1.   Introduction

The metal-to-insulator transition (MIT) achieved in 3d-correlated transitional metal oxides prompts abrupt variations in the electronic and/or optical properties beyond conventional semiconductors, enabling applications such as critical temperature resistance thermistors, correlated logical devices, and thermochromism [1–6]. Among the MIT material family, rare-earth nickelates (RENiO₃) exhibit an exceptionally complex electronic phase diagram and multiple electronic phase transitions. For instance, a reversible charge disproportionation (or antidisproportionation) related to the Ni³⁺ ↔ Ni^(3±δ)+ occurs across a critical temperature (T_MIT) that prompts the conventional MIT properties of RENiO₃. An overwhelming advantage in the MIT of RENiO₃ is the wide and continuous tunability in T_MIT (e.g., within 100–600 K) by simply adjusting its rare-earth composition. For example, reducing the rare-earth ionic radius (r_RE) within RENiO₃ weakens the orbital overlapping between Ni 3d and O 2p, which increases the relative stability of the insulating phase compared with the metallic phase and thus increases T_MIT [7]. Aside from the conventional MIT, the 3d orbital configuration of Ni can be further switched among Ni³⁺ (${\mathrm{t}}_{2\mathrm{g}}^{6}{\mathrm{e}}_{\mathrm{g}}^{1}$), Ni²⁺ (${\mathrm{t}}_{2\mathrm{g}}^{6}{\mathrm{e}}_{\mathrm{g}}^{2}$) and Ni¹⁺ (${\mathrm{t}}_{2\mathrm{g}}^{6}{\mathrm{e}}_{\mathrm{g}}^{3}$) via a hydrogen (or lithium)-related Motronic process. Direct manipulation of the orbital occupancy leads to the further discovery of novel electronic phases within RENiO₃, such as the electronically highly correlated phase associated with Ni²⁺ and the nickelate superconductor corresponding to Ni⁺ [8]. The abundant electronic states of RENiO₃ provide a new paradigm for further exploring emerging electronic applications such as proton-doped memory devices [9–11], synaptic devices [12–14], and reconfigurable perovskite devices [15].

Nevertheless, the present application of RENiO₃ merely focuses on a single device pixel at a destined rare-earth composition, while their further array application via integrating RENiO₃ with different rare-earth compositions has not yet been fully explored. It is worth noting that precise control in the ground electronic phase of RENiO₃ can be realized by adjusting the co-occupation ratio of two rare-earth elements, which determines the average magnitude of r_RE. This was previously shown by the linear variation in the T_MIT of Nd_xSm_1−xNiO₃ with the rare-earth co-occupation atomic ratio x. Thus, establishing arrays of multiple RENiO₃ pixels offers new freedom from the perspective of their electronic permutation and combination in further extending their correlated electronic applications.

In this work, a method was proposed to achieve precise temperature measurement using the negative temperature coefficient resistance (NTCR) of RENiO₃ within the MIT region. Compared with a conventional NTCR thermistor, RENiO₃ displays a larger temperature coefficient resistance (TCR) due to the abrupt transition from an insulator to a metal when reaching a critical temperature. Thus, it is anticipated that the use of RENiO₃ for temperature measurement would give a higher accuracy and sensitivity. However, hysteresis exists in the resistance curves as a function of temperature for some RENiO₃ with a single rare-earth composition (e.g., NdNiO₃). As a consequence, we selected and prepared Nd_xSm_1−xNiO₃ with different Nd/Sm atomic ratios (x = 0.2, 0.33, 0.4, 0.5, 0.6, 0.67, 0.75, 0.8, and 1). Their temperature dependence on resistance (R–T) exhibits sharp MIT variation across a temperature range of 200–400 K with no hysteresis except for x =1. In addition, for Nd_xSm_1−xNiO₃ with various x, their total resistance after paralleling an optimum constant resistor indeed varied linearly with temperature within a specific temperature range. In addition, the linear coefficient of total resistance with temperature was much larger in Nd_xSm_1−xNiO₃ than that in conventional NTCR thermistors because of larger magnitudes of NTCR that originate from their sharp MIT and is unachievable in conventional NTCR thermistor materials. Thus, it can be expected that integrating an array of Nd_xSm_1−xNiO₃ with different x values can be applied to measure the temperature with high sensitivity and precision in the future.

2.   Experimental

2.1   Sample preparation

Nd_xSm_1−xNiO₃ powders with x = 0, 0.2, 0.33, 0.4, 0.5, 0.6, 0.67, 0.75, 0.8, and 1 were prepared using the KCl molten salt–assisted high oxygen pressure reaction approach reported in our previous work [16]. Typically, rare-earth oxides (Nd₂O₃, Sm₂O₃), NiO, and KCl were mixed and ground at a molar ratio of 1:2:2 in an agate mortar. The obtained mixture was transferred into a quartz tube and then calcined at 800°C and 10 MPa oxygen pressure for 24 h. Next, the obtained powder was rinsed with deionized water and ethanol to remove KCl and residual water. Finally, the powder was pressed into pellets and further sintered at 800°C and 10 MPa oxygen pressure for 24 h.

2.2   Characterization

The crystal structures of the as-prepared Nd_xSm_1−xNiO₃ powders were verified by X-ray diffraction (XRD, Rigaku SmartLab). The lattice constants of Nd_xSm_1−xNiO₃ were measured by Rietveld refinement using FullProf software. The morphologies of the samples were characterized by scanning electron microscopy (SEM) with a JEOL JSM 6510A electron microscope. Near-edge X-ray absorption fine structure (NEXAFS) was conducted at the BSRF-4b9b beamline of the Beijing Synchrotron Radiation Facility, Institute of High Energy Physics, Chinese Academy of Sciences to reflect the electronic structures of Nd_xSm_1−xNiO₃. The resistance as a function of temperature for Nd_xSm_1−xNiO₃ was determined from 5 to 300 K using a physical property measurement system (Quantum Design Inc., China) and above 300 K using a commercialized CTA system.

3.   Results and discussion

3.1   Initial concepts

As illustrated in Fig. 1(a), Nd_xSm_1−xNiO₃ shows a distorted perovskite and continuously tunable T_MIT by simply adjusting the Nd/Sm ratios (x from 1 to 0) within 200–400 K. It is worth noting that the resistivity of RENiO₃ exhibits an abrupt drop across the transition temperature T_MIT by 1–3 order within a wide temperature range compared with the conventional NTCR thermistor [17–19], as shown in Fig. 1(b). Moreover, another characteristic of RENiO₃ with middle and heavy rare-earth composition is that there is nearly no thermally induced hysteresis in their MIT when measuring the temperature dependence of resistivity by heating up and cooling down. Thus, compared with the conventional NTCR thermistor, the larger decrease of resistance in the MIT region from Fig. 1(b) would result in higher TCR, where the resistance of RENiO₃ is more sensitive to temperature changes. Hence, it is expected that using RENiO₃ for temperature measurements will lead to higher accuracy and sensitivity. However, it is worth mentioning that the nonlinear variation of resistance with the temperature of RENiO₃ in the MIT region would be unfavorable for actual temperature measurements. To solve this problem, an optimum constant resistor was paralleled to RENiO₃, which causes the total resistance to show a linear change within a certain temperature range. Then, it can be utilized to determine the temperature using the typical electric circuit diagram shown in Fig. 1(c). The circuit (red box in Fig. 1(c)) is a constant voltage source that gives a stable voltage for the bridge voltage sampling part (blue box). The bridge voltage sampling part is employed to obtain the total resistance of rare-earth nickelates and R_p in parallel. Once the total resistance is probed, the temperature can be determined on the basis of the assumption that the total resistance varies linearly with temperature. Afterward, the voltage amplification part (purple box) magnifies the weak voltage signals between V1 and V2. Also, it would be better to integrate an array containing a series of RENiO₃ with different rare-earth compositions to realize precise temperature measurements over a wide temperature range.

Figure  1.  (a) Schematic of the crystal structures of RENiO₃ with different Nd/Sm ratios. (b) Comparison of resistance versus temperature for RENiO₃ and conventional NTCR. (c) Possible electric circuit for temperature measurement using linearized RENiO₃ (VCC—Volt current condenser; GND—ground; TEMP-AD—temperature analog-to-digital converter; V1 and V2—voltage 1 and voltage 2; R_p and ${\boldsymbol{R}}_{\mathbf{R}\mathbf{E}\mathbf{N}\mathbf{i}\mathbf{O}_{\boldsymbol{3}}}$—parallel resistance and resistance of RENiO₃).

下载: 全尺寸图片 幻灯片

3.2   Material structure, morphology, and electronic structure

To achieve the above concept, a series of Nd_xSm_1−xNiO₃ powders with different rare-earth compositions (x = 0, 0.2, 0.33, 0.4, 0.5, 0.6, 0.67, 0.75, 0.8, and 1) were synthesized and then pressed into pellets (more details in the Section 2). Fig. 2(a) displays the XRD patterns of the as-synthesized Nd_xSm_1−xNiO₃ samples. Almost all peaks for each Nd_xSm_1−xNiO₃ sample are in accordance with the standard PDF card (PDF#51-0391), demonstrating a typical distorted perovskite structure. As a typical example, the XRD spectrum of Nd_0.8Sm_0.2NiO₃ was refined using the Rietveld method, as shown in Fig. 2(b) (see more Rietveld refinements for other Nd_xSm_1−xNiO₃ in Fig. S1). Some parameters taken from the Rietveld refinements are summarized in Table S1. R_P and χ² factors are less than 5.6% and 3.5%, respectively, for Nd_xSm_1−xNiO₃, which shows the effectiveness of the Rietveld refinements. In addition, the lattice constants of the c-axis (c) obtained from the Rietveld refinements of the as-made Nd_xSm_1−xNiO₃ samples increase with the improvement of the Nd content, as shown in Fig. 2(c). These results align with the fact that SmNiO₃ has a smaller lattice parameter than NdNiO₃ [20].

Figure  2.  (a) XRD patterns of the as-prepared Nd_xSm_1−xNiO₃ powders (PDF#51-0391) covered with different rare-earth compositions (x = 0.2, 0.25, 0.33, 0.5, 0.67, 0.75, and 0.8). (b) Rietveld refinement patterns obtained from the measured XRD spectra for Nd_0.8Sm_0.2NiO₃ (Yobs and Ycalc are measured XRD pattern and rietveld refinement pattern, respectively). The red lines indicate the obtained XRD data, the black lines indicate the calculated pattern, the blue lines refer to the observed calculated pattern, and the green vertical ticks indicate the positions of the Bragg reflections. (c) Variation of the c-axis lattice parameters of the as-prepared Nd_xSm_1−xNiO₃ powder with increasing Nd content, which was obtained from the Rietveld refinement of the X-ray diffraction spectra.

下载: 全尺寸图片 幻灯片

The morphologies of the Nd_xSm_1−xNiO₃ samples were further observed by using SEM. As a representative example, Fig. 3(a) and (b) shows the morphologies of the as-synthesized Nd_0.67Sm_0.33NiO₃ powder and the corresponding cross-section, respectively. The morphology of the Nd_0.67Sm_0.33NiO₃ powder in Fig. 3(a) shows a typical cubic shape. The size is about several micrometers, which is close to our previous results [7]. Also, based on the energy dispersive spectroscopy results, the practical composition of the as-synthesized powders agrees with their respective nominal compositions, as shown in Fig. S2 for Nd_0.67Sm_0.33NiO₃ powder. In addition, a compact cross-section was observed for the sintered pellets, as shown in Fig. 3(b). It is worth noting that although grains of different sizes are observed in the cross-section of the pellet, their composition is the same, as shown in Fig. S3. The dense pellet also gives convenience for measuring electrical transportation.

Figure  3.  Scanning electron microscope images of (a) the Nd_0.67Sm_0.33NiO₃ powder and (b) the cross-section of the Nd_0.67Sm_0.33NiO₃ pellet. NEXAFS analysis of the (c) O-K edge and (d) Ni-L₃ edge of Nd_xSm_1−xNiO₃ with x = 1, 0.8, and 0.2. In (c), peak A, also called the prepeak, is associated with the transition from O 1s to the Ni 3d-O 2p hybridized states. Peak B results from the 4f and 5d hybridized states of rare-earth ions with O 2p. Peak C corresponds to the Ni 4s and 4p hybridized with O 2p. In (d), the Ni-L₃ spectrum originates from the Ni 2p to Ni 3d transition and splits into peaks A and B. The proportion of the ${\mathbf{t}}_{{\boldsymbol{2}}\mathbf{g}}^{{\boldsymbol{6}}}{\mathbf{e}}_{\mathbf{g}}^{{\boldsymbol{1}}}$ (Ni³⁺) orbital configurations compared with the ${\mathbf{t}}_{{\boldsymbol{2}}\mathbf{g}}^{{\boldsymbol{6}}}{\mathbf{e}}_{\mathbf{g}}^{{\boldsymbol{2}}}$ (Ni²⁺) configuration is indicated by the relative height of peak B split from peak A in the Ni-L₃ NEXAFS spectrum.

下载: 全尺寸图片 幻灯片

Next, the electronic structures of Nd_xSm_1−xNiO₃ with different rare-earth compositions were analyzed by NEXAFS, and the O-K edge and Ni-L₃ edge spectra are presented in Fig. 3(c) and (d), respectively. Comparing the O-K edge of NdNiO₃ and Nd_0.8Sm_0.2NiO₃ with that of Nd_0.2Sm_0.8NiO₃, an improvement of the prepeak (marked as A) can be observed in Fig. 3(c) when increasing the Nd content. It has been reported in previous literature [21] that the intensity of prepeak A in their O-K edge spectra reflects the overlapping between Ni 3d and O 2p orbitals. Thus, a higher intensity of prepeak A is associated with a larger overlap between Ni 3d and O 2p and Ni–O–Ni bond angle, which stabilizes the metallic phase of RENiO₃ and results in a lower T_MIT. Further consistency can be verified from the variations in their Ni-L₃ spectra (Fig. 3(d)). The relative intensity of peak A compared with that of peak B shows the proportion of the Ni³⁺ orbital configurations [22]. In Fig. 3(d), an increasing intensity of peak A is observed with increasing Nd content. In general, the Ni³⁺ orbital configuration is associated with metallic states, causing a decrease in T_MIT when Nd content is increased [19].

3.3   MIT properties and potential applications for temperature sensing

In addition, the resistance (R) as a function of temperature (T) (R–T) curves for the as-synthesized Nd_xSm_1−xNiO₃ during both heating and cooling processes are presented in Fig. 4(a). All samples show significant MIT behaviors, and the decrease in resistance in the MIT region reaches approximately two orders of magnitude. Furthermore, it can be noted that their electrical transport curves measured by heating and cooling procedures overlap well, except for NdNiO₃, which offers the prerequisite for temperature measurement because the presence of a gap between the heating and cooling curves makes the measurement of temperature inaccurate. Thus, the subsequent discussion will not include NdNiO₃. The decreasing tendency of T_MIT with increasing Nd in Fig. 4(b) proves the correct analysis of the electronic structure of the as-synthesized Nd_xSm_1−xNiO₃. In addition, Fig. 4(b) displays the variations (R_I/R_M)/ΔT (i.e., S/ΔT) with the improvement of Nd percentage, to some degree, indicating the sharpness of MIT. Note that a more abrupt MIT tendency can be found when increasing the Nd proportion, but interestingly, the sharpness of the MIT behavior increases slightly when the portion of Nd continues to decrease to below 30at%.

Figure  4.  (a) R–T curves of rare-earth nickelates Nd_xSm_1−xNiO₃ pellets with different rare-earth compositions (x = 1, 0.8, 0.75, 0.67, 0.6, 0.5, 0.4, 0.33, 0.2, and 0). The data source of x = 1 is taken elsewhere [23]. Solid lines refer to the heating process, while the dashed lines refer to the cooling process. (b) Variation of T_MIT and S/ΔT with the Nd content. Fig. S4 shows the detailed process of calculating using S/ΔT, and S = R_I/R_M, where R_I and R_M stand the resistances of the insulating phase I and the metallic phase M, respectively.

下载: 全尺寸图片 幻灯片

Nevertheless, the nonlinear resistance variation with temperature for Nd_xSm_1−xNiO₃ within the MIT region is not beneficial for measuring the temperature directly and accurately. Thus, a suitable constant resistor is needed in parallel with Nd_xSm_1−xNiO₃ to linearize their R–T curves within a specific temperature range of MIT. Concrete linearization details can be found in Section 3 of the supporting information. To demonstrate the possibility of the linearization method, we selected Nd_0.8Sm_0.2NiO₃ as a representative example to show the linearization process within the MIT region. Fig. 5(a) illustrates the resistance of Nd_0.8Sm_0.2NiO₃ as a function of temperature, displaying significant MIT behavior, as shown in the red box. Combined with the linearization method in the supplementary material, we chose this R–T curve included in the red box to be linearized, and the obtained results are shown in Fig. 5(b). It can be clearly observed that the R–T curve of Nd_0.8Sm_0.2NiO₃ deviates from the linear relationship without paralleling a constant resistor, whereas a linear tendency of resistance with temperature appears with the addition of an optimum constant resistor. This linear tendency proves the validity of the linearization method. However, it is also worth mentioning that the linear variation in the temperature dependence of resistance across the MIT of a single Nd_1−xSm_xNiO₃ composition usually covers an effective temperature range of several tens of Kelvin. Thus, a series of Nd_1−xSm_xNiO₃ (x = 0.2, 0.33, 0.4, 0.5, 0.6, 0.67, 0.75, 0.8) can be connected in parallel as an array to cover the entire temperature range of 210–360 K. By further applying this linear method to Nd_xSm_1−xNiO₃ with other rare-earth compositions, some sensing temperature parameters such as linear slopes, magnitude of the parallel resistor, and linear temperature ranges can be determined [24]. The relationships among these parameters are summarized in Fig. 5(c). Since the magnitudes of slope k are associated with the resistance of Nd_xSm_1−xNiO₃ at the linearization temperature, as elucidated in Section 3 of the supplementary material, we introduced the ratio k/R_T as a novel and scientifically derived parameter to reflect the sensitivity of the total resistance to temperature variation, where R_T is the resistance of Nd_xSm_1−xNiO₃ at the linearization temperature, which is different for each x value. Nevertheless, if the linearization temperature is determined for a certain x value, R_T is a constant. The variation tendency of k/R_T with Nd content is consistent with the evolution of S/ΔT as a function of temperature, as shown in Fig. 4(c). Meanwhile, we also calculated the TCR for the as-synthesized Nd_1−xSm_xNiO₃ shown in Fig. S6. The maximum TCRs of every sample are shown in Fig. 5(d). It can be observed that the variations of TCR and k/R_T as a function of Nd are consistent, which agrees well with our expectation that a larger TCR means a higher temperature sensitivity. Furthermore, the same TCR values are difficult to achieve in a conventional NTCR thermistor. In other words, integrating an array of Nd_xSm_1−xNiO₃ with various x values allows for precise and highly sensitive temperature measurements in the future. The process of temperature measurement can be explained more specifically as follows: the target temperature range can be first detected by the temperature dependence of RENiO₃ with a heavier rare-earth composition (e.g., GdNiO₃) that shows insulator transportation. Subsequently, the system is switched to the specific composition of Nd_xSm_1−xNiO₃ to realize accurate temperature sensing within the low-temperature range.

Figure  5.  (a) Representative temperature dependence of the resistance of Nd_0.8Sm_0.2NiO₃ for linearity in temperature sensing. The red dashed box means that this part of the R–T curve is selected to achieve linearization by paralleling the constant resistor. (b) Before and after paralleling a specific constant resistor, the variations of the total resistance R_t with temperature. The black dashed line indicates that the curve will deviate from the linear relationship without paralleling a constant resistance in the same temperature range. The blue solid line is the linear fitting curve, and k is its slope. R_t is the total resistance of Nd_0.8Sm_0.2NiO₃ or Nd_0.8Sm_0.2NiO₃ parallel to R_p. (c) Linear temperature ranges and k/R_T for various rare-earth nickelates. (d) Temperature coefficient resistance (TCR) maxima and linear temperature range for different rare-earth nickelates.

下载: 全尺寸图片 幻灯片

4.   Conclusion

In conclusion, we propose a strategy to sense low-range temperatures using the MIT properties of array-like RENiO₃, where high magnitudes of TCR can be realized beyond conventional NTCR materials. A series of Nd_xSm_1−xNiO₃ samples with different rare-earth compositions of x = 0–1 were synthesized, while the critical temperature related to the MIT without thermally induced hysteresis was achieved for x = 0–0.8. The linearity in the temperature dependence of resistivity across MIT of Nd_xSm_1−xNiO₃ was achieved by paralleling a constant resistor, while the magnitudes of such parallel resistors compared with Nd_xSm_1−xNiO₃ were presented. The entire temperature range of 210–360 K can be covered by integrating the Nd_xSm_1−xNiO₃ with different Nd/Sm atomic ratios from x = 0.2 to x = 0.8. Compared with conventional NTCR materials, such array-like Nd_xSm_1−xNiO₃ achieves higher magnitudes of TCR (e.g., 7%–14%), and this is anticipated to be of utility in accurate temperature sensing at low-temperature ranges.