Introduction

Nickel ferrite (NiFe2O4) is one of the most important spinel ferrites as well as a typical spin soft-magnetic ferrite (Ref 1). Its cubic inverse spinel structure shows ferromagnetism, which originates from magnetic moment of anti-parallel spins between Fe3+ ions at tetrahedral sites and Ni2+ ions at octahedral sites (Ref 2, 3). NiFe2O4 composites, as semiconductor-natured polycrystalline spinel ferrites, are applicable in many magnetic devices because of their low electrical conductivity values compared to those of magnetic materials (Ref 4-7). Because of its high initial electrical resistivity at ambient temperatures, NiFe2O4 composite was seldom employed to serve as electrode material (Ref 8). However, the cubic inverse spinel structure provides better chemical stability in the molten cryolite-alumina bath (Ref 9). Hence, it is possible for nickel ferrite (NiFe2O4) to be used as a promising candidate for green inert anodes at high temperatures. In order to examine this possibility, many efforts have been made to study the NiFe2O4 composites working as inert anodes, and some progress has been achieved (Ref 9-12). It is a proven fact that it can produce environment-friendly O2 gas during electrolysis instead of greenhouse gases and fluorocarbons (Ref 12). Unfortunately, unsatisfactory fracture toughness and thermal shock resistance are still the obstacles preventing NiFe2O4 composite ceramics from being widely used, especially for applications in thermal shock conditions with high transfer and/or rapid environmental temperature changes, such as those experienced in electrolytic bath (Ref 13). One method for improving thermal shock resistance is to tailor the structure on multiple length scales to produce architectures that are engineered to enhance thermal shock resistance and improve fracture toughness while maintaining load-bearing capability. Another feasible method is to introduce a toughening phase, such as particles, whiskers, and fibers, into the ceramic matrix. It was recognized that fiber with certain aspect ratio is a promising candidate due to the complex toughening mechanisms (Ref 14). Since the discovery of the transformation toughening in ZrO2 (Ref 15), a considerable effort has been performed to understand the microstructure and mechanical properties of ZrO2-based ceramics. ZrO2 is of particular interest as it undergoes at least three crystallographic transformations (monoclinic, tetragonal, and cubic) when it cools from high temperature to room temperature (Ref 16).

It has been shown that the ceramics reinforced by continuous fibers have promising potential for high-temperature application, especially for the ceramics reinforced by Y2O3-stabilized ZrO2 fibers. It is due to the remarkable fracture toughening characteristic of the Y2O3-stabilized ZrO2f through three mechanisms: fiber bridging, fiber pullout, and crack deflection (Ref 17-19). It was expected that the toughness can further be improved when both fiber toughening and phase transformation are incorporated into the ceramic matrix. Nevertheless, there are few reports about ZrO2 fiber-toughened NiFe2O4 composites with respect to the cooperative toughening.

In this paper, to further improve the fracture toughness of NiFe2O4 composite ceramics as inert anodes in severe environments, NiFe2O4 composite ceramics toughened with ZrO2 fiber were fabricated with a two-step sintering process. The microstructure and the mechanical properties of NiFe2O4 composite ceramic reinforced with ZrO2 fiber (ZrO2f) were investigated and discussed in detail. Meanwhile, the pyroconductivity of ZrO2 fiber-reinforced NiFe2O4 composites that serve as inert anodes working in molten cryolite is also studied.

Experimental

Commercially available NiO powder (>99.98%, Guoyao Co. Ltd., PR China), Fe2O3 powder (>99.3%, Xincheng Co. Ltd., PR China) were used as the starting powders. Yttria (Y2O3; 6.66%, mole)-stabilized commercial ZrO2 fiber with 3-8 μm in mean diameter and 200 μm in medium length was also employed for the preparation of the composite ceramic material.

Raw materials (NiO power and Fe2O3 powder) were ground in distilled water via a ball-milling process using polypropylene jars with yttria-stabilized zirconia balls for 24 h and then dried thoroughly. The molar ratio of NiO to Fe2O3 was 1.87:1 in the powder mixture. Then, the mixtures were ground with 4 vol.% polyvinyl alcohol (PVA) binder and cold pressed at 160 MPa into a cylinder (ϕ100 mm × 20 mm) using a stainless steel die. The green ceramic bodies were calcined at 1000 °C for 6 h in air to produce NiFe2O4 ceramic matrix. The calcined matrix products were crushed thoroughly and ball-mixed with different amounts of ZrO2f (0-4 wt.%) using ethanol as the grinding media. The properties of ZrO2f adopted in this paper are listed in Table 1. After mixing for 24 h, the slurry was dried thoroughly in a vacuum oven. The resulting mixtures with 4 vol.% PVA binder were cold-pressed into rectangular bars (70 mm × 15 mm × 8 mm) using a stainless steel die under a uniaxial load of 200 MPa for 5 min. The bars were sintered at 1300 °C for 6 h in air atmosphere with a heating rate of 10 °C/min to produce ZrO2f-reinforced NiFe2O4 composite ceramics and then cooled down to ambient temperature.

Table 1 Properties of ZrO2 fibers adopted in this paper

The microstructural features of sintered composites were observed using scanning electron microscope (SEM) (SSX-550, Japan) with simultaneous chemical analysis by energy-dispersive spectroscopy (EDS; EDAX Inc.). The phase composition was determined by a D/max 2RB x-ray diffraction (XRD, Japan) with Cu Kα radiation, pip voltage 40 kV, and current 100 mA. The bulk density of the specimens was measured by the Archimedes method. Bending strength was tested by three-point bending method on 70 mm × 15 mm × 8 mm bars, using a 30-mm span and a crosshead speed of 0.5 mm/min. The edges of all the specimens were chamfered to minimize the effect of stress concentration due to machining flaws. Fracture toughness (K IC) was evaluated by a single-edge notched beam test with a 30-mm span and a crosshead speed of 0.05 mm/min using 70 mm × 15 mm × 8 mm bars, on the same jig used for the flexural strength measurements. All flexural and fracture bars were cut with tensile surface perpendicular to the cold-pressed direction. A minimum number of five specimens were tested for each experimental condition.

The electrical pyroconductivity of solid materials was generally measured using a special device, which was generally combined with a heating furnace and based on the standard principles of DC two-point, four-point, or four-electrode technique (Ref 10, 20). Because of the difficulties in obtaining a good electrical contact at both ends of the specimen with the two-point technique, the four-point or four-electrode technique was adopted to determine the electrical pyroconductivity of NiFe2O4 composite ceramics, using the measuring apparatus as described in detail by Lai et al. (Ref 21).

The electrical conductivity σ is determined by the following equation (Ref 22):

$$ \upsigma = IL/US $$
(1)

where σ is the conductivity; I is the current intensity through the tested sample; L and U are the interval and voltage drop between two electrodes, respectively; and S is cross section of tested sample.

The measurements of conductivity for NiFe2O4 composites were carried out at temperatures ranging from 600 °C to 1000 °C with a temperature interval of 50 °C.

Results and Discussion

Microstructure and Phase Identification

The green cold-pressed NiO-Fe2O3-ZrO2f ceramic billets with bulk densities of ~3.13 g/cm3 were used to produce ZrO2f-reinforced NiFe2O4 composite ceramics. The fracture surface of ZrO2f-reinforced ceramic samples sintered at 1300 °C for 6 h is shown in Fig. 1.

Fig. 1
figure 1

SEM images of NiFe2O4 composites with different amounts of ZrO2 fibers: (a) 1 wt.%; (b) 2 wt.%; (c) 3 wt.%; (d) 4 wt.%; and (e) higher magnification SEM image of polished surface from 3 wt.% ZrO2f-reinforced NiFe2O4 composites

It can be seen that when the ZrO2f content is lower, the dispersion of fibers is more uniform. There is no obvious occurrence of agglomeration phenomena for ZrO2f in the ceramic matrix, such as in the cases of 1 and 2 wt.%, as shown in Fig. 1(a) and (b), respectively. For 3 wt.% ZrO2f-reinforced ceramic samples, ZrO2 fibers were distributed uniformly at the ceramic matrix. Trajectories of many holes for fiber pull-out can be seen (Fig. 1c), which was confirmed by SEM. When 4 wt.% ZrO2f was adopted in the ceramic matrix, obvious agglomeration and uneven distribution for ZrO2f can be observed. It is the immediate cause of increase in porosity of sintered samples (Fig. 1d). In order to observe more clearly the interface between ZrO2f and ceramic matrix, a larger magnification SEM image of polished surface from 3 wt.% ZrO2f-reinforced NiFe2O4 composites is illustrated, in Fig. 1(e). It can be seen that ZrO2 fiber is bonded tightly to NiFe2O4 composite ceramic matrix.

In order to further identify the phase transformation in the sintered samples, XRD patterns of ZrO2 fibers adopted in this study and the synthesized NiFe2O4 composite ceramics with 3 wt.% ZrO2f were also detected, and the detection results are shown in Fig. 2 and 3, respectively.

Fig. 2
figure 2

XRD analysis of ZrO2 fibers

Fig. 3
figure 3

XRD patterns of 3 wt.% ZrO2f-reinforced NiFe2O4 composites with different sintering temperatures

Besides Y2O3, only Y2O3-stabilized tetragonal zirconia polycrystals (Y-TZP) were observed for the ZrO2 fibers adopted in this study. The XRD detection results for the polished surface of 3 wt.% ZrO2f-reinforced ceramics, sintered at 1200, 1250, 1300, 1350, and 1400 °C for 6 h, are shown in Fig. 3. Besides the synthesized NiFe2O4 and excessive NiO in matrix, only the tetragonal phases for ZrO2 are present. No phase transformation for ZrO2 fibers occurred during the sintering process with different temperatures. It is demonstrated that well-stabilized ZrO2 fibers are obtained by Y2O3 (6.66%, mole) here.

Mechanical Properties

Results of the density and mechanical properties for NiO-Fe2O3-ZrO2f ceramics sintered at 1300 °C for 6 h are listed in Table 2. It can be seen from Table 2 that the volume density decreases with increasing ZrO2f content from the overall perspective. For all the samples sintered under the same condition, there is no difference in sintering driving force of NiO-Fe2O3-ZrO2f system. Increasing ZrO2f content would induce a higher average shear stress to the sintered ceramic matrix, which would result in a decrease in the sintering velocity. Simultaneously, it is difficult to make ZrO2f disperse uniformly in the matrix. It is relatively easier for ZrO2f to form three-dimensional rigid network skeleton producing bridges and springbacks among ZrO2 fibers, which leads to structural defects and lower volume densities of sintered samples (Ref 14).

Table 2 Density and mechanical properties of NiO-Fe2O3-ZrO2f ceramics

Bending strength increases obviously with ZrO2f content increasing from 0 to 3 wt.%, but above 3.0 wt.%, a small change in value has been observed. For example, the bending strength increases from ~67.1 MPa for undoped samples to ~85.8 MPa for 3 wt.% ZrO2f-doped samples. The variation trend of bending strength versus ZrO2f content in this study is different from that of ZrO2f-reinforced ZrB2-SiC ceramics (Ref 23). It may be attributed to the different sintering temperatures of NiFe2O4 composites compared with ZrB2-SiC-ZrO2f system.

Generally speaking, for fiber or whisker-reinforced composites, in an optimal situation with fiber in a unidirectional arrangement, the strength σ can be represented by the following equation:

$$ \upsigma = \upsigma_{\text{m}} V_{\text{m}} + \upsigma_{\text{f}} V_{\text{f}} $$
(2)

where σm is matrix strength, V m is the volume fraction of matrix, σf is fiber strength, and V f is the volume fraction of fiber. There is a critical fiber volume fraction (P c) to form rigid network skeletons in matrix (Ref 24). When V f < P c, σ will increase with increasing fiber content for the domination of fiber reinforcement on ceramic matrix; when V f ≥ P c, σ is dependent on the densification of composites. In this paper, the variation trend of bending strength versus ZrO2f content is in accordance with Ref 24.

It can obviously be seen from Table 2 that the fracture toughness was influenced by ZrO2f content. With increasing ZrO2f content from 0 to 3 wt.%, the fracture toughness increased from ~2.57 MPa m1/2 to the maximum value of ~.4.67 MPa m1/2. Upon increasing the ZrO2f content to 4 wt.%, the fracture toughness attains a value of 4.38 MPa m1/2, which is lower than that obtained with 3 wt.% ZrO2f-reinforced NiFe2O4 composites. There is a decreasing trend in the fracture toughness for ZrO2f-reinforced NiFe2O4 composites when ZrO2f content is increased from 3 to 4 wt.%. The possible reason is that fiber agglomeration occurred in the ceramic matrix with 4 wt.% ZrO2 fiber, which might lead to structural defects in the composites. The fracture toughness was decreased when the ceramic composite was subjected to an external load. It is demonstrated that an optimal level of ZrO2 fiber is of advantage to fracture toughness of NiFe2O4 composites, i.e., 3 wt.% ZrO2 fiber in this case.

The effective toughening mechanisms of ZrO2 fibers in composite ceramics including fiber debonding, fiber pull-out, crack deflection, crack bridging, and phase transformation were identified (Ref 23, 25-27). According to the results from XRD patterns from Fig. 2 and 3, it can be concluded that the phase transformation for ZrO2 fibers to reinforce the ceramic matrix did not happen here. ZrO2 fibers always remain tetragonal phases during sintering process here. SEM images of the fractured surface of 3 wt.% ZrO2f-reinforced NiFe2O4 composites are shown in Fig. 4. A uniform distribution of short ZrO2 fibers in NiFe2O4 matrix can be seen in Fig. 4(a). Figure 4(b) is a sketch of partial enlargement, which is marked as B in Fig. 4(a). According to Ref 27, in most cases, the crack propagates straight through the fibers with little deflection because of the low fracture strength of the fibers. However, on the fractured surface of the NiFe2O4 ceramics, some short fibers’ debonding and no cracks propagating straight through the fibers were observed, as shown in Fig. 4(b). Because of the debonding at the fiber/matrix interface, the matrix crack did not penetrate the fibers, which then remained to bridge the matrix crack and restrained the crack from propagating (Ref 28). Addition of ZrO2 fibers is beneficial for the enhancement of the fracture toughness of the NiFe2O4 ceramics.

Fig. 4
figure 4

SEM images of the fractured surface of 3 wt.% ZrO2f-reinforced NiFe2O4 composites at different magnifications

Fibers’ pull-out and rough fracture surfaces of the ceramics were also detected, as shown in Fig. 4(c). It can be seen from Fig. 4(d) that a perfect interface between ZrO2 fiber and the matrix was observed, which also indicated that there was no obvious reaction between NiFe2O4 ceramics and ZrO2 fibers. It is in accordance with the XRD results from Fig. 3. The enhancement in the mechanical properties of the ceramic matrix, which is provided by perfect interface between ZrO2 fibers and the synthesized NiFe2O4, is similar to that observed in ZrB2-SiC composites (Ref 29).

The analysis from Fig. 4 indicated that the crack acquired not only the propagating resistance of matrix but also the interfacial shear resistance to cause the fibers pull-out and debonding. At the same time, the rough fracture surface of the ceramics (Fig. 4d) implied that crack deflection might be another toughening mechanism because the crack swerving and twisting along fiber/matrix interface would expend more energy than that expended by crack propagating straight (Ref 30). Thus, the extensive crack interactions among the short fibers during fracturing process increased the fracture surface energy and thus contributed to the improvement in the toughness, marked as the region A in Fig. 4(a).

Pyroconductivity

Besides the analysis of reinforcement for ZrO2 fiber on the mechanical properties of NiFe2O4 composites, the pyroconductivity of ZrO2f-reinforced NiFe2O4 as inert anodes was also studied. Arrhenius equation for the conductivity of normal semiconductors, like NiFe2O4, was suggested as follows (Ref 31):

$$ \uprho_{\text{DC}} = \uprho_{0} \exp \left( \frac{E}{kT} \right) $$
(3)

where ρDC is DC electrical resistivity; E is the activation energy in electron volts (eV), which is needed to release an electron from one ion to the neighboring one giving rise to the electrical conductivity (Ref 32); k is the Boltzmann’s constant; and T is the absolute temperature. Conductivity value of σ is the inverse of ρDC. The conductivity versus temperature curve of the ZrO2f-reinforced NiFe2O4 composites is shown in Fig. 5. The same temperature interval of 50 °C was applied to study the conductivity of each ZrO2f-reinforced NiFe2O4 sample at temperatures ranging from 600 to 1000 °C.

Fig. 5
figure 5

The conductivity vs. temperature of ZrO2f-reinforced NiFe2O4 composites

From Fig. 5, the conductive behavior is observed when the temperature is greater than 600 °C. Also, conductivity increases with increasing temperature up to 850 °C, a threshold marked as a small square in Fig. 5. In the range from 600 to 850 °C, the conductivity of the undoped samples is greater than that of the ZrO2f-reinforced NiFe2O4 samples. However, it is interesting to note that the conductivity of the ZrO2f-reinforced samples is much better than that of the undoped samples when the temperature is greater than 850 °C. The more the content of ZrO2 fiber that was introduced into the ceramic matrix, the higher the value of conductivity that could be obtained. When temperature reaches up to 950 °C, the conductivity value of the samples reinforced with 4 wt.% ZrO2 fibers reaches 0.63 S/cm, which shows an increment of 37.8% compared with 0.45 S/cm obtained for the undoped samples.

However, nearly no conductive behavior for ZrO2 fiber is obviously observed at low temperature, as the introduction of ZrO2 fiber into the ceramic matrix decreased the conductivity of the composites. It can be concluded that the conductivity of ceramic composites is dependent on NiFe2O4 matrix. Also with increasing ZrO2f content, a higher porosity will be induced in the matrix, which is a barrier for good interaction among conductive phases. It is detrimental to the improvement of conductivity. However, ZrO2f-reinforced NiFe2O4 obtains better conductivity at higher temperature (>850 °C). This phenomenon is probably attributed to the admittance of partial Zr4+ ions into the lattice of NiFe2O4, which will substitute partial Fe3+ ions to form a solid solution. It can lead to lattice transformation of NiFe2O4. This transformation could be confirmed by the differences in the values of the lattice parameters a, b, and c: 0.83370, 0.83370, 0.83370 nm for the undoped and 0.83379, 0.83379, 0.83379 nm, for the 3 wt.% ZrO2f-reinforced ceramic samples, respectively. Hence, the substitution may produce quasi-free electrons as follows (Ref 33):

$$ 3{\text{ZrO}}_{2} \mathop{\longrightarrow}^{{{\text{NiFe}}_{2} {\text{O}}_{4} }}3{\text{Zr}}^{\prime}_{\text{Fe}} + 6{\text{O}}_{\text{O}}^{X} + V^{\prime\prime\prime}_{\text{Fe}} $$
(4)

where \( {\text{Zr}}^{\prime}_{\text{Fe}} \) represents the occupancy of Fe3+ ions by Zr4+ ions, \( {\text{O}}_{\text{O}}^{X} \) means the oxygen ions at normal lattice point, and \( V^{\prime\prime\prime}_{\text{Fe}} \) represents the cation vacancy. The substitution effect of partial Zr4+ ions to produce quasi-free electrons for the enhancement of electrical conductivity is limited by the concentration of conductive ions produced in this way.

There is another conductive mechanism—oxygen ionic conduction in ZrO2f for improving the conductivity of the ceramic matrix. For the Y2O3-stabilized ZrO2f, Y3+ ions occupied the sites of Zr4+ ions after the addition of Y2O3 to ZrO2f, as shown by the following equation (Ref 34):

$${\text{Y}}_{2} {\text{O}}_{3} \mathop{\longrightarrow}^{{{\text{ZrO}}_{2} }}2{\text{Y}}^{\prime}_{{{\text{Zr}}}} + V_{{{\ddot{\text{O}}}}} + 3{\text{O}}$$
(5)

In the crystal lattice with a couple of Y3+ substitutions, an oxygen vacancy could be produced. Zr-O spacing is 0.211 nm, which is 0.017 nm smaller than that of Y-O, and so anionic vacancy is close to Zr4+. Vacancy distribution around Zr4+ in ZrO2 crystal lattice will decrease the partial repelling force of O-O and result in coordination sphere distortion, which is beneficial to filling the lattice defects of ZrO2 (Ref 35). Consequently, the crystal transformation of ZrO2 could be avoided, and the Y2O3-stabilized ZrO2 can remain as tetragonal phase during the sintering process. Also, the ionic conductance (σi) of the Y2O3-stabilized tetragonal ZrO2 is strongly dependent on crystal structure and temperature. Ionic conductance (σi) of tetragonal ZrO2 is much higher than that of monoclinic and cubic phases, and σi increased exponentially with the increasing temperature (Ref 36). Hence, when electrolyte temperature is increased up to 1000 °C, the conductivity of ZrO2(f)/NiFe2O4 composites is increase to much higher than that of NiFe2O4 samples.

Conclusions

In the present study, ZrO2f-NiO-Fe2O3 composites were prepared through a two-step sintering process. XRD of the prepared samples shows that NiFe2O4, NiO, and ZrO2 phases are the main phases present in the sintered ceramic matrix. ZrO2 fiber remains as tetragonal phase during the sintering process, and no phase transformation was detected. Among all the doping levels, there is a uniform distribution of short ZrO2 fibers in 3 wt.% ZrO2f-reinforced NiFe2O4 matrix. The volume density decreases with the increasing ZrO2f content from the overall perspective. The values for both the bending strength and fracture strength of 3 wt.% ZrO2f-reinforced NiFe2O4 samples reached their respective maximum values of ~89.0 MPa and ~4.67 MPa m1/2. The observed toughening mechanisms were attributed to the fiber debonding, fiber bridging, crack deflection, and fiber pull-out toughening. The conductivity of ZrO2f-reinforced NiFe2O4 is dependent on temperature and ZrO2f content. When the electrolytic temperature is increased up to 950 °C, the conductivity value of the samples reinforced with 4 wt.% ZrO2 fibers attains 0.63 S/cm, which shows an increment of 37.8% compared with 0.45 S/cm obtained for the undoped samples. The two main conductive mechanisms of ZrO2 fiber in the matrix are the substitution of Fe3+ to produce quasi-free electrons and oxygen ionic conducting mechanism.