteristic elemental footprints, Zn, W, and Ti, were found in the absence of carbon content (Fig. S3, ESI†).

Fig. 1 (a) SEM image of nanofibers after calcination at 600 °C. (b) HRTEM of (a). (c–e) HRTEM of selected areas in (b). (f) Nanorod on the nanofiber. (g) HRTEM of (f).

The effect of thermal treatment was further rationalized at the crystal structure level. X-ray diffraction (XRD) patterns of the metal oxide nanofibers from three varying calcination temperatures are illustrated in Fig. S4 (ESI†). In the absence of zinc oxide or tungsten oxide diffraction footprints, only a TiO₂ anatase crystalline structure formed in the nanofibers after calcination at 500 °C. At the higher calcination temperature (600 °C), both characteristic crystalline structures of TiO₂, anatase and rutile, were distinguished as major phases, with ZnWO₄ in a sanmatinite crystalline structure³² as a minor phase. At the highest calcination temperature (700 °C), clear characteristic peaks of sanmatinite phase ZnWO₄ were still discernable while the rutile phase of TiO₂ dominated over the anatase phase. From FWHM (full width at half maximum), the average anatase crystalline size of TiO₂ after calcination at 500, 600, and 700 °C was 11.5, 13.7, and 40.4 nanometers, respectively. The increase in crystal size was hypothesized to be due to the higher energy given to the samples at higher temperatures to have their bulk reach structural equilibrium.

The morphology of the metal oxide nanofibers was further demonstrated by TEM. Under a scanning and transmission mode, all three crystalline domains of TiO₂ and ZnWO₄ were observed using HRTEM (Fig. 1b) and SAED measurements (Fig. S5, ESI†). Fig. 1c illustrates the diffraction spacing of the (011) planes of ZnWO₄ while Fig. 1d and e shows the diffraction spacing of the (004) planes of anatase and the (101) planes of rutile crystal structures of TiO₂, respectively. In addition, TEM-EDX confirmed the presence of the three key elements, Ti, Zn and W, as expected from such a triple metal composite (Fig. S5f, ESI†).

The uniquely present rod-like structures were also scrutinized using HRTEM. It was primarily observed that the minute structure embodied a well-defined rod shape with a persistent diameter of 60 nanometers, stemming almost vertically from the fiber’s surface (Fig. 1f). To elaborate on the crystallographic structure of the nanorod, SAED measurements (Fig. 1f inset) and HRTEM (Fig. 1g) helped identify the interplanar spacings of the lattice fringes. The result suggested that these nanorods conveyed a diffraction spacing that belonged to the (121) planes of ZnWO₄, confirming the existence of ZnWO₄/mixed-phased TiO₂ nanofibers as an output. Interestingly, under the TEM-EDX mode, it was unveiled that titanium took no part in the chemical composition of those nanorods (Fig. S6f, ESI†). However, to the best of our knowledge, this was the first time that nanorods could be grown from the surface of metal oxide nanofibers via a single step electrospinning process and subsequent thermal treatment. From the above crystal developments, it was hypothesized that phase separations among the emerging stable and distinctive single crystals could be a driving force behind the formation of these nanorods at the surface. Consequently, in-depth characterization of ZnWO₄/mixed-phased TiO₂ nanofibers was further carried out using XANES (X-ray absorption near edge structure).^33–35 In the vicinity of the Zn K-edge (9659 eV), the measurement was completed under fluorescence mode by using a Ge (220) monochromator at BL-5 of the Synchrotron Light Research Institute (SLRI). Fig. 2a shows the XANES spectra of the electrospun nanofibers before and after calcination at three various temperatures compared with Zn foil and ZnO references. Typically, the edge energy positions shifted to a higher level when the oxidation state of the metal ions increased. It could be clearly seen that all the edge energies of all the ZnWO₄/mixed-phased TiO₂ samples (Fig. 2a(i)) were located close to that of the ZnO reference, insinuating the existence of Zn²⁺. In addition, the adsorption intensity or white line position (Fig. 2a(ii)) of all the samples gradually increased in unison with the increase of the calcination temperature. This phenomenon is strongly associated with the growth of ZnWO₄ crystal size at higher calcination temperatures. Then, it could be said that the change in white line intensity might be influenced by the increase of the crystal sizes. The results above suggested that the existence of Zn²⁺ persisted during the physical transformation from solution to electrospun nanofibers to metal oxide nanofibers. Therefore, it could be inferred from these results that metal oxides had emerged in the initial state. As the same types of the metal oxide persisted during thermal evolution, phase separations among the emerging stable and distinctive single crystals became a dominant mechanism behind the formation of the nanorods.

Fig. 2 (a) Zn K-edge XANES spectra of ZnWO₄/mixed-phased TiO₂ nanofibers at different calcination temperatures. (b, c) TEM and HRTEM images of WO₃ nanofibers after calcination at 600 °C. (c) ZnWO₄/mixed-phased TiO₂ nanofibers after calcination at 600 °C.

Furthermore, the unusual dimensional stability of the resultant ZnWO₄/mixed-phased TiO₂ nanofibers inspired an in depth investigation of the nanofibers’ structure to determine the physical origin of their physical integrity. In the following structural analysis, brittle WO₃ nanofibers were employed as references. Their physical appearance at the nanometer length scale was closely inspected and compared with that of the ZnWO₄/mixed-phased TiO₂ nanofibers. Under TEM transmission mode, the WO₃ nanostructures held a low degree of structural homogeneity, as fully developed single crystals were predominant (Fig. 2b). A stark contrast was observed for the well-defined nanofibrous structure of the ZnWO₄/mixed-phased TiO₂ composite which showed densely packed bulk areas along its secure arrangement (Fig. 2c). This could be confirmed using HRTEM of both structures as the WO₃ nanofibers are unmistakably composed of loosely packed single crystal building blocks. Such rough nanodomains were completely absent in the firm ZnWO₄/mixed-phased TiO₂ nanofibers.

3.2. Development of stable and flexible metal oxide nanofibrous membranes

In the subsequent investigations, stability aspects were considered systematically during (i) the annealing process, (ii) the calcination process and when considering (iii) the characteristics of the membrane fabrication method. It was first hypothesized that the final structures of the metal oxide nanofibrous membranes could be strongly influenced by the thermal treatment process. Experimentally, the structural development towards metal oxide nanofibers during the thermal treatment process was monitored as a function of heating methods. First, the dimensions of the electrospun membrane samples were examined before the treatment (Fig. 3a). In the first approach, after calcination at 600 °C (Fig. 3b), the membrane contracted and crumbled with fragments observed near the original membrane. In the second approach, an original electrospun membrane was placed inside the furnace for annealing at 100 °C followed by calcination at 600 °C (100-AC processes). The resulting calcined membrane showed a smaller degree of shrinkage in comparison to the previous example. However, while contracted, the calcined membrane held its original shape with no discernable fragments (Fig. 3c). For the sample annealed at 200 °C before calcination at 600 °C (200-AC processes), a similar degree of shrinkage was observed (Fig. 3d) compared to the original electrospun nanofibers. The better stability of the latter sample could be an effect of the given relaxation time for the polymer during annealing at temperatures above its glass transition. From the above results, it could be inferred that thermally induced relaxation might also help stabilize the nanofibrous metal oxide membrane assembly during high temperature transformations.

Fig. 3 Pictures of electrospun nanofibrous membranes after (a) calcination at 600 °C, (b) 100-AC processes, (c) 200-AC processes, (d) 100-AC processes while sandwiched between two pieces of fiberglass fabric, (e) 200-AC processes while sandwiched between two glass slides, (f) 100-AC processes and (g) 200-AC processes while sandwiched between two pieces of fiberglass fabric.

Nevertheless, the resultant metal oxide nanofibrous membranes showed a discernable degree of physical surface irregularities that could impede future high quality device fabrications. Thus, we investigated the factors pertinent to structural control towards well-defined metal oxide nanofibrous membranes. First, it was argued that the spatially unconfined electrospun sample thermally and freely transformed into nanofibrous metal oxide membranes with random dimensional evolution. It was then hypothesized that physical confinement could counteract the undesirable physical distortion.

In a subsequent experiment, three identical and physically well-defined electrospun nanofibrous membranes were subjected to high temperature treatment under different degrees of confinement. The first sample was annealed and calcined while sandwiched between two glass slides. Sandwiched between two pieces of fiberglass fabric, the second sample was annealed at 100 °C and calcined at 600 °C, while the third was annealed at 200 °C and calcined at 600 °C.

Under high temperature treatment between glass slides, the calcined nanofibrous membrane retained its flat structure with a smooth surface (Fig. 3e). However, the sample became so fragile and unstable that picking it up using tweezers was found to be impossible as it would readily dissemble into small pieces. In contrast, the other two samples that were thermally treated between two pieces of fiberglass fabric at 100 °C/600 °C (Fig. 3f) and 200 °C/600 °C (Fig. 3g) appeared physically stable with relatively rougher surfaces. It was interesting to note that, for these two samples, thermal treatment caused a 70% reduction in the membranes’ area (Fig. S7 and S8, ESI†). This dimensional shrinkage could be informative during the future planning of such membrane production. Therefore, it could be said that the confinement method during thermal treatment played an influential role on the stability of the final metal oxide nanofibrous membranes.

It was then hypothesized that the form of confinement during the thermal transformation imparted an effect on the structure of the membrane. Consequently, more intriguing characteristics of the thermally treated sandwiched samples were discovered after careful inspection at the nanometer scale as shown in Fig. 4a–d. While the two samples appeared almost identical in terms of the fibers’ characteristic size, shape and degree of surface homogeneity, there was one distinct difference. That was the way the metal oxide nanofibers in the two samples intermingled differently.

Fig. 4 (a and b) SEM images of calcined nanofibers while sandwiched between glass slides and (inset of a) TEM image showing the merging point. (c and d) SEM images of calcined nanofibers while sandwiched between two pieces of fiberglass fabric and (inset of c) TEM image showing no merging point. Pictures of electrospun nanofibrous membranes after calcination by (e) 200-AC processes with controlled bending into shape by fiberglass fabric in a beaker and (f) 200-AC processes with controlled folding into shape by fiberglass fabric.

For samples from the glass slide sandwiches, the nanofibers were interwoven with physical crosslinks between any two contacting points. It was also observed that the fiber’s contour between these physical crosslinks appeared straight (Fig. 4a and b). It was believed that these crosslinks might have been forced to form because of the rigid and flat surface of the glass slide that introduced absolute spatial confinement to the membranes during and after the calcination. As a consequence, it could be said that such confinement led to reduced stress relaxation in the fiber during annealing at temperatures above the materials’ T_g, forcing the fibers to make contact with each other and finally coalesce. Furthermore, physical crosslinking within the resulting metal oxide nanofibers might have prevented stress transfer along the fibers’ contour that led to stress concentration, making the sample prone to breakage with only slight mechanical perturbation.

Different nanostructures were identified for the metal oxide samples from the fiberglass fabric sandwich. Evidently, the nanofibers were freely interwoven without fixed crosslinks between two contacting points (Fig. 4c and d). Furthermore, the fiber’s contour between any two slightly touching points appeared either curved or moderately digressed away from their original direction. It was hypothesized that the absence of crosslinking could be attributed to the porous surface of the fiberglass fabric which offered a certain degree of spatial freedom to the membranes during the calcination. Arguably, such spatial freedom led to enhanced stress relaxation in the fiber during thermal annealing above the materials’ T_g. Moreover, the looser structure promoted stress transfer along the fibers’ contour and prevented stress concentration, enabling the calcined sample to withstand certain degrees of mechanical stress and so become less likely to fracture.

From the above structure–property understanding, more stable metal oxide nanofibrous membranes could be made into different shapes. For example, we attempted to study the structure of metal oxide nanofibrous membranes under fiberglass sandwich confinement that had been manipulated into various 3D structures. It was noted that such nonconventional conformations might open up new possibilities in future smart device fabrication. The nanofibrous membrane was first sandwiched between two pieces of fiberglass fabric and then put in a beaker prior to high temperature treatment (Fig. S7m, ESI†). It was found that the resultant shape-controlled membrane was not only stable but also slightly flexible in the original bent shape (Fig. 4e). Even more complex metal oxide nanofibrous membrane structures could be successfully created by the same thermal processes but with different supporting sandwich architectures. Fig. S7o (ESI†) showed a membrane before calcination sandwiched between two wavy-shaped pieces of fiberglass fabric. After the calcination, the sample was not only able to maintain its nanofibrous membrane structure but also retained its original wavy shape with high stability and slight flexibility (Fig. 4f). To the best of our knowledge, these observations of thermally and mechanically free-standing metal oxide nanofibrous membranes have never been reported before. It could be said that this novel and yet simple fabrication of stable and flexible 3D metal oxide nanofibrous membranes was a promising step towards overcoming the most crucial drawback of metal oxide membranes, i.e. fragility, which has impeded the scaling up process towards an industrial scale. (The SEM images of nanofibers from each AC process are provided in Fig. S8, ESI†). Various applications such as nanofiltration and flow-through catalytic systems can take advantage of these as-designed and stand-alone membrane designs to fulfil demands previously unmet by their powder counterparts.

3.3. Pd/Pt-decorated solar light active ZnWO₄/mixed-phased TiO₂ nanofibers

It was part of our initial plan to utilize our metal oxide, i.e. ZnWO₄/mixed-phased TiO₂ nanofibers, as a versatile platform for metal decoration. However, the challenge was how to develop a process to deposit well-defined functional metals onto the given metal oxide nanofibrous membrane. In this specific study, incorporating nanostructured metals on the nanofibers’ surface could be an excellent choice for the fabrication of hybrid metal/metal oxide catalysts with even more functionalities and broader utilization. For instance, it has been found that metal nanoparticles could act as electron-trapping sites that impede electron–hole recombination in photocatalysis.¹⁷ The successfully fabricated ZnWO₄/mixed-phased TiO₂ nanofibrous membranes were hypothesized to be photocatalyst active under visible, UV and solar light. Furthermore, palladium and platinum were employed as model noble metals in the material design.

Experimentally, photodeposition of Pd and Pt onto the metal oxide nanofibrous membranes was performed under visible, UV and natural sunlight. The resulting metal/metal oxide hybrid materials after photodeposition of the two metals under laboratory generated visible and UV light irradiation were characterized using SEM-EDX, where the presence of the key elements, Pd and Pt, was confirmed (Fig. S9, ESI†). From SEM, it was ascertained that the nanofibrous membrane structures remained intact after the photodeposition reactions (Fig. S10, ESI†). Careful inspection unveiled small dots decorating the surface of the nanofibers. HRTEM further confirmed that these were metal nanoparticles that were also deposited on the ZnWO₄ nanorods by photodeposition under visible and UV light irradiation (Fig. S10, ESI†).

After successful photodeposition under laboratory generated visible and UV light, a similar reaction was performed under natural sunlight of 85 300 Lux in intensity (Fig. S11, ESI†). As a consequence, the resulting hybrid materials retained their nanofibrous structure with small rods adhering to the surface (Fig. 5a). Using SEM-EDX, it was found that all three of the above samples contained elemental footprints for both Pd and Pt, confirming successful deposition of the two noble metals (Fig. 5b).

Fig. 5 (a) SEM and (b) SEM-EDX of photodeposited Pd and Pt nanoparticles on ZnWO₄/mixed-phased TiO₂ nanofibers. (c–e) HRTEM images and SAED patterns of Pd and (f–h) Pt nanoparticles on ZnWO₄/mixed-phased TiO₂ nanofibers after the sunlight photodeposition process.

The TEM and HRTEM images, and SAED patterns (Fig. 5c–h) revealed nanosized particles with an average diameter of less than 7 nm attached to the surface of the main fiber and the stemming ZnWO₄ nanorods. The interplanar spacing of the round-shaped particles on the surface of the ZnWO₄ nanorod was found to be 0.223 and 0.227 nm, corresponding to the (111) crystal planes of Pd and Pt, respectively.

3.4. Catalytic applications

The photocatalytic performance of the metal oxide nanofibers was evaluated using three methods, including the in-house photocatalyst recyclability under visible light, the outdoor photocatalytic activity under sunlight and the gas-phase decomposition of a benzene/methanol mixture. In the first approach, the ZnWO₄/mixed-phased TiO₂ nanofibers were tested for the degradation of MB under visible light for 5 cycles. Fig. 6 illustrates the catalyst’s performance against the number of catalytic cycles. It shows the varying concentration of MB as a function of the reaction time after 0 (black), 1 (red), 3 (blue) and 6 hours (pink) of visible light irradiation over 5 cycles. The first set of columns showed the average decrease of the reference MB concentration in the absence of catalyst which did not show any significant reduction throughout the experiment. In addition, the next five sets of columns showed the photocatalytic performance during the subsequent five repeats. As was evident, the performances during the first hours were consistent for all cycles where only a small degree of MB reduction was observed. However, significant decreases and moderate variance in the MB concentrations were observed after 3 hours for each cycle with C/C₀ values of 0.6012 in the second cycle and 0.7031 in the third cycle. For all cycles, additional reductions were observed after 6 hours of irradiation time. It could be seen from these results that the catalyst could be recycled.

Fig. 6 Recycling of ZnWO₄/mixed-phased TiO₂ nanofibers against photocatalytic degradation of 5 ppm MB solution under visible light. (Inset) Vials showing fading color of the methylene blue solutions after 6 hours of light irradiation; from left to right: control MB, 1st cycle MB, 2nd cycle MB, 3rd cycle MB, 4th cycle MB and 5th cycle MB.

The second photocatalytic performance evaluation involved a demonstrative application of water treatment where ZnWO₄/mixed-phased TiO₂ nanofibers and Pd/Pt-decorated ZnWO₄/mixed-phased TiO₂ nanofibers were examined through degradation of methylene blue (MB) in the aqueous phase with WO₃ nanofibers as a reference (Fig. 7). The experiment was performed on a sunny day with a clear sky where the sunlight intensity was measured to be close to 83 000 Lux. The three catalysts showed different degradation rates with sunlight exposure. All catalysts showed their highest rates at 15 minutes after starting the reaction, with WO₃ nanofibers showing a much lower performance throughout the 6 hours of irradiation time. It was clearly shown that both types of newly developed hybrid nanofibers performed far better than the reference materials, with Pd/Pt-decorated ZnWO₄/mixed-phased TiO₂ nanofibers showing the most impressive result.

Fig. 7 Photocatalytic degradation of MB under sunlight by WO₃, ZnWO₄/mixed-phased TiO₂, and Pd/Pt–ZnWO₄/mixed-phased TiO₂ nanofibers. (Inset) Experimental set up under natural sunlight.

The third evaluation approach represented a demonstrative application of the catalyst in air treatment via degradation of a benzene/methanol mixture in the gas phase. Due to the high risk associated with the model pollutants, the catalytic degradation was conducted under laboratory-generated visible light with WO₃ nanofibers as a reference. Visible light was selected over UV light because it represents 95% of natural solar light.

The three catalysts showed different degradation rates under visible light irradiation (Fig. 8). For air treatment of the two gases, it was found that WO₃ nanofibers did not show any activity. In contrast, ZnWO₄/mixed-phased TiO₂ nanofibers showed a 43% degradation efficiency against 500 ppm benzene. Even though Pd/Pt-decorated ZnWO₄/mixed-phased TiO₂ nanofibers showed a lower benzene degradation efficiency, they could transform methanol into methyl formate at room temperature.

Fig. 8 Benzene decomposition and methyl formate conversion at room temperature. Benzene/methanol mixture with (a) no catalyst as reference, (b) WO₃, (c) ZnWO₄/mixed-phased TiO₂ and (d) Pd/Pt–ZnWO₄/mixed-phased TiO₂. (Inset) Experimental set up under visible light.

The results from these model tests suggested that the metal oxide nanofibers could be effective as photocatalysts under visible light for stable air pollutants such as benzene. Moreover, they could also act as a platform for further metal decoration via a facile photodeposition method utilizing their photocatalytic features. The resulting metal/metal oxide hybrids could not only hold the original well-defined nanofibrous membrane, but also offer a versatile catalytic substrate for endless chemical reactions.

4. Conclusion

Using Pd/Pt-decorated solar light active ZnWO₄/mixed-phased TiO₂ nanofibers as a model subject, we showed that three crucial stability aspects must be realized towards the construction of mechanically stable metal oxide nanofibrous membranes. These investigations probed the stabilities of the organic–inorganic hybrid nanofibers’ spinnability, the thermal transformation of the hybrid nanofibers and the physical structures of the subsequently all-inorganic membranes formed. We observed two crucial phenomena which influenced the mechanical performances of the final nanofibrous membranes. First, the thermally-induced single crystals in our stable metal oxide nanofibers appeared densely packed, while those in a thermally treated and brittle WO₃ nanofiber, used as a reference, induced nanosized voids throughout its bulk. Second, in our undesirable and unstable nanofibrous samples, the metal oxide nanofibers were interwoven with distinct evidence of physical crosslinking, an observation unseen in their stable counterpart. We further demonstrated that specific combination of metal precursors led to not only uniquely stable multicomponent metal oxide nanofibers and membranes, but also a photocatalytic platform for solar light induced noble metal nanoparticle deposition. The resulting metal/metal oxide hybrids with a single layer of fully dispersed and supported metal nanoparticles could be of special interest in catalytic applications as demonstrated through both water and air treatments. Finally, both ZnWO₄/mixed-phased TiO₂ nanofibers and Pd/Pt-decorated ZnWO₄/mixed-phased TiO₂ nanofibers showed excellent photocatalyic degradation of model water- and airborne pollutants. Furthermore, the latter showed the ability to transform methanol into methyl formate, supporting the proof of concept that this hybrid material could be amenable towards versatile chemical reactivity.

Acknowledgements

The authors would like to thank the Thailand Research Fund (Grant Number RSA5780067) and National Nanotechnology Center for financial support.

References

1. D.-H. Lien, J. R. D. Retamal, J.-J. Ke, C.-F. Kang and J.-H. He, Nanoscale, 2015, 7(47), 19874–19884.
2. D. R. Miller, S. A. Akbar and P. A. Morris, Sens. Actuators, B, 2014, 204, 250–272.
3. R. Ramachandran, S.-M. Chen and G. P. G. Kumar, Int. J. Electrochem. Sci., 2015, 10(10), 8581–8606.
4. P. F. Manicone, P. R. Iommetti and L. Raffaelli, J. Dent., 2007, 35(11), 819–826.
5. H. K. Yu, S. Kim, B. Koo, G. H. Jung, B. Lee, J. Ham and J. L. Lee, Nanoscale, 2012, 4(21), 6831–6834.
6. R. Raghavan, J. Elias, R. Erni, M. Parlinska, L. Philippe and J. Michler, Thin Solid Films, 2015, 578, 174–179.
7. K. L. Garner and A. A. Keller, J. Nanopart. Res., 2014, 16, 2503.
8. G. Li and Z. Tang, Nanoscale, 2014, 6(8), 3995–4011.
9. W. Yaipimai, N. Subjalearndee, G. Tumcharern and V. Intasanta, J. Mater. Sci., 2015, 50(23), 7681–7697.
10. M. C. Daniel and D. Astruc, Chem. Rev., 2004, 104(1), 293–346.
11. F. E. Ahmed, B. S. Lalia and R. Hashaikeh, Desalination, 2015, 356, 15–30.
12. C. Srisitthiratkul, W. Yaipimai and V. Intasanta, Appl. Surf. Sci., 2012, 259, 349–355.
13. R. Fagan, D. E. McCormack, D. D. Dionysiou and S. C. Pillai, Mater. Sci. Semicond. Process., 2016, 42, 2–14.
14. M. Muruganandham, Y. F. Zhang, R. Suri, G. J. Lee, P. K. Chen, S. H. Hsieh, M. Sillanpaa and J. J. Wu, J. Nanosci. Nanotechnol., 2015, 15(9), 6900–6913.
15. D. Sudha and P. Sivakumar, Chem. Eng. Process., 2015, 97, 112–133.
16. V. Pongsorrarith, C. Srisitthiratkul, K. Laohhasurayotin and N. Intasanta, Mater. Lett., 2012, 67(1), 1–4.
17. C. Srisitthiratkul, V. Pongsorrarith and N. Intasanta, Appl. Surf. Sci., 2011, 257(21), 8850–8856.
18. C. Deraedt and D. Astruc, Acc. Chem. Res., 2014, 47(2), 494–503.
19. L. Li, L. Hu, J. Li and Z. Wei, Nano Res., 2015, 8(2), 418–440.
20. S. Leong, A. Razmjou, K. Wang, K. Hapgood, X. Zhang and H. Wang, J. Membr. Sci., 2014, 472, 167–184.
21. R. Molinari, C. Lavorato and P. Argurio, Chem. Eng. J., 2015, 274, 307–316.
22. M.-C. Wu, P.-H. Lee and D.-L. Lee, Int. J. Hydrogen Energy, 2015, 40(13), 4558–4566.
23. A. Zielinska-Jurek and J. Hupka, Catal. Today, 2014, 230, 181–187.
24. A. G. Fane, R. Wang and M. X. Hu, Angew. Chem., Int. Ed., 2015, 54(11), 3368–3386.
25. X. Liang, S. Liu, S. Wang, Y. Guo and S. Jiang, J. Chromatogr. A, 2014, 1357, 53–67.
26. A. P. Lim and A. Z. Aris, Rev. Environ. Sci. Bio/Technol., 2014, 13(2), 163–181.
27. H. Anlauf, Sep. Purif. Technol., 2007, 58(2), 242–246.
28. I. Bekri-Abbes, S. Bayoudh and M. Baklouti, Desalination, 2007, 204(1–3), 198–203.
29. N. Maximova and O. Dahl, Curr. Opin. Colloid Interface Sci., 2006, 11(4), 246–266.
30. A. Cincinelli, T. Martellini, E. Coppini, D. Fibbi and A. Katsoyiannis, J. Nanosci. Nanotechnol., 2015, 15(5), 3333–3347.
31. W. Yaipimai and V. Intasanta, Sci. Adv. Mater., 2014, 6(3), 448–458.
32. H. B. Fu, J. Lin, L. W. Zhang and Y. F. Zhu, Appl. Catal., A, 2006, 306, 58–67.
33. A. Kuzmin and J. Chaboy, IUCrJ, 2014, 1, 571–589.
34. T. K. Sham, Int. J. Nanotechnol., 2008, 5(9–12), 1194–1246.
35. G. Sigircik, O. Erken, T. Tuken, C. Gumus, O. M. Ozkendir and Y. Ufuktepe, Appl. Surf. Sci., 2015, 340, 1–8.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra15086a

---