A facile method to prepare zirconia electrospun fibers with different morphologies and their novel composites based on cyanate ester resin

Abstract

Using zirconium propoxide as the precursor, a new method combining the advantages of electrospinning and sol–gel approaches is set up to facilely fabricate zirconia (ZrO₂) nanofibers with controllable chemical and morphological structures in batch size on a standard electrospinning equipment. Interestingly, zirconia (ZrO₂) nanofibers with different morphologies were prepared, the influence of the preparing parameters on the structure of the ZrO₂ nanofibers were investigated. Results show that by adjusting the preparing parameter, the ZrO₂ nanofibers can be porous or compact fibers, in addition the crystalline structure, and dimensions of pores can be also changed. Based on the successful preparation of these ZrO₂ electrospun fibers, novel composites consisting of ZrO₂ electrospun fibers and cyanate ester (CE) resin were developed, which show significantly reduced curing temperature compared to CE owing to the presence of hydroxyl groups on the fibers. The dynamic mechanical and dielectric properties of ZrO₂/CE composites are closely related to the morphological structure of ZrO₂ fibers because the latter determines the chemical structure and crosslinking density of the matrix as well as the interfacial adhesion of the corresponding composites. These interesting results demonstrate that the method proposed herein provide a new approach to design and prepare ceramic nanofibers and corresponding composites with controlled structure and expected performance for cutting-edge industries.

---

1 Introduction

Traditionally, materials are divided into three kinds, of which polymer and ceramic belongs to two sorts, each of them has unique properties, and thus shows special applications.^1,2 However, with the rapid development of modern industries, an individual polymer or ceramic is found to be difficult to meet the requirements of cutting-edge fields, hence ceramic/polymer composites have been received increasing attentions because they not only combine the advantages of both polymer and ceramic, and may have new unique properties which are not possessed by either polymers or ceramics.^3,4

There are two methods to prepare ceramic/polymer composites, one is dispersing inorganic fillers into a polymeric matrix by surface modification and/or thorough stirring to form 0–3 type composites,^5,6 and the other is using long fibers.⁷ The results of many researches presented in literature show that to get a complete and uniform dispersion of fillers (especially nanofillers) in polymer matrix is still a difficult and challenge topic today;^8,9 while the preparation of long ceramic fibers is usually not easy and expensive. Therefore, it is of great interesting to develop new methods to prepare novel ceramic/polymer composites.

Nanocomposites based on non-woven mats maybe a good choice to solve above problems because non-woven mats show the advantages of both nano-materials and long fibers. Electrospinning has been considered as the simplest and versatile nanotechnology for preparing nanofibers and non-woven mats. Many non-woven mats based on nanofibers have been successfully fabricated by electrospinning.^10–12 However, few articles focused on preparing electrospinning fiber/polymer composites.

On the other hand, it is well known that the interfacial nature plays a crucial influence on the performance of composites,^13,14 hence many methods have been developed to achieve desirable interfacial adhesion. For a long time, the chemical technique has been widely employed to improve the interfacial nature between different phases of a composite.¹⁵ However, the physical method is attracting increasing attention of researchers worldwide recently because it may overcome the drawback of chemical methods, and thus provides the greatest possibility for maintaining the outstanding properties of the original fillers/fibers.^16,17 Increasing the surface roughness of fibers is the typical reflection of physical methods,¹⁸ however, this is not easy to perform on the nanofibers, so we propose to design and synthesis nanofibers with pores, these pores are expected to be filled by the resin matrix, and thus provide a physical interaction between fibers and the resin matrix, this maybe an easy way to achieve a good physical interface although no literature has reported this so far.

Principles of composites point out that the properties of ceramic/polymer composites are greatly dependent on the nature of the ceramic and polymer, so they should be carefully selected. Zirconia (ZrO₂) is an important category of advanced ceramic which has been widely studied because of unique properties such as high toughness and chemical stability, good refractory properties and ionic conductivity.¹⁹ Many efforts have been devoted to the synthesis and application of ZrO₂ fibers by electrospinning,^20–27 however they used zirconyl chloride or zirconia as the precursor, this approach has some obvious drawbacks. The hydrolysis speed of either zirconyl chloride or zirconia is not rapid enough, hence the whole hydrolysis process will continue even after the electrospinning solution is ejected from the electrospinning setup orifice to form nanofibers on the collection set. That is to say, the hydrolysis will proceed continuously, and then makes the nanofibers formed initially to conglutinate together and partially lose their fibrous morphology during the continuous hydrolysis and condensation of the precursor nanofibers.²⁸ Hence the morphology of the resultant nanofibers is not easy to control; moreover, the production of zirconia electrospun fibers in batch size is not possible. Therefore, developing a new method to overcome the above disadvantages for facile preparation of ZrO₂ nanofibers is one target of our work presented herein.

With regard to polymeric matrix, thermosetting resins exhibit unique advantages on the processing characteristics over thermoplastics.^30,31Cyanate ester (CE) resin offers advantages as the composite matrix owing to its outstanding thermal stability, excellent dielectric property, high mechanical property, good moisture and radiation resistance; hence it is widely regarded as the right candidate with the great potential to prepare structural/functional materials of this century.^32,33 However, unfortunately, no literature discusses the preparation and characterization of ZrO₂/CE composites.

In this paper, a new method is set up to facilely fabricate zirconia (ZrO₂) nanofibers with controllable chemical and morphological structures in batch size on a standard electrospinning equipment, which employs zirconium propoxide as the precursor, and combines the advantages of electrospinning and sol–gel approaches. Based on this development, zirconia (ZrO₂) nanofibers with different morphologies were prepared, the influence of the preparing parameters on the structure of the ZrO₂ nanofibers was investigated; in addition, novel composites based on CE resin and non-woven mats of ZrO₂ nanofibers with different structure were prepared, and the influences of the surface nature and morphologies of ZrO₂ nanofibers on the typical properties of composites were systematically studied. The aim of this paper is developing a facile method to prepare ZrO₂ nanofibers with controlled structure, and corresponding composites with desirable high performance for cutting-edge industries.

2. Experimental

2.1 Materials

2,2′-Bis(4-cyanatophenyl)isopropylidene (CE) was bought from Zhejiang Shangyu Chemical Ltd., China. Zirconium propoxide (Zr(OCH₂CH₂CH₃)₄), polyvinylpyrrolidone (PVP, Mw≈1300000), acetic acid, and ethanol were all commercial reagents with analytic grades, and used as received.

2.2 Preparation of electrospinning solution

In a typical procedure, 0.6 g PVP was mixed with 2.0 mL of acetic acid and 10.0 mL of ethanol in a glove box to form a solution. After 30 min, 6.0 mL zirconium propoxide was added to the solution, followed by magnetic stirring for 2 h to get the electrospinning solution.

2.3 Preparation of ZrO₂ fibers and performs

The electrospinning solution was subsequently electrospun using a standard electrospinning equipment at a feeding rate of 9.0 mL h⁻¹ through a needle with a diameter of 1.23 mm. The accelerating voltage was 23.50 kV. The distance between the needle tip and the collector (a piece of flat aluminum foil) was 15 cm. The obtained as-spun fibers form non-woven mats, which were cut out into sheets with the suitable dimensions, and then piled up together to form a fiber congeries. The fiber congeries was subsequently dried at 60 °C for 12 h before calcination.

The fiber congeries was put into a calcinator at a preset calcination temperature for 2 h (the heating rate from room temperature to the calcination temperature is 10 °C min⁻¹), and then naturally cooled overnight in the calcinator to form a fiber preform. The resultant fiber preform is coded as ZrO₂(n), where n represents the calcination temperature, n is 600, 800, 1000, and 1200.

2.4 Preparation of ZrO₂/CE composites

Appropriate quantities of CE were heated to 150 °C and maintained at that temperature with stirring till a clear liquid was obtained.

ZrO₂(n) fiber preform was filled into the mould, and then heated to 150 °C. After that the melting CE was immediately cast into the mould, followed by degassing under vacuum at 150 °C for 10 min. The mold was then put into an oven for curing and postcuring taking the programmed procedures of 150 °C/2 h + 180 °C/2 h + 200 °C/2 h + 220 °C/2 h, and 240 °C/4 h, successively. The resultant composite was coded as ZrO₂(n)/CE.

The content of fibers can be calculated by the weights of the composite before and after calcination at 1200 °C for 2 h, which varies from 10.47 to 11.68 wt%, indicating that all composites prepared herein have the similar content of fibers, and thus the influence of the content of fibers on the properties of composites can be ignored.

2.5 Characterization

Field Emission Scanning Electron Microscopy (FESEM) (S4800, HITACHI Company, Japan) coupled with energy disperse X-ray spectrometer (EDS) was employed to observe the morphology of samples. All samples should be dried at 100 °C for 6 h, and then coated with gold before tests.

Powder X-ray Diffraction (XRD) measurements were performed on an X-ray diffractometer (X'Pert-Pro MPD, PANalytical B.V. Corporation, Netherlands) with Ni–filtered Cu Ka radiation. The XRD spectra were obtained by scanning over a 2θ angle range from 24 to 36° at a scanning speed of 2°/min and a step width of 0.028.

Fourier Transform Infrared (FTIR) spectra were recorded using a Nicolet 5700 FT-IR spectrometer (Thermo Electron Corporation, USA) with KBr for solid samples. The corresponding spectrum of pure KBr was regarded as the background spectrum.

Differential Scanning Calorimeter (DSC) analyses were done using a Diamond DSC (PE Corporation, USA) with a heating rate of 10 °C min⁻¹ in a nitrogen atmosphere.

The thermogravimetric (TG) analysis was inspected in an air atmosphere by a Simultaneous Differential Thermal Analyzer (SDT) (Q600, TA Instruments Company, USA) in the temperature range from 50 to 1200 °C with a heating rate of 10 °C min⁻¹

Dynamic Mechanical Analyses (DMA) were performed on a Q800 DMA (TA Instrument Company, USA) apparatus in a single-cantilever bending mode with a heating rate of 3 °C min⁻¹ and a frequency of 1 Hz.

The dielectric property was tested using a Broadband Dielectric Spectrometer (Novocontrol Concept 80, Germany) at room temperature over the frequency ranged from 1 to 10⁶ Hz. The dimensions of the sample were (10 ± 0.2) × (10 ± 0.2) × (1.5 ± 0.01) mm³.

3. Results and discussion

3.1 Synthesis and characterization of ZrO₂ fibers

As stated in the Introduction part, the physical interaction between fibers and the resin matrix is an important sort of interfacial adhesion of composites, so it is interesting to design and synthesize fibers with pores, these pores are expected to be filled by the resin matrix and thus provide a physical interaction between fibers and the resin matrix. However, in the case of ZrO₂ fibers, many efforts have been devoted to synthesize electrospun ZrO₂ fibers, but some undesirable problems exist as described in the Introduction part. In addition, it is worthy to point out that no ZrO₂ fibers with pores have been reported in the literature.

In this paper, we develop a new method exhibiting obvious advantages over previous techniques, which utilizes zirconium propoxide as the precursor, and combines the advantages of electrospinning and sol–gel approaches, consequently, the new method can facilely produce ZrO₂ electrospun fibers with controllable chemical and morphological structures in batch size on a standard electrospinning equipment. This attractive feature is attributed to the very fast hydrolysis speed of zirconium propoxide by moisture in the air, especially, in the case of the huge contact area between nanofibers and moisture in the air, so continuous networks gels of precursors are able to form once the electrospinning solution is ejected from the electrospinning setup orifice.²⁹ This means that the hydrolysis process will complete before the nanofibers reach the collect set, hence the chemical and morphological structures of nanofibers on the collection set do not change. Therefore, we can facilely produce zirconia electrospun fibers with controllable chemical and morphological structures in batch size using this method.

In order to evaluate the chemistry of ZrO₂(n) fibers calcined, the FTIR spectra of PVP and ZrO₂(n) fibers calcined at different temperatures were recorded and are shown in Fig. 1. All fibers show similar spectra, but in which no characteristic peaks (1663 cm⁻¹, 1291 cm⁻¹) reflecting PVP can be observed, indicating that organic materials have been removed from the fibers after calcination because PVP and carbon will completely decompose when the temperature is higher than 680 °C as shown in Fig. 2. On the other hand, all fibers exhibit a typical strong and wide peak at 3450 cm⁻¹, and a peak at 1640 cm⁻¹, assigned to the stretching vibration and the bending vibration of hydroxyl groups, respectively,³⁵ indicating that all fibers have hydroxyl groups which will a play important role in the curing behaviour of CE resin based composites as will be discussed later.

Fig. 1 FTIR spectra of PVP and ZrO₂ fibers calcined at different temperatures.

Fig. 2 TG curve of PVP in an air atmosphere.

Fig. 3 shows SEM images of as-spun zirconium propoxide/PVP fibers and their calcined products (ZrO₂(n) fibers), it can be seen that as-spun zirconium propoxide/PVP fibers have smooth surfaces and uniform distribution (Fig. 3a). After calcination at 600 °C, 800 °C or 1000 °C, the resultant ZrO₂(n) fibers have an average diameter of several hundred nanometres; importantly, they have pores but different pore dimensions, specifically, the dimensions of ZrO₂(600) and ZrO₂(800) fibers are only several nanometres (Fig. 3b–3c), while those of ZrO₂(1000) fibers are about tens of nanometres (Fig.3d). However, when the calcination temperature increases to 1200 °C, pores can not be observed on either surfaces or the internal parts of the resultant ZrO₂(1200) fibers as shown in Fig. 3e.

Fig. 3 SEM micrographs of electrospinning fibers calcinated at different calcination temperatures

The appearance of pores on the ZrO₂ fibers results from the thermal decomposition of PVP in the as-spun nanofibers during calcination. However, when the calcination temperature increases to 1200 °C, sintering will occur, and thus the resultant ZrO₂(1200) fibers do not have pores. These interesting results reveal a facile method to synthesize ZrO₂ nanofibers with nanometre pores on the interior or surface by a controlled calcination process. Based on the above phenomena, the non-woven mats made of these ZrO₂ nanofibers are expected to exhibit two types of pores, that is, the relatively large pores formed through the interconnection of the ceramic nanofibers, and the smaller pores within individual nanofiber.

It is known that ZrO₂ has three kinds of crystalline structures, they are tetragonal, monoclinic and cubic phases, so ZrO₂ with different crystalline structures will have different performances, and it is interesting to characterize the crystalline structure of ZrO₂ fibers synthesized herein. XRD measurements were carried out, and corresponding XRD patterns of ZrO₂(n) fibers are plotted in Fig. 4. ZrO₂(600) fiber has one sharp peak at 30.2°, which is the characteristic peak of t(111) for tetragonal phase.³⁴ The peak also appears in the XRD curve of ZrO₂(800); moreover, two new peaks at 28.2° and 31.5° can also be observed, they are the characteristic peaks of m (1̄11) and m(111) for the monoclinic phase, respectively, indicating that ZrO₂(800) fiber has both tetragonal and monoclinic crystalline structures. With regard to the XRD curves of ZrO₂(1000) and ZrO₂(1200) fibers, the intensities of the characteristic peaks of monoclinic phase significantly increase, but that of tetragonal phase greatly decreases, reflecting that the structure of ZrO₂(1000) and ZrO₂(1200) fibers mainly consist of monoclinic phase with a small amount of tetragonal phase.

Fig. 4 XRD curves of ZrO₂ fibers calcined at different temperatures.

The content of monoclinic phase in ZrO₂ fibers, coded as X_m, can be calculated by using eqn (1)

where I_m(1̄11) and I_m(111) are the intensities of monoclinic peak of ZrO₂, and I_t(111) is the intensity of tetragonal peak of ZrO₂.

Because ZrO₂(600) only has tetragonal crystalline structure, so its X_m value is 0. In the case of other three ZrO₂ fibers, all of them have both tetragonal and monoclinic phases. The X_m values of ZrO₂(800), ZrO₂(1000) and ZrO₂(1200) fibers are 0.26, 0.97 and 0.99, respectively. In other words, with the increase of the calcination temperature from 600 to 1200 °C, the crystalline structure of ZrO₂ fibers almost changes from a tetragonal phase to a monoclinic crystalline structure, while the crystalline grain tends to grow up at high temperature. These results are similar to the investigation reported by Erin Davies's group, although the fibers are not porous fibers.²¹

3.2 Effect of the nature of ZrO₂ fibers on the curing behavior and cured structure of composites

For a thermosetting system, its curing behavior determines the structure, and thus the performance of the cured resin, so it is necessary to first investigate the effect of ZrO₂ fibers on the curing behavior and that on cured structure of the crosslinked network. Fig. 5 gives the DSC curves of neat CE and uncured ZrO₂/CE composites. It can be seen that each sample shows a single exothermal peak, indicating that the addition of ZrO₂ fibers does not change the curing mechanism of CE, but the whole peak of either composite appears at obviously lower temperature compared with that of CE. For example, with the addition of ZrO₂ fibers to CE resin, the peak curing temperature of the resultant ZrO₂(600)/CE, ZrO₂(800)/CE, ZrO₂(1000)/CE and ZrO₂(1200)/CE composites shifts from 312 °C to 290, 274, 285, and 268 °C, respectively, suggesting that ZrO₂ fibers significantly accelerate the whole curing process of CE, but the catalyzing degree is closely related to the nature of ZrO₂ fibers.

Fig. 5 DSC curves of neat CE and uncured ZrO₂/CE composites.

As shown in the FTIR spectra of all ZrO₂ fibers (Fig.1) all fibers have hydroxyl groups, hence it is reasonable to observe the significantly decreased curing temperature because hydroxyl groups have a remarkable catalytic effect on the cyclotrimerization of –OCN to form crosslinked triazine rings.³⁶ On the other hand, Jung's group found that the concentration of hydroxyl groups on the surface of monoclinic ZrO₂ is generally higher than that on tetragonal ZrO₂,³⁷ hence it is not surprising to observe that ZrO₂ fibers with higher calcination temperature exhibit a bigger catalyzing effect on the curing of CE.

Generally, high curing and postcuring temperature is one of the key disadvantages of thermal resistant resins including CE resin,³⁸ how to overcome this drawback has been an important topic on developing high performance resins and related composites.^39,40 Therefore the improved curing behavior of ZrO₂/CE is very attractive for applications.

In fact, no reaction takes place if thoroughly purified cyanate esters are heated, however, cyanate esters are very sensitive to the materials surroundings, and thus can be applied for thermal curing. The curing mechanism of CE monomer is to form a triazine ring network by thermal cyclotrimerization, which can be catalyzed by active hydrogen including hydroxyl groups.⁴¹ The curing mechanism of CE with the presence of hydroxyl groups is complicated, specifically, the –OCN group of CE can react with hydroxyl group under the heating condition to form the iminocarbonate, a reactive intermediate product, which can further react with more molecules of CE to generate triazine ring.^42,43 Note that the hydroxyl group may become the end group of a triazine ring and terminate the further growth of the triazine ring network, hence the whole cured structure of the matrix in ZrO₂/CE composite is still the crosslinked network made up of triazine rings, but which contains different parts with different molecular weights rather than a whole homogenous bulk.

3.3 Effect of the structure of ZrO₂ fibers on properties of ZrO₂/CE composites

Fig. 6 shows the SEM images of the cross-sections for ZrO₂/CE composites, it can be seen that ZrO₂ fibers are randomly and well dispersed in the CE resin matrix, hence each composite is isotropic. Fig. 7 shows an overlay plots of storage modulus-temperature for cured CE resin and ZrO₂/CE composites. All composites exhibit higher storage moduli in glassy state than in the cured CE resin, this result is expected because ZrO₂ fibers have a higher rigidity than the CE resin.⁴⁴ In addition, the presence of inorganic fillers will bring a big restriction on the molecular movement. With regard to composites, it is interesting to find that the composite based on fibers calcined at higher temperature tends to have high storage modulus, however ZrO₂(600) fiber does not follow this trend, which does not exhibit the lowest modulus but shows a slightly higher storage modulus than ZrO₂(1000)/CE. This phenomenon can be explained from the influence of the structure of ZrO₂ fibers on the performance. First, many scholars researched the relationship between crystalline structure and mechanical property of ZrO₂, and found that ZrO₂ with higher content of tetragonal phase has higher strength, modulus and toughness.⁴⁵ In the case of ZrO₂ fibers synthesized herein, as the calcination temperature increases, the content of tetragonal phase gradually decreases as discussed above, hence the corresponding ZrO₂/CE composite is expected to have a gradually increased storage modulus. Second, besides the difference in the crystalline structure, whether there are pores on the internal and (or) surfaces of fibers is another difference among these fibers calcined at different temperatures. Generally, a structure with a bigger compact degree will have a higher modulus, meaning that the addition of ZrO₂(1200) to CE resin tends to prepare composite with the highest modulus, while that of ZrO₂(600) will lead to the opposite result. Third, TG and EDS analyses (Fig. 1 and 8) show that ZrO₂(600) fibers contain carbon, this is attributed to the char yield of PVP during calcinations. The weight loss (4.35%) from 600 °C to 1200 °C is caused by the oxidation of carbon. Because carbon has higher modulus than ZrO₂,⁴⁶ so the existence of residual carbon element on ZrO₂(600) fibers will make some contribution to increase the modulus. Obviously, the above two parameters play a combined role on the value of storage modulus, and consequently lead to the experimental results shown in Fig. 7.

Fig. 6 SEM micrographs of the cross-sections for ZrO₂/CE composites based on ZrO₂ fibers calcined at different temperatures.

Fig. 7 Overlay plots of storage modulus–temperature for cured CE resin and ZrO₂/CE composites.

Fig. 8 EDS curve of ZrO₂(600) fibers.

The effectiveness of fillers on the modulus of the composites can be represented by a coefficient C, which is calculated by using eqn (2):⁴⁷

where E'_g and E'_r are the storage modulus in the glassy and rubbery region, respectively. The measured E' values at (Tg − 250) °C and (Tg + 50) °C were employed as E'_g and E'_r, respectively, the calculated C values are summarized in Table 1. It can be seen that when the calcination temperature increases from 600 to 1000 °C, the resultant fibers show increased effectiveness on the modulus of resultant ZrO₂/CE composite reflected by the decreased C value; however, with further increasing the calcination temperature to 1200 °C, the resultant ZrO₂(1200) fibers have the biggest C value among the four kinds of fibers, and thus the weakest effectiveness on the modulus of composite.

Table 1 Typical properties of cured CE resin and ZrO₂/CE composites obtained from DMA tests

Sample	Coefficient C	Tg (°C)	Mc (g mol^−1)
CE resin	—	311	631.7
ZrO2(600)/CE	0.2465	291, 315	126.9
ZrO2(800)/CE	0.1818	300, 326	103.5
ZrO2(1000)/CE	0.1766	293, 326	102.3
ZrO2(1200)/CE	0.3564	291, 319	169.4

---

It is explained that as described above, ZrO₂(1000) fibers are porous with tens of nanometres diameter pores, so it is possible for more CE molecules to enter the pores of ZrO₂ fibers, triggering a greater physical interaction between the ZrO₂ fibers and the CE resin matrix. Comparatively, the smaller dimensions of pores for ZrO₂(600) and ZrO₂(800) fibers provide smaller physical interaction.

Fig. 9 shows overlay tanδ–temperature curves of cured CE resin and ZrO₂/CE composites. Since the tanδ peak occurs in a temperature range over which the polymer changes from a glassy state to an elastic state, hence the peak temperature of which is often taken as the glass transition temperature of a polymer. In addition, the shape of tanδ peak may be used as a convenient indicator of the morphology for multiphase materials.⁴⁷

Fig. 9 Overlay plots of tanδ versus temperature for cured CE resin and ZrO₂/CE composites.

From the view of molecular motion, the intensity of the tanδ peak represents the ability of the molecular motion, and the greater the intensity, the bigger the ability of the molecular movement.⁴⁸ Compared with the CE resin, the molecules in composites are not easy to move owing to the restriction of ZrO₂ fibers and the increased crosslinking density (which will be discussed below), therefore it is reasonable to find that the tanδ peaks of all composites have remarkable lower intensities than that of CE resin.

CE resin has a strong and sharp peak at about 311 °C, indicating that CE resin possess a single-phase structure; while each ZrO₂/CE composite exhibits a broad and complex shape of tanδ, which can be divided into two peaks appearing at 315–326 °C and 291–300 °C, respectively, by the Gaussian fitting,⁴⁹ reflecting that each composite has a multi-phase structure, and thus two Tg values. These results can be explained by the following reasons. First, it is known that CE monomer cures thorough the thermal cyclotrimerization to form triazine ring network. In the case of the ZrO₂/CE composites, they contain different parts including a triazine ring network as that of the pure CE resin, but also that terminated by hydroxyl groups as discussed above. Second, ZrO₂(600), ZrO₂(800) and ZrO₂(1000) fibers have pores which will be filled by the resin matrix during the fabrication process of the composites. Obviously, these filled resins have different transition behaviors from those outside the pores because the former is supposed to obtain a bigger restriction resulting from the limited space. Different from the first factor, the second factor can be regarded as a physical interaction. The above two factors are responsible for the broadened and complex tanδ peaks owned by ZrO₂/CE composites. Note that ZrO₂(1200) fibers do not have pores, meaning that the fibers only have the first parameter, and thus the ZrO₂(1200)/CE composite has the highest intensity among the four composites.

The crosslinking density is another important property of a thermosetting resin and related materials, which can be described by M_c, the molecular weight between entanglements or crosslinked sites.⁴¹ According to the physical meaning of M_c, it can be seen that M_c value is inversely proportional to the crosslinking density. The M_c value can be calculated by eqn (3):^50,51

where, q is the front factor, and is usually equal to 1; d is the density of the material (1.2 g cm⁻³ for ZrO₂/CE composites); R is the universal gas constant (8.314 J K⁻¹mol⁻¹); E'_{Tg+50 K} is the storage modulus at temperature T which is selected as (Tg + 50) °C from DMA tests.

The values of M_c calculated from eqn (3) are summarized in Table 1. All composites have a remarkably smaller M_c values than the CE resin, suggesting that the former has a significantly higher crosslinking densities than the latter. This can be explained from chemical and physical crosslinking aspects. As discussion above that –OH can catalyze the cyclotrimation of –OCN groups, meaning that under the same curing and postcuring procedures, the resin matrix in the composites has a bigger crosslinking density than the CE resin. On the other hand, ZrO₂(600), ZrO₂(800) and ZrO₂(1000) fibers have pores which allow much contact of the CE resin with –OH groups on the fibers, and thus increases the formation of hydrogen bonds between the –OH groups of the fibers, oxygen and/or nitrogen atoms of the triazine rings, this physical “crosslinking” also makes contribution to the whole crosslinking density. Because ZrO₂(1200) fibers do not have pores, the ZrO₂(1200)/CE composite has smaller crosslinking density than the other three composites.

Fig. 10 shows the dependence of dielectric constant (ε') on the frequency for neat CE resin and ZrO₂/CE composites over a wide frequency range at room temperature. The dielectric constant and its frequency dependence of ZrO₂/CE composites are determined by the structure of the ZrO₂ fibers. Specifically, the dielectric constant of ZrO₂(600)/CE composite shows obvious dependence on the frequency, this can be ascribed to the existence of carbon on the ZrO₂(600) fibers (Fig. 8). It is known that carbon is an electric conductor, which can easily induce space polarization owing to the charge accumulation at the interfaces, and affect the dielectric behavior as the frequency changes.^46,52

Fig. 10 Frequency dependence of dielectric constant of cured CE resin and ZrO₂/CE.

The above phenomenon does not appear for the ZrO₂(800)/CE, ZrO₂(1000)/CE and ZrO₂(1200)/CE composites, in opposite, they show similar and excellent stability of dielectric constant on frequency as the CE resin. Because ZrO₂ has big dielectric constant (∼23),⁵³ hence it is not a surprise to find that the ZrO₂/CE composites have higher dielectric constant than CE resin. However, it is interesting to note that the experimental values of dielectric constant for all ZrO₂/CE composites are smaller than their corresponding theoretical values calculated by the common “Rule of Mixture” (Table 2). Because the “Rule of Mixture” proposes that there is no interaction between two components,⁵⁴ obviously, this does not fit with the situation of ZrO₂/CE composites as discussed above. In fact, this phenomenon also confirms that there is a good interaction between ZrO₂ and CE resin. On the other hand, the pores of fibers contain air, of which the dielectric constant is as low as 1.0, hence the presence of air would reduce the dielectric constant of ZrO₂/CE composites. There is no pore in the ZrO₂(1200) fibers, so the dielectric constant of ZrO₂(1200)/CE composite is relatively close to the theoretical value.

Table 2 The theoretical and experimental dielectric constants of ZrO₂/CE composites

Sample	Dielectric constant (at 1 kHz)
	Theoretical	Experimental
CE resin	—	3.01
ZrO2(600)/CE	3.67	2.84
ZrO2(800)/CE	3.68	2.86
ZrO2(1000)/CE	3.75	3.16
ZrO2(1200)/CE	3.71	3.46
ZrO2 fibers	—	23

---

4. Conclusions

A new facile method is set up to fabricate zirconia (ZrO₂) nanofibers with controllably chemical and morphological structures in batch size on a standard electrospinning equipment, which employs zirconium propoxide as the precursor, and combines the advantages of electrospinning and sol–gel approaches.

By adjusting the calcination temperature from 600 to 1200 °C, ZrO₂ nanofibers with different morphologies can be facilely prepared, the different morphologies mainly include the existence of pores or not, the dimensions of the pores, and the crystalline structure. Although all ZrO₂ fibers have a significant catalytic effect on the curing of CE, the catalyzing effect is related to the structure of ZrO₂ fibers, so does the cured structure of the crosslinked network; in addition, the composites show different physically interfacial adhesion and macro-performance. Compared with neat CE resin, ZrO₂/CE composites have higher glassy and rubbery storage moduli, bigger crosslinking densities. The experimental values of dielectric constant for all ZrO₂/CE composites are smaller than their corresponding theoretical values calculated by the common “Rule of Mixture” owing to the good physical interaction between ZrO₂ and CE resin. These interesting results demonstrate that the method proposed herein provides a new approach to design and prepare ceramic nanofibers and corresponding composites with controlled structure and expected performance for cutting-edge industries.

Acknowledgements

The authors thank the National Natural Science Foundation of China (50873073, 20974076, 50903058), A Project Funded by the Priority Academic Program Development of Jiangsu Higher Education Institutions, the Major Program of Natural Science Fundamental Research Project of Jiangsu Colleges and Universities (11KJA43001), Provincial Natural Science Foundation of Jiangsu (BK2009124) and “Qing Lan Project” (2008) of Jiangsu Province in China for financially supporting this project.

References

1. A. Cravino and N. S. Sariciftci, J. Mater. Chem., 2002, 12, 1931.
2. C. Elissalde and J. Ravez, J. Mater. Chem., 2001, 11, 1957.
3. B. Liu, Y. Zou, S. Ye, Y. He and K. Zhou, RSC Adv., 2011, 1, 424.
4. J. Huang and T. Kunitake, J. Mater. Chem., 2006, 16, 4257.
5. P. G. Ren, G. Z. Liang, Z. P. Zhang and T. Lu, Composites, Part A, 2006, 37, 46.
6. T. Berzina, K. Gorshkov, A. Pucci, G. Ruggeri and V. Erokhin, RSC Adv., 2011 10.1039/C1RA00584G.
7. K. Sever, S. Erden, H. A. Gulec, Y. Sekid and M. Sarikanatb, Mater. Chem. Phys., 2011, 129, 275.
8. T. Paunikallio, M. Suvanto and T. T. Pakkanen, J. Appl. Polym. Sci., 2004, 91, 2676.
9. L. B. Marie and O. Kristiina, J. Thermoplast. Compos., 2009, 22, 115.
10. D. H. Reneker and I. Chun, Nanotechnology, 1996, 7, 216.
11. D. Li and Y. Xia, Nano Lett., 2004, 4, 933.
12. M. Macias, A. Chacko, J. P. Ferraris and K. J. Balkus, Microporous Mesoporous Mater., 2005, 86, 1.
13. R. Rusli and S. J. Eichhorn, Nanotechnology, 2011, 22, 325706.
14. J. L. Yao, C. X. Xiong, L. J. Dong, C. Chen, Y. A. Lei, L. Chen, R. Li, Q. M. Zhu and X. F. Liu, J. Mater. Chem., 2009, 19, 2817.
15. W. Z. Nie, X. Z. Li and F. F. Sun, J. Mater. Eng. Perform., 2010, 19, 1240.
16. P. G. Ren, G. Z. Liang, Z. P. Zhang and T. L. Liu, Composites, Part A, 2006, 37, 46.
17. Z. X. Jiang, Y. D. Huang, L. Liu and J. Long, Appl. Surf. Sci., 2007, 253, 9357.
18. W. Song, A. J. Gu, G. Z. Liang and L. Yuan, Appl. Surf. Sci., 2011, 257, 4069.
19. L. Kumari, G. H. Du, W. Z. Li, R. S. Vennilab, S. K. Saxena and D. Z. Wang, Ceram. Int., 2009, 35, 2401.
20. H. B. Zhang and M. J. Edirisinghew, J. Am. Ceram. Soc., 2006, 89, 1870.
21. E. Davies, A. Lowew, M. Sterns, K. Fujihara and S. Ramakrishna, J. Am. Ceram. Soc., 2008, 91, 1115.
22. A. M. Azad, Mater. Lett., 2006, 60, 67.
23. A. M. Azad, T. Matthews and J. Swary, Mater. Sci. Eng., B, 2005, 123, 252.
24. C. L. Shao, H. Y. Guan, Y. C. Liu, J. Gong, N. Yu and X. H. Yang, J. Cryst. Growth, 2004, 267, 380.
25. F. R. Lamastra, A. Bianco, A. Meriggi, G. Montesperelli, F. Nanni and G. Gusmano, Chem. Eng. J., 2008, 145, 169.
26. Y. Y. Zhao, Y. F. Tang, Y. C. Guo and X. Y. Bao, Fibers Polym., 2010, 11, 1119.
27. L. F. Yin, J. F. Niu, Z.Y. Shen, Y. P. Bao and S. Y. Ding, Mater. Lett., 2011, 65, 3131.
28. S. P. Wen, L. Liu and L. F. Zhang, Mater. Lett., 2010, 64, 1517.
29. D. Li and Y. Xia, Nano Lett., 2003, 3, 555.
30. M. Jayabalan, K. T. Shalumon, M. K. Mitha, K. Ganesan and M. Epple, Biomed. Mater., 2010, 5, 1.
31. Q. L. Zhang, X. Y. Ma, G. Z. Liang, X. H. Qu, Y. Huang, S. H. Wang and K. C. Kou, J. Polym. Sci., Part B: Polym. Phys., 2008, 46, 1243.
32. H. J. Hwang, C.H. Li and C. S. Wang, Polymer, 2006, 47, 1291.
33. M. Laskoski, D. D. Dominguez and T. M. Keller, J. Mater. Chem., 2005, 15, 1611.
34. C. L. Shao, H. Y. Guan, Y. C. Liu, J. Gong, N. Yu and X. H. Yang, J. Cryst. Growth, 2004, 267, 380.
35. W. Hertl, Langmuir, 1989, 5, 96.
36. M. C. Lu and J. L. Hong, Polymer, 1994, 35, 2822.
37. K. T. Jung and A. T. Bell, J. Mol. Catal. A: Chem., 2000, 163, 27.
38. S. K. Dai, D. X. Zhuo, A. J. Gu, G. Z. Liang and L. Yuan, Polym. Eng. Sci., 2011 DOI:10.1002/pen.21995.
39. C. P. R. Nair, D. A. Mathew and K. N. Ninan, Adv. Polym. Sci., 2001, 155, 1.
40. C. Marieta, M. del Rio, I. Harismendy and I. Mondragon, Eur. Polym. J., 2000, 36, 1445.
41. I. Hamerton, Chemistry and technology of cyanate ester resins. Blackie Academic & Professional, London, 1994, ch. 6, 155–167.
42. X. Sheng, M. Akinc and M. R. Kessler, Polym. Eng. Sci., 2010, 50, 1075.
43. M. F. Loustalot and C. L. Grenier, Eur. Polym. J., 1995, 31, 1139.
44. D. G. Purton, R. M. Love and N. P. Chandler, Oper. Dent., 2000, 25, 223.
45. C. L. Maria, S. S. Scherrer, P. Ammann, M. Jobin and H. W. Wiskott, Acta Biomater., 2011, 7, 858.
46. C. F. Han, A. J. Gu, G. Z. Liang and L. Yuan, Composites, Part A, 2010, 41, 1321.
47. L. A. Pothan, Z. Oommen and S. Thomas, Compos. Sci. Technol., 2003, 63, 283.
48. W. Ling, A. J. Gu, G. Z. Liang, L. Yuan and J. Liu, Polym. Adv. Technol., 2010, 21, 365.
49. M. A. Stephen and G. Steve, Langmuir, 2009, 25, 8152.
50. X. Sheng, J. K. Lee and M. R. Kessler, Polymer, 2009, 50, 1264.
51. G. Levita, S. D. Petris, A. Marchetti and A. Lazzeri, J. Mater. Sci., 1991, 26, 2348.
52. C. H. Lin, C. N. Hsiao, C. H. Li and C. S. Wang, J. Polym. Sci., Part A: Polym. Chem., 2004, 42, 3986.
53. D. P. Thompson, A. M. Dickins and J. S. Thorp, J. Mater. Sci., 1992, 27, 2267.
54. K. A. Malini, E. M. Mohammed, S. Sindhu, P. A. Joy, S. K. Date, S. D. Kulkarni, P. Kurian and M. R. Anantharaman, J. Mater. Sci., 2001, 36, 5551.

---