The discovery of topological semimetals (TSMs) has attracted broad attention in condensed matter physics and material science because of the topologically protected band structure. TSMs give rise to some exotic physical properties, including giant magnetoresistance (MR), ultrahigh carrier mobility, nontrivial Berry phase, and negative longitudinal MR.^[1–4] According to the degeneracy of the band crossing points and their distribution in the Brillouin zone (BZ), TSMs can classify into Dirac semimetals (DSMs), Weyl semimetals (WSMs), and nodal-line semimetals (NLSMs).^[5] The band crossing points are four-fold degenerate for DSMs, such as Na₃Bi and Cd₃As₂.^[6–8] WSMs exhibit two-fold degenerate band crossing points with definite chirality, and it is located at an even number of discrete points in the BZ, such as TaAs^[9,10] and WTe₂.^[11,12] Different from the discrete points in DSMs and WSMs, the band crossing points form closed loops in the BZ of NLSMs.^[13,14]

Recently, it has been reported that black phosphorus (BP) exhibits a topological phase transition from band semiconductor to zero-gap Dirac semimetal state^[15,16] under hydrostatic pressure. The semimetal phase in BP is characterized by the large MR, the coexistence of electron and hole pockets, low carrier densities, small cyclotron masses, weak localization-weak antilocalization transition at low magnetic fields and the emergence of a nontrivial Berry phase of π detected by Shubnikov-de Haas (SdH) magneto-oscillations. BP has a puckered honeycomb lattice,^[17–19] in which the direct band gap is mainly determined by the out-of-plane p_z-like orbital of P and can be further reduced by tuning the interlayer coupling.^[15,20,21] Interestingly, in CaAs₃, the two-dimensional infinite polyanionic nets are derivatives of the BP structure. While a deeper and systematic study on CaAs₃ single crystals is still missing except for the semiconducting behavior.^[22,23] More recently, CaAs₃ was predicted theoretically to be a candidate for topological insulator (TI) or NLSMs when the spin–orbit coupling (SOC) was taken into consideration or not and the topological transition could be engineered via strain, or chemical doping.^[5,24] While it should be degenerate to a strong topological insulator (TI) with taking the SOC into consideration.^[5,24] Hydrostatic pressure has been recognized as a powerful tool for studying band structures of semiconductors. Therefore, combined with the theoretical prediction, it is necessary to study the pressure effect on the transport properties of CaAs₃.

In this work, we have grown the single crystals of CaAs₃ and performed the studies of transport properties at atmospheric and hydrostatic pressures. CaAs₃ shows a semiconducting-like behavior at atmospheric pressure. A large hump in the resistivity is also observed for some crystals containing tiny defects or impurities. Hydrostatic pressure can effectively reduce its resistivity, but it is not fully metallized up to 2 GPa. However, a small drop of resistivity is observed above 1.5 GPa for some samples, indicating the onset of superconducting transition.

Single crystals of CaAs₃ were grown by the modified Bridgman method. Firstly, stoichiometric amounts of Ca and As granule were put into an Al₂O₃ crucible, and sealed in an evacuated quartz ampule, which was then heated extremely slowly and kept at 900 °C for 10 h before turning the furnace off. Secondly, the precursor was grinded, tableted, put into an Al₂O₃ crucible, and then sealed again as described in the first step. The quartz ampule was heated to 800 °C, kept at this temperature for 50 h, and then cooled to 400 °C slowly. The phase of obtained single crystals was characterized by x-ray diffraction (XRD) on a PANalytical diffractometer with Cu radiation at room temperature and atmospheric pressure. The elemental compositions were checked by Oxford X-Max energy dispersive x-ray (EDX) spectroscopy analysis in a Hitachi S-4800 scanning electron microscope. Electrical resistivity was measured by the conventional dc four-probe method and the measurements were performed on a Quantum Design physical properties measurement system (PPMS-9). Samples were pressurized in a piston cylinder device made of Be–Cu alloy. The pressure transmitting medium was Daphene Oil 7373. The pressure inside the device was determined by shifting the superconducting transition temperature of Pb strip.

In Fig. 1(a), the powder XRD pattern of CaAs₃ can be well indexed to the triclinic structure with the lattice parameters a = 5.88(1) Å, b = 5.85(1) Å, and c = 5.93(1) Å, α = 70.05(1)°, β = 80.16(1)°, and γ = 75.83(1)°, which are in agreement with the previous reported values.^[25] figure 1(b) shows the single crystal XRD pattern of the CaAs₃. All of peaks can be identified as the (0k0) reflections. For comparison, we also show the crystal structure of BP in Fig. 1(c). BP is a two-dimensional (2D) layered material which is stacked together by weak van der Waals force. Each P atom is covalently bonded with three adjacent P atoms to form a puckered honeycomb structure in a single layer. One can clearly see that the crystal structure of CaAs₃ is similar to that of BP. Here CaAs₃ can be viewed as a list of 2D infinite puckered polyanionic layers stacking along the b-axis and Ca²⁺ cations are inserted into the interstitial sites, in which the puckered arsenic layers are directly derived from the orthorhombic structure of BP by removing 1/4 of the P atoms in an ordered way.^[23]

Fig. 1.

	Figure Option ViewDownloadNew Window
	Fig. 1. (a) Powder XRD pattern of CaAs3 obtained by grinding the single crystals at room temperature and the refinement results. (b) The single crystal XRD of CaAs3. (c) and (d) The crystal structure of BP and CaAs3, respectively.

All the samples we measured were taken from different parts of a large crystal of CaAs₃. We find that the transport properties show a large sample dependence. The crystals peeled from the bottom exhibit semiconducting-like behavior, while some crystals taken from the top show “bad metal” behavior with a big hump at the temperature T₁, as shown in Figs. 2(a) and 2(b), respectively. It is noted that the absolute value of resistivity is essentially same between samples #1 and #2 at 300 K. In Fig. 2(a), the ρ(T) in the range from 195 K to 265 K can be fitted by Arrhenius equation, , where k_B is the Boltzmann constant. The estimated activation energy E_g is ∼73 meV, which is consistent to the previous report.^[22]

Fig. 2.

	Figure Option ViewDownloadNew Window
	Fig. 2. Resistivity ρ plotted as a function of temperature T measured on #1 crystal (a) and #2 (b) at atmospheric pressure, respectively.

To further distinguish between the two type resistivity behaviors in the single crystals (#1 and #2) of CaAs₃, we performed the Hall-effect measurements at various temperatures on the same samples, as shown in Figs. 3(a) and 3(d), respectively. The magnetic field is applied perpendicular to the ac-plane. Figure 3(a) presents the magnetic field dependence of the Hall resistivity of the crystal (#1) at several temperatures from 2 K to 300 K. The slope of the curve is all negative, suggesting that the main carriers should be electron-type. One can find that the shows nonlinear behavior at 2 K, indicating the coexistence of two types of carriers at very low temperatures.^[8,9] Different from sample #1, the slope of for sample #2 changes from negative to positive with decreasing temperatures, as shown in Fig. 3(d), indicating the transition from electrons dominated transport to holes dominated transport in this sample. All these show that CaAs₃ is a multiband system. We try to fit the Hall data obtained at 2 K for sample #1 using the semi-classical two-band model, in which Hall conductivity can be expressed as the following equation:^[26] 1 where n_h (or n_e) is the carrier density of the holes (or the electrons) and μ_h (or μ_e) is the mobility of the holes (or the electrons), respectively. For high temperatures, the appears to be approximately linear, and fitting with two-band model could bring uncertainty. Hence, we use the single-band model to estimate the carrier density , and carrier mobility (0 T), where t = e, h indicate electron-type and hole-type carriers, respectively, and the Hall coefficient R_H is estimated by the slope of (B) at low fields ( T). By fitting with the single-band and two-band model for different temperatures, the carrier densities and mobilities for sample #1 are summarized in Figs. 3(b) and 3(c), respectively. A large discrepancy is found between the results obtained from the two methods. At 2 K, the concentration and mobility for two kinds of carriers have little difference ( , , , and ). However, the results of fitting with single-band and two-band model are not consistent between 2 K and other temperatures. Above 20 K, μ_e increases monotonically with increasing temperature, and up to around at 300 K, while n_e shows a little temperature dependence from 4.2 ×10¹⁷ cm⁻³ at 20 K to 9.6 ×10¹⁷ cm⁻³ at 300 K. The temperature dependence of the carrier densities and mobilities for sample #2 are shown in Figs. 3(e) and 3(f), respectively. We find that the carrier density is larger than that of sample #1 by 1 order of magnitude. The carriers change from hole-type to electron-type as the temperature goes up to around 230 K, and the transition point is in accord with T₁ in the ρ(T) curve. Remarkably, the obvious hump in the ρ(T) curve of the #2 crystal could be attributed to dominated carriers change. We noted that the resistivity curve of HfTe₅ also showed a huge hump at about 65 K, at which the dominated carrier type changed, and which was ascribed to the occurrence of the critical topological phase transition.^[27] However, in this special case of layered crystal structure, the anomalous transport properties observed in some CaAs₃ could be reasonably attributed to the presence of impurities or crystal defects during the crystal growth.

Fig. 3.

	Figure Option ViewDownloadNew Window
	Fig. 3. (a) and (d) The magnetic field dependence of Hall resistivity on samples #1 and #2 at various temperatures, respectively. (b), (c), (e), and (f) The temperature dependence of carrier mobilities and carrier densities of electrons and holes measured on samples #1 and #2, respectively.

CaAs₃ was predicted to hold engineered topological phase transitions by applying external tensile or compressive strains.^[24] In order to test the pressure effect of CaAs₃, we use the piston cylinder device to apply compressive strains to two samples #3 and #4 which are selected with two typical transport behaviors. Figure 4(a) and 4(b) display the temperature dependence of the resistivity ρ(T) under varying hydrostatic pressures for samples #3 and #4, respectively. The values of ρ(T) both for samples #3 and #4 decrease in its entirety under pressure. At low temperatures, the ρ(T) both for two samples changes more significantly, but not metallized up to 2 GPa. In Fig. 4(b), the peak temperature, T₁, also shifts to lower temperatures with increasing pressure, while we find that ρ(T) of sample #4 drops with the lowering temperature above 1.5 GPa as shown in the inset. The transition moves to higher temperature with the increasing pressure. It seems natural to ascribe this drop to the onset of superconductivity in such “doped” sample, which provides the possibility for CaAs₃ to realize a topological nodal-line semimetal or even topological superconductivity under much higher pressure in the future work.

Fig. 4.

	Figure Option ViewDownloadNew Window
	Fig. 4. (a) and (b) Resistivity ρ of CaAs3 plotted as a function of temperature T on #3 and #4 crystals at various hydrostatic pressures, respectively. Inset of (b): Resistivity ρ above 1.2 GPa in the low temperature region (2 K–6 K).

In summary, we have synthesized the single crystals of CaAs₃, which has been predicted to be a topological nodal-line semimetal candidate. CaAs₃ shows an n-type semiconducting-like behavior at atmospheric pressure. While for some samples with the presence of crystal defect and/or impurities, the resistivity shows an obvious hump around 230 K, at which the dominated carriers change from hole-type to electron-type. The application of hydrostatic pressure results in decrease of the resistivity for both samples, but they are not fully metallized on the whole. A possible superconducting transition is also observed in some “doped” samples above 1.5 GPa. Our work will provide valuable clues for further exploration of the topological phase in CaAs₃, and in-depth studies are needed in future.