WorksPhysics and Astronomy
Sign change in c-axis thermal expansion and lattice collapse by Ni substitution in Co1-xNixZr2 superconductors
Abstract
We investigated the structural, electronic, and superconducting properties of Co1-xNixZr2, in which c-axis thermal expansion is systematically controlled. At x (smaller than) 0.3, c-axis negative thermal expansion (NTE) was observed, and the thermal expansion constant αc approached zero with increasing x. At x = 0.4-0.6, zero thermal expansion was observed, and positive thermal expansion (PTE) appeared for x (greater than) 0.7. By analyzing the c/a ratio, we observed a possible collapsed transition in the tetragonal lattice at around x = 0.6-0.8. The lattice collapse results in c-axis PTE and the suppression of bulk superconductivity.
Impact Statement
Systematic Ni substitution in the CoNiZr superconductor achieves sign change in -axis thermal expansion constant and collapsed tetragonal phases; those structural changes are linked to suppression of superconductivity.
Introduction
Thermal expansions are structural properties of materials. In the case of normal (positive) thermal expansion (PTE), an axis and/or volume expand with increasing temperature. In contrast, materials with negative thermal expansion (NTE), those contract with increasing temperature. Importantly, zero thermal expansion (ZTE) can be achieved by fabricating a composite using PTE and NTE materials, and the ZTE materials have been used in various structural materials and devices in which ultraprecision of positions is required [1–5]. However, achievement of ZTE in a single material is quite rare [6] but has potential merits for development of ZTE application. Development of ZTE in a superconductor is particularly interesting because it will be available in superconducting devices like Josephson junctions with a strength to temperature cycle.
Recently, we reported anomalous axis thermal expansion in CuAl-type (tetragonal) transition-metal zirconide superconductors TrZr (Tr: transition metal) [7–9]. In CoZr, for example, the -axis shows NTE in a wide temperature range, while the -axis exhibits PTE. Owing to the contrasting axis thermal expansion, CoZr and similar TrZr show volume ZTE in a limited temperature range. In addition, we revealed that the axis ratio is the potential factor for switching the character of the -axis expansion [9]. In this study, we focus on CoZr and NiZr with a large and small ratio, respectively. CoZr exhibits a -axis NTE and is a superconductor with a transition temperature () of ~6 K [7,10,11]. NiZr exhibits PTE in both and axes. In previous works [12–14], synthesis and physical properties of a solid solution system of CoNiZr were reported with its superconducting properties. Here, we show that the -axis thermal expansion character in CoNiZr is systematically changed from NTE, ZTE, and PTE with increasing Ni concentration .
Methods
Polycrystalline samples of CoNiZr ( = 0, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1.0) were synthesized by arc melting in an Ar atmosphere. Powders of pure transition metals (Tr) of Co (99%, Kojundo Kagaku) and Ni (99.9%, Kojundo Kagaku) with a nominal composition were mixed and pelletized. The Tr pellet and plates of pure Zr (99.2%, Nilaco) were used as starting materials. The samples were melted five times and turned over after melting to homogenize the sample.
X-ray diffraction (XRD) patterns were collected by - method with Cu-K radiation on a Miniflex-600 (RIGAKU) diffractometer equipped with a high-resolution semiconductor detector D/tex-Ultra. For High-temperature XRD on a Miniflex-600, the sample temperature was controlled by a BTS 500 attachment. The obtained XRD patterns were refined by the Rietveld method using RIETAN-FP [15], and the schematic images of the crystal structure were depicted using VESTA [16]. The actual compositions of the samples were investigated using energy-dispersive x-ray spectrometry (EDX, Swift-ED, Oxford) on a scanning electron microscope (SEM, TM3030, Hitachi Hightech).
The temperature dependence of magnetization was measured both after zero-field cooling (ZFC) and field cooling (FC) using a superconducting quantum interference device (SQUID) on an MPMS3 (Quantum Design).
The first principles band calculations were performed using the WIEN2k package [17] within the PBE-GGA exchange-correlation functional [18]. The virtual crystal approximation is adopted to take into account the effect of the elemental substitution of Ni for Co. We used the experimentally determined lattice parameters shown in Table 1. The atomic coordinates of Zr were theoretically optimized. and the -mesh were set to 8 and , respectively.
| Nominal | (EDX) | (Å) at 303 K | (Å) at 303 K | at 303 K | (μK) | (μK) | (μK) | (K) |
| 0 | 0 | 6.360(3) | 5.514(3) | 0.8670(6) | 25.6(6) | -20(1) | 31(1) | 5.9 |
| 0.1 | 0.111(3) | 6.375(3) | 5.475(3) | 0.8589(6) | 20.2(8) | -14(1) | 27(2) | 6.4 |
| 0.2 | 0.216(6) | 6.3750(9) | 5.428(1) | 0.8514(2) | 21.1(6) | -4.6(9) | 37(1) | 6.1 |
| 0.3 | 0.312(2) | 6.377(3) | 5.391(2) | 0.8454(5) | 14.9(6) | -3(1) | 27(1) | 5.1 |
| 0.4 | 0.417(3) | 6.415(1) | 5.384(2) | 0.8393(3) | 14.0(7) | -1(1) | 27(2) | 4.1 |
| 0.5 | 0.544(6) | 6.425(1) | 5.350(1) | 0.8327(3) | 15.1(5) | 1.4(9) | 32(1) | 3.2 |
| 0.6 | 0.635(5) | 6.450(2) | 5.336(3) | 0.8273(8) | 23.6(7) | 2(2) | 48(2) | 2.5 |
| 0.7 | 0.752(4) | 6.469(3) | 5.288(3) | 0.8175(6) | 14.4(9) | 15(2) | 43(3) | (2.4) |
| 0.8 | 0.824(2) | 6.481(3) | 5.279(3) | 0.8145(6) | 7(1) | 11(2) | 25(3) | (2.5) |
| 0.9 | 0.937(3) | 6.491(4) | 5.261(3) | 0.8105(7) | 10.5(6) | 12(1) | 33(2) | (2.5) |
| 1 | 1 | 6.509(6) | 5.259(4) | 0.8081(9) | 15.8(5) | 17(1) | 49(1) | (2) |
Table 1. Results of chemical analyses, evolutions of lattice constants and thermal expansion constants, and in examined CoNiZr. with bracket indicates filamentary superconductivity.
Results and Discussion
Crystal structure analysis and axis thermal expansion
The obtained actual compositions at the site are comparable to the nominal values and summarized in Table 1. Figure 1(a) shows the schematic images of crystal structure of . Figure S1-Figure S1 (supporting materials) are XRD patterns for –. These compounds have a tetragonal -type structure (), and the main peaks could be indexed with the structural model. Small impurity peaks of the orthorhombic phase are seen as indicated by asterisks as reported in Ref. 7, and the amount of the impurity decreased with increasing Ni concentration (). We estimated lattice constants by Rietveld refinements, and the obtained parameters are plotted in Figure 1(b) and summarized in Table 1. The obtained trend of lattice constants is consistent with a previous study [14].

Figure 1. (a) Schematic images of the crystal structure of . (b) Ni concentration () dependence of lattice constants and at 303 K. The error bars are standard deviations estimated by the Rietveld refinement.

Figure S1. XRD patterns of CoNiZr for = (a) 0–0.3, (b) 0.4–0.6, (c) 0.7–1.0. Asterisks indicate the impure phases of .
We investigated the thermal expansion properties by high-temperature XRD. As reported in Ref. 7, exhibits -axis NTE, while the -axis exhibits PTE. To investigate how the anisotropic axis thermal expansion changes with Ni doping, we collected the data for all the samples between 303-573 K with a temperature increment of about 30 K and used it for the calculation of the linear thermal expansion coefficients along the -axis () and -axis () and the volumetric expansion coefficient (). The -axis NTE was observed for , 0.1, 0.2, 0.3; hence, we show the results for as an example. Figure S2-Figure S2 (supporting materials) show the temperature dependence of lattice constants , , and volume () for the sample.

Figure S2. (a, b, c) The temperature dependence of lattice constants , , and for sample. The estimated linear thermal expansion constants are , . The volumetric expansion coefficient is . (d) High-temperature XRD patterns on , , , K for sample. The dashed line in this graph indicates that the 002 peak shifts to the higher angle side.
Figure S2 (supporting materials) shows the typical high-temperature XRD patterns. The 002 peak shifts to the higher angle side with increasing temperature, indicating that the sample still contracts along the -axis upon heating. The estimated values of , , and using the formulas , and are , , , respectively The magnitude of for is smaller than that of CoZr, which also exhibits NTE along the -axis with [7]. This suggests that the substitution of Ni for the Co site suppresses the NTE along the -axis, and the switching between PTE and NTE is controlled by adjusting the value. Figure 2(a)-(Figure 2i) show the temperature dependence of the normalized rate of change in the lattice constants , , and from 303 K for all the samples. For all , the lattice constant and gradually increase with heating. The samples with a lower Ni amount (–) show NTE along the -axis as shown in Figure 2(d). On the other hand, the samples with a larger Ni amount (–) show PTE along the -axis as shown in Figure 2(f). For the samples with medium Ni amount (–), ZTE trend was observed as shown in Figure 2(e). Figure 3 shows the dependence of the linear thermal expansion coefficient along the -axis, which shows a successful control of the switching of NTE and PTE along the -axis by tuning . The turning point for the NTE and PTE is estimated between and 0.6. Therefore, there is a possibility to synthesize the sample which exhibits the perfect ZTE along the -axis by optimizing the Ni amount doped at the Co site. As well, materials that exhibit anisotropic thermal expansion have been reported, such as -Eucryptite (LiAlSiO) [19], Ag[Co(CN)] [20], and CaRuO [21]. The mechanisms of NTE are diverse [1,22]. For example, the cause of the NTE on the monoclinic CaRuO is orbital ordering and disordering [21]. Not only electronic contributions but also structural properties contribute to the NTE mechanisms. In another study on -(CuZn)VO, the chemical substitution of Cu by Zr decreases the free space for the transverse vibrations, which suppresses NTE along the -axis [23]. Furthermore, MnCuGeN exhibits giant negative thermal expansion due to the local lattice distortion triggered by Ge dope [24]. These facts will help us to understand the mechanisms of the anomalous (anisotropic) -axis thermal expansion in the current system. Recently, we reported that the NTE along the -axis for TrZr was caused by the robust -Zr distance to the temperature change and the flexible bonding of the polyhedron units. In addition, the ratio is found to be an essential parameter that determines the polyhedron shape and the thermal expansion characteristics [7,9]. Therefore, further studies on electronic and/or orbital characteristics and local structures of will be striking in determination of the mechanisms of the emergence of -axis NTE in .

Figure 2. (a-i) Temperature dependence of the normalized rate of change of the lattice constants , , and from 303 K.

Figure 3. The dependence of for .
Superconducting properties
A superconducting transition was observed for all samples, as shown in Figure 4a. The large diamagnetic signals observed for suggests the emergence of bulk superconductivity. In contrast, the signals for are quite small, which indicates that the observed diamagnetic signals are caused by filamentary (trace) superconductivity states in those samples. The tends to decrease with increasing ; however, the samples with and 0.2 have slightly higher than that for . Figure 4b shows the enlarged view near the for , 0.1, 0.2, 0.3, 0.4. The highest of 6.39 K was observed for . To discuss about the electronic origins on this behavior, we performed first-principles calculations for . Figure 5(a) shows the dependence of density of states near Fermi level, . The dependence of the calculated looks consistent with the evolution of if we assumed conventional phonon-mediated superconductivity [25], because a large achieves a higher in a conventional superconductor. Figure 5(b) shows the dependence of . As we mentioned above, the evolution of at is consistent to the behavior where the value is smaller than . However, we cannot explain the change in with behavior where is larger than . As mentioned above, the samples with exhibit filamentary superconductivity; in Figure 5(b), we indicated the boundary between bulk superconductivity (Bulk SC) and filamentary superconductivity (Filamentary SC). According the discussion above, we suggest that the bulk superconductivity observed for is positively linked to DOS(), which would suggest the importance of phonon in the superconductivity mechanism in .

Figure 5. (a) The dependence of DOS() for CoNiZr. (b) The dependence of superconducting transition temperature for CoNiZr. (c) The dependence of for CoNiZr at 303 K. The estimated slope values for – and – are , respectively.
For , bulk superconductivity is suppressed, while the DOS() is comparable or higher than . To explore possible cause of the suppression of superconductivity, we estimated the ratio of using the data at 303 K and plotted in Figure 5(c) as a function of . Although linearly decreases with increasing for , the slope clearly changes at around –. For –, another slope can guide the evolution of . We propose that the change in the ratio is a kind of transition to collapsed tetragonal phases as observed in iron-based superconductors and and related layered compound [26-30]. The electronic structure is generally affected by a collapsed tetragonal transition, which affects superconductivity as well [31,32], we assume that the disappearance of bulk superconductivity by Ni heavy doping is related to the collapsed transition. In our previous work, we suggested the trend that a higher of is achieved with increasing lattice constant [33]. This fact is also consistent with the above scenario because the -axis is largely compressed at around . To obtain further evidence on the collapsed transition and its relation to electronic structure, superconductivity, and axis thermal expansion, further investigations with different probes are needed.

(a)

(b)
Figure 4. (a) Temperature dependence of ZFC susceptibility for . (b) Enlarged view near the superconducting transition temperature for , , , , .
Conclusion
We investigated the crystal structure, axis thermal expansion, electronic structure, and superconducting properties of transition-metal zirconide superconductor CoNiZr. The samples were synthesized by arc melting and characterized by powder XRD and EDX. At , -axis NTE was observed, and the thermal expansion constant () approached zero with increasing . At –, -axis thermal expansion close to ZTE was observed, and PTE appeared for . Those results confirm that the -axis NTE can be controlled by Ni substitution (tuning ratio) and switched to PTE. On the superconducting properties, we observed bulk superconductivity for , and bulk nature of superconductivity is suppressed by Ni heavy doping. For , the evolution of the electronic DOS() well explains the change in , but it cannot explain the disappearance of bulk superconductivity at . By analyzing the ratio, we revealed a possible transition to collapsed tetragonal phases with a boundary concentration of – by Ni heavy doping. The lattice collapse would affect electronic structure and be negatively linked to superconductivity in .
In addition, the lattice collapse seems to be linked to the appearance of -axis PTE. Since superconductivity in would be mediated by phonon, the correlation between axis thermal expansion, emergence of superconductivity, and the lattice collapse transition is one of the notable features of this system. Thus, is a suitable platform to study anomalous axis thermal expansion and the method to systematically control the thermal expansion. Furthermore, the relationship between lattice collapse and/or anomalous axis thermal expansion and emergence of superconductivity would provide us with new strategy on exploration of new superconductors.
Acknowledgements
The authors thank Md. R. Kasem, A. Yamashita, O. Miura for supports in experiments and D. Louca for fruitful discussion on local structures. The work has been partly supported by JSPS KAKENHI (Grant Number: 21H00151, 21K18834), Tokyo Metropolitan Government Advanced Research (Grant Number: H31-1), TMU strategic fund for young scientists, and TMU strategic fund for multidisciplinary collaboration.
Author statements
All the authors declare no competing interests.
Graphical Abstract

Figure 6 (PDF p. 14). CoNiZr Tetragonal (); Collapsed tetragonal; PTE; NTE; ( K); in CoNiZr; (K); Bulk SC; Filamentary SC; CoZr; NiZr.
Supporting materials
The XRD patterns for -1.0
We measured XRD patterns for all samples. Main peaks could be indexed with a tetragonal CuAl-type structure (), however, some impurity peaks developed as shown in asterisks. These impurity peaks are indicated with an orthorhombic (space group #63) phase [1].
Thermal expansion results for
We observed the thermal expansion for all samples, here, we show the results of . This sample exhibits -axis, volume PTE and -axis NTE as shown in Supplementary Figure 2(a)-2(c). Supplementary Figure 2(d) shows High-temperature XRD patterns and the gradual shift of the 002 peak to the higher angle side. This behavior is consistent with the result of -axis contract.
References
- [1][S1] Mizuguchi Y, Kasem Md R, Ikeda Y. Anomalous Thermal Expansion in a CuAl₂-type Superconductor CoZr₂. J Phys Soc Jpn. 2022;91:103601.DOI