Influence of V/Ni-doping on electrical properties and aging stability of ZnFe₂O₄-based NTC ceramics

Abstract

V/Ni co-doped zinc ferrite (Zn_1-yNi_yFe_2-xV_xO₄, x = 0.01, 002, 0.03, 0.04, 0.05; y = 0.05, 015, 0.25) ceramics were prepared by solid-state reaction method. The effect of V and Ni contents on the electrical property and aging property of the ceramics were investigated. All the ceramics have a cubic spinel structure and show typical characteristics of negative temperature coefficient (NTC) of resistivity. For being doped with various V and Ni contents, the room temperature resistivities (ρ₂₅) of the ceramics range from 0.037 MΩ·cm to 4.8 MΩ·cm, and NTC material constants (B values) vary between 3255 K and 6614 K. The Zn_1-yNi_yFe_2-xV_xO₄ ceramic with x = 0.04 and y = 0.15 has high electrical stability with a resistance drift rate of 0.19% after having been aged at 150 °C for 500 h in air. The aging characteristic was investigated by analyzing cationic valences using X-ray photoelectron spectroscopy before and after aging treatment.

1 Introduction

Negative temperature coefficient (NTC) thermistors have characteristics of decreasing resistivity with temperature increasing. For their unique properties, NTC thermistors are widely applied in fields such as temperature monitoring, temperature compensation, surge current suppression, bolometric detection, and infrared imaging [1,2,3]. In recent years, the investigations of NTC thermistors are mostly based on AB₂O₄ spinel manganates and perovskite compounds [4,5,6,7]. NTC ceramics based on simple oxide-based semiconductors have also attracted attention due to their unique characteristics [8,9,10]. Material constant (B value) representing temperature sensitivity, room temperature resistivity (ρ₂₅) and electrical stability are important parameters for NTC thermistors. Different B values and ρ₂₅ are required for various applications. It is reported that the properties of NTC thermistors can be adjusted and improved by cation doping or process improvement [11,12,13,14,15,16,17]. For examples, Xiong et al. reported that the Mn_1.5−0.5xCo_0.9−0.3xNi_0.6−0.2xCu_xO₄ ceramics prepared by heterogeneous precipitation have adjustable B values and ρ₂₅ by substitution with different contents of Cu ions [15]. Wang et al. reported that the B value and ρ₂₅ of Co_2.77Mn_1.71Fe_1.10Zn_0.42O₈ ceramics could be adjusted by changing sintering temperatures [17].

In the meanwhile, electrical stability is also a key property for NTC thermistors. Many studies have reported on the aging phenomena of NTC ceramics. Grone et al. proposed two aging models basing on the study on the Mn–Ni–Fe–O system [18, 19], i.e., absorption or release of oxygen, and redistribution of cations at the sublattices. The effect of chemical composition on the aging of Ni_xMn_3−xO_4+δ (0.56 ≤ x ≤ 1.0) ceramics was studied by Fang et al. [20], the ceramics with a nickel content of less than 0.7 showed high electrical stability with resistance drift rates less than 0.5%. The composition dependence of aging performance is proved by using the conductive mechanism of small polaron theory and the thermodynamic optimal position of cation redistribution [20]. Recently, Zhao et al. studied the influence of Cu ions on aging property of Ni_0.7Mn_2.3-xCu_xO₂ (x = 0.1, 0.2, 0.3, and 0.4) NTC ceramics by analyzing the change of cationic valences using X-ray photoelectron spectroscopy (XPS) [21]. Their results showed that Cu-ionic dopant greatly reduced the resistivity and B value, but, the rise of Cu-ionic content increased the adding-induced drift in both resistivity and B value, resulting from the change of Cu-valence from Cu⁺ to Cu²⁺ as well as the decrease of Mn⁴⁺ content after aging.

Spinel-type ferrites (denoted as MFe₂O₄, M = Ni, Co, Mg, Cu, Mn, Zn, etc.) such as CuFe₂O₄ and MnFe₂O₄-based ceramics also have good NTC properties [22, 23]. The CuFe₂O₄-based ceramics reported by Liu et al. showed that the B value can be adjusted from 2971 K to 4180 K by altering the content of the Mo ions additive [22]. As reported by Guo et al. [23], the properties of MnFe₂O₄-based ceramics can also be modified by using Mo ions additive and the B value reached about 6100 K. The conduction mechanism of those compounds was proposed to follow the polaron hopping model, namely Fe³⁺  + Fe²⁺  ↔ Fe²⁺  + Fe³⁺ and Mn³⁺  + Mn²⁺  ↔ Mn²⁺  + Mn³⁺. However, to our knowledge, there has no report on the aging stability of the NTC thermistors based on spinel-type ferrites so far. In this work, V/Ni co-doped ZnFe₂O₄ ceramics were synthesized, the electrical properties and aging characteristics were investigated. The results show that ZnFe₂O₄ ceramics have a typical NTC effect, and room temperature resistivity (ρ₂₅) and temperature sensitivity constant (B value) can be adjusted in a certain range. The co-doped samples show better aging stability, the composition and valence of elements were analyzed by XPS, and the aging effect of ZnFe₂O₄-based ceramics were discussed.

2 Experimental

2.1 Sample preparation

The nominal composition of ZnFe_2-xV_xO₄ (x = 0, 0.01, 0.02, 0.03, 0.04 and 0.05; denoted as A0, A1, A2, A3, A4 and A5, respectively) and Zn_1-yNi_yFe_1.96V_0.04O₄ (y = 0.05, 0.15 and 0.25; denoted as S1A4, S2A4 and S3A4, respectively) ceramics were prepared by conventional solid-state reaction method. Zinc oxide (ZnO, ≥ 99.0%, Chinese pharmaceutical chemical reagent, China), iron citrate (FeC₆H₅O₇·5H₂O, ≥ 97%, Chinese pharmaceutical chemical reagent, China), vanadium pentoxide (V₂O₅, > 99%, Sinopharm Chemical Reagent Co, Ltd, China) and basic nickel carbonate (NiCO₃·2Ni(OH)₂·4H₂O, ≥ 99.0%, Xilong Scientific, China) were used as the raw materials. For each batch of the experiment, stoichiometric raw materials were weighed according to the designed composition. The starting materials were mixed with ball-milling, and the slurry was dried at 110 °C for 24 h. Then, mixtures were calcined in air at 900 °C for 5 h. The calcined powders were granulated with polyvinyl alcohol solution and pressed into disks with a diameter of 12 mm and thickness of about 2 mm. The green pellets were sintered in air at 1350 °C for 1 h. To test the electrical properties, two opposite surfaces of the as-sintered ceramics were polished and then coated with silver paste followed by being heated at 600 °C for 5 min to make the ohmic electrodes.

2.2 Characterization

Phase components of the sintered ceramics were detected by X-ray diffraction (XRD, Rigaku D/max 2500, Japan) at room temperature. The diffraction angle 2θ was selected in the range between 10° and 80°. The densities of sintered samples were measured by the Archimedes drainage method. The microstructure of ceramics was observed by using a field emission scanning electron microscopy (FESEM, sirion200). In SEM observation, the sintered ceramics were broken into pieces and the fracture morphology was observed. X-ray photoelectron spectroscopy (XPS, ESCALAB 250Xi, Thermo Fisher, USA) was used to analyze possible valences of cations in ceramics. The standard C 1 s peak (binding energy of 248.8 eV) was used for calibration.

Temperature dependence of resistance of all ceramics was measured at 25 °C–250 °C using a resistance temperature measurement system (ZWX-C, China) under direct current (DC) condition. The resistivity (ρ) was calculated based on the measured resistance (R) according to the equation of ρ = R(S/h), where S and h are, respectively, electrode area (cm²) and thickness (cm) of the measured sample. An electrochemical test system (Gamry reference 600, USA) was used to measure complex impedance spectra of ceramics. Impedance data were analyzed by Gamry Echem Analyst program. To test electrical stability, the aging treatment of samples with ohmic electrodes was performed at 150 °C for 500 h in air. The resistances were measure at room temperature (25 °C). The resistance drift rate was calculated by [(R_A–R₀)/R₀] × 100%, where R₀ is original resistance and R_A is resistance after being aged.

3 Results and discussion

3.1 Phase and microstructure

Figure 1 shows XRD patterns of some typical ceramics. All XRD patterns show cubic spinel phase with space group Fd $\overline{3 }$ m (227) (Ref. PDF No. 79–1150). In the partially enlarged illustration shown in the inset, slight impurity peaks reflected from the NiV₂O₆ phase (Ref. PDF No. 76–0359) can be detected only in the S3A4 sample. Refined by the Jade-6 analysis program, the related lattice parameters are obtained and shown in Table 1. The lattice parameters changed due to doping with V or Ni ions, indicating that V and Ni ions have substituted into the spinel lattice. The lattice parameters of ZnFe₂O₄ ceramics decrease with the increase of V-ionic content, mainly because the radius of V⁵⁺ ion (0.059 nm) [24] is smaller than that of Fe³⁺ ion (0.064 nm) [22] and Zn²⁺ ion (0.074 nm) [27]. With the increase of Ni ionic content, the lattice parameter continues to decrease. When y is 0.25 in Zn_1-yNi_yFe_1.96V_0.04O₄, impurity NiV₂O₆ was detected, indicating the content of Ni for y = 0.25 exceeds the related solid solubility in the studied ceramic.

Fig. 1

XRD patterns of the as-sintered ceramics with various contents of V and Ni ions. The inset is a partially enlarged view

Table 1 Lattice parameters (a) and relative densities of Zn_1−yNi_yFe_2−xV_xO₄ ceramics with various contents of V and Ni ions

SEM images of samples A0, A4, S2A4 and S3A4 are shown in Fig. 2. It can be seen that the grains are closely connected to each other, evenly distributed, and few pores can be found. According to measurement by the Archimedes method, relative densities of ceramics were obtained and are shown in Table 1. The relative density of doped samples is higher than that of undoped sample, which should be attributed to the existence of V₂O₅ which has low melting point and promotes the sintering.

Fig. 2

SEM images of as prepared ceramics, a ZnFe₂O₄, b ZnFe_1.96V_0.04O₄, c Zn_0.85Ni_0.15Fe_1.96V_0.04O₄, d Zn_0.75Ni_0.25Fe_1.96V_0.04O₄

3.2 Electrical properties

Figure 3a shows the temperature dependence of resistivity of ZnFe_2-xV_xO₄ with different V contents. All ceramics show typical NTC characteristics. Each plot is approximately linear and follows the Arrhenius relation as shown in Eq. (1).

$$\rho = \rho_{0} \exp \left( {\frac{{E_{{\text{a}}} }}{kT}} \right) = \rho_{0} \exp \left( \frac{B}{T} \right)$$

(1)

Fig. 3

Electrical properties of Zn_1−yNi_yFe_2−xV_xO₄ ceramics with various contents of V ions, a temperature dependence of resistivity, b V concentration dependence of ρ₂₅ and B_25/85

where ρ is resistivity at temperature T (in Kelvin), ρ₀ is a factor related to the material characteristics, k is the Boltzmann constant, E_a is the activation energy of conduction, and B is a material constant representing the temperature sensitivity. B value can be calculated by Eq. (2).

$$B = \frac{{\ln \rho_{1} - \ln \rho_{2} }}{{1/T_{1} - 1/T_{2} }}$$

(2)

where ρ₁ and ρ₂ are resistivities at T₁ and T₂, respectively. T₁ and T₂ are often chosen as 298 K (25 °C) and 358 K (85 °C). So the B value is usually written as B_25/85.

Figure 3b shows ρ₂₅ and B as a function of V content. The detailed data are listed in Table 2. All B are more than 3000 K. Doping of V⁵⁺ ions significantly reduced ρ₂₅. This should result from the doping effect of semiconductor for that the substitution of V⁵⁺ ions introduced electrons. The related defect reaction can be described as Eq. (3).

$${\text{V}}_{2} {\text{O}}_{5} \mathop{\longrightarrow}\limits^{{{\text{ZnFe}}_{2} {\text{O}}_{4} }}2V_{{{\text{Fe}}}}^{ \cdot \cdot } + 4{\text{e}}^{\prime} + 3{\text{O}}_{{\text{O}}} + {\text{O}}_{2}$$

(3)

Table 2 Resistivities (ρ₂₅), temperature sensitivity (B value) and activation energy (E_a) of Zn_1−yNi_yFe_2−xV_xO₄ ceramics doped with various contents of V and Ni ions

The introduced electrons act as weakly bound electrons and locate at the donor level, and can be thermally activated to jump to the conduction band, resulting in increase of conductivity. In the meanwhile, doping with V ions, B values increased. For the ionic substitution, the different radii between V and Fe ions resulted in lattice distortion and increase of lattice potential. The carrier migration barrier is increased, resulting in higher E_a and B.

Figure 4a shows the resistivity as a function of temperature for V/Ni co-doped ZnFe₂O₄ (Zn_1-yNi_yFe_1.96V_0.04O₄) ceramics with different Ni content. The ceramics show typical NTC characteristics. The resistivity increases with the increase of Ni content. When the Ni contents (y) are 0, 0.05, 0.15 and 0.25, ρ₂₅ are 37.1, 744.6, 3050.3 and 4800.4 kΩ·cm, respectively. Figure 4b and Table 2 show the ρ₂₅ and B dependence of Ni-ion concentration in Zn_1-yNi_yFe_1.96V_0.04O₄ ceramics. It can be seen that Ni substitution increased ρ₂₅ and B. All B values are higher than 3000 K and are suitable for practical applications (2000 – 5000 K) [1].

Fig. 4

Electrical properties of Zn_1−yNi_yFe_1.96V_0.04O₄ ceramics with various contents of Ni ions, a temperature dependence of resistivity, b Ni concentration dependence of ρ₂₅ and B_25/85

Figure 5 shows the complex impedance spectra in Cole–Cole plots of A0, A4 and S2A4 ceramics measured at room temperature. Where, the impedances of both real part and imaginary part were calculated taking account into the sample size, so the unit of impedance is kΩ·cm. An equivalent circuit shown in the inset is used to fit the impedance data. Among them, R_g and R_gb represent the impedance of grain boundary effect and grain boundary effect, respectively. CPE1 and CPE2 represent the constant phase elements of grain effect and grain boundary effect, respectively, and are explained as the deviation of capacitance from ideal behavior. Fitted with the equivalent circuit, the resistivities of grain effect and grain boundary effect of each sample are shown in Table 3. Doping of V ion reduced the resistivities of both grain effect and grain boundary effect in the ceramics, while, Ni-doping increased the resistance of both grain effect and grain boundary effect. These also indicate that V and Ni ions have substituted into the lattice, although impurity phase of NiV₂O₆ in S3A4 ceramic were detected as shown in Fig. 1. The impurities might locate at grain boundary and result in the resistance increase.

Fig. 5

Complex impedance spectra in Nyquist plots of the as-sintered A0, A4 and S2A4 ceramics recorded at room temperature

Table 3 Resistivities of grain effect (ρ_g), grain boundary effect (ρ_gb) and total one (ρ_g + ρ_gb) of A0, A4 and S2A4 ceramics fitted using the equivalent circuit as a shown inset in Fig. 5

To determine the effect of cationic valences on the electrical properties of the ceramics, XPS spectra were analyzed as shown in Fig. 6. The full XPS spectra are shown in Fig. 6a. The characteristic peaks of Zn, Fe, O and Ni, V can be observed. The XPS spectra were analyzed in detail by Gaussian Lorentz curve fitting. Figure 6b shows V 2p orbital energy peaks of A4 and S2A4 samples. The XPS peak of V 2p_3/2 is composed of two splitting peaks (516.27 eV and 517.21 eV), which correspond to V⁴⁺ and V⁵⁺ ions [25]. The ratios of XPS peaks of V⁴⁺/V⁵⁺ are 1.34:1 and 1.32:1 in A4 and S2A4 ceramics, respectively. This indicates that the introduction of Ni ions has almost no effect on the concentrations of V⁴⁺ and V⁵⁺ ions.

Fig. 6

XPS analysis of Zn_1−yNi_yFe_2−xV_xO₄ ceramics, a full spectrum, b V 2p, c Fe 2p, d Ni 2p

The Fe 2p spectra of A0, A4 and S2A4 samples are shown in Fig. 6c. Each 2p_3/2 orbital can be fitted by two characteristic peaks, indicating that there are two valence states of Fe ion (Fe²⁺ and Fe³⁺) in each sample. The splitting peak positions are at 710.15 eV and 711.46 eV, corresponding to Fe²⁺ and Fe³⁺, respectively [26]. The ratios of XPS peaks of Fe²⁺/Fe³⁺ in A0, A4 and S2A4 ceramics are 0.54:1, 0.95:1 and 0.69:1, respectively. The existence of high content of Fe²⁺ ions in the A0 (ZnFe₂O₄) ceramic might originate from the processes such as calcining and sintering under an insufficient oxygen atmosphere, leading to the formation of oxygen vacancies. The defect reaction can be described as Eq. (4).

$${\text{ZnFe}}_{2} {\text{O}}_{4} \to {\text{Zn}}_{{{\text{Zn}}}} + 2{\text{F}}e^{\prime}_{{{\text{Fe}}}} + {\text{V}}_{{\text{O}}}^{ \cdot \cdot } + 3{\text{O}}_{{\text{O}}} + \frac{1}{2}{\text{O}}_{2}$$

(4)

So the formula of ZnFe₂O₄ could be expressed as ${\text{ZnFe}}_{\varepsilon }^{2+}{\text{Fe}}_{2-\varepsilon }^{3+}{\text{O}}_{4-\frac{\varepsilon }{2}}$, where ε is the quantity of oxygen vacancies. The occurrence of Fe²⁺ ions enhanced the conductivity for the polaron hopping model [22, 23]. These should be the main reason for the conduction of undoped ZnFe₂O₄ ceramic as shown in Fig. 3. The introduction of V ions (V⁴⁺ and V⁵⁺) induced the increase of Fe²⁺ ion concentration for compensating electric neutrality of ceramic. The increase of Fe²⁺/Fe³⁺ ionic ratio induced by V ions provided high polaron hopping chance between Fe²⁺ and Fe³⁺ ions, combining with V-induced production of weakly bound electrons as discussed in Eq. (3). So the conductivity was enhanced by doping with V ions as shown in Fig. 3.

Figure 6d shows the Ni 2p narrow XPS spectrum in S2A4 ceramic. The spectrum consists of a satellite peak and Ni 2p_3/2 peak. The Ni 2p_3/2 peak was fitted to be composed of two splitting peaks with binding energies of 854.50 eV and 856.20 eV, which correspond to Ni²⁺ and Ni³⁺ valence states, respectively [27]. The transition from Ni²⁺ to Ni³⁺ ions should be accompanied by the formation of holes. At the same time, for compensating electric neutrality, the content of Fe²⁺ ions was reduced comparing with A4 ceramic.

3.3 Aging properties

The resistance drift rates (ΔR/R₀) of Zn_1-yNi_yFe_2-xV_xO₄ ceramics (x = 0.01, 002, 0.03, 0.04, 0.05; y = 0.05, 015, 0.25) after being aged at 150 °C for 500 h are shown in Fig. 7. In the initial stage, the resistance increased with the increase of aging time and tended to be stable after 200 h. The resistance drift rate of the A4 sample reached about 6.20% and the one A0 sample is 4.68% after 500 h aging treatment, while the resistance drift rate of S2A4 sample is the smallest one, about 0.19%. It is obvious that Ni-doping may improve obviously electrical stability.

Fig. 7

Aging induced resistance drift rates of samples A0, A4, S1A4, S2A4 and S3A4

XPS was used to analyze the ionic valence states of the aged ceramics. Figure 8a shows XPS comparison analysis of V ions in A4 and S2A4 samples after aging. Through fitting and calculation, the V⁴⁺/V⁵⁺ peak ratios of A4 and S2A4 ceramics are 1.43:1 and 1.36:1, respectively. Compared to the results of the samples before aging as shown in Fig. 6b, the concentration of V⁵⁺ ions decreased after aging treatment. The change quantity of V⁵⁺ concentration in S2A4 ceramic is smaller than that in A4 ceramic.

Fig. 8

Comparison of XPS analysis, a V 2p peaks of A4 and S2A4 ceramics after 500 h aging, b Ni 2p peaks of S2A4 ceramics before and after 500 h aging, c Fe 2p peaks of A4 ceramics before and after 500 h aging, d Fe 2p of S2A4 ceramics before and after 500 h aging

The valence states of Ni ions of S2A4 before and after 500 h aging were analyzed as shown in Fig. 8b. Each Ni 2p_3/2 peak is composed of two splitting peaks with the binding energies of 854.50 eV and 856.20 eV, respectively. These indicate that Ni²⁺ and Ni³⁺ ions co-exist in V/Ni-doped ZnFe₂O₄-based ceramic. The content ratios of Ni²⁺/Ni³⁺ were calculated to be 5.96:1 and 4.87:1 before and after aging, respectively. After aging, the quantity of Ni²⁺ ions decreased and Ni³⁺ ions increased.

The valence states of Fe ions in A4 and S2A4 samples before and after aging were analyzed as shown in Fig. 8c and d. The concentration ratios of Fe²⁺ to Fe³⁺ in A4 and S2A4 samples were calculated, respectively, to be 0.95:1 and 0.69:1 before aging, while the ratios changed to be 0.82:1 and 0.61:1 after aging. These indicate that Ni-doping in the ZnFe₂O₄-based ceramics enhanced the valent stability of Fe ions.

According to the XPS analysis, the introduction of Ni-ion contributed to the stability of cationic valence. This might result from the electrovalence compensating and rearrangement of ion occupying. Fe, V and Ni are transition metals, and their valent states may change when they are exposed to oxidation or reduction atmosphere, or surface adsorption of gas. As reported by Gao et al. [28], cationic oxidation should be one of the main contributions for resistance drift during the aging process. According to the reports by Bârsan et al. and Dai et al. [29, 30], oxygen molecules (O₂) in atmospheric environment adhere to crystal surfaces, then the O₂ capture electrons from ceramic surface or interfaces (also be grain boundaries) to turn into the adsorbed oxygen (O_adⁿ⁻, n ≤ 2). When the adsorbed oxygen diffused into the ceramic body, cations may be oxidized. As the results, cations with higher valence produced, e.g., Ni²⁺ was oxidized to Ni³⁺ and Fe²⁺ was oxidized to Fe³⁺. For electrovalence compensating, the concentration of V⁵⁺ decreased as shown in Fig. 8a. On the other hand, the ZnFe₂O₄ itself has an inverse spinel structure, Ni²⁺, Ni³⁺, Fe²⁺ and Fe³⁺ ions may occupy at either oxygen octahedral interstices or oxygen tetrahedral interstices. So cation redistribution might take place during aging, leading to the conductivity change and resistance drift. As analyzed in Fig. 8c and d, the introduction of Ni-ion restrained valence evolution of Fe ions, reduced the extent of cation redistribution, resulting in higher electrical stability during aging.

4 Conclusion

The V/Ni co-modified ZnFe₂O₄ ceramics (ZnFe_2-xV_xO₄, x = 0, 0.01, 0.02, 0.03, 0.04 and 0.05; Zn_1-yNi_yFe_1.96V_0.04O₄, y = 0.05, 0.15 and 0.25) with spinel structure were prepared by solid-state reaction method. By changing the contents of Ni and V ions, the room temperature resistivity ranging between 0.037 MΩ·cm and 4.8 MΩ·cm and NTC material constant of B ranging between 3255 K and 6614 K were obtained. V/Ni co-modifying contributes to higher electrical stability of ZnFe₂O₄-based ceramics during aging. The Zn_0.85Ni_0.15Fe_1.96V_0.04O₄ ceramic showed the highest electrical stability with a resistance drift rate of 0.19% after being aged at 150 °C for 500 h in air, while the ZnFe₂O₄ and ZnFe_1.96V_0.04O₄ sample have 4.68% and 6.20% of resistance drift rates. The introduction of Ni-ion reduced the oxidation extent of Fe ions and is proposed to restrain the extent of cation redistribution, resulting in higher electrical stability during aging. It shows that Zn_1-yNi_yFe_2-xV_xO₄ spinel system has great potential as a candidate for applications for NTC thermistors.