Coexistence of Superconductivity and Antiferromagentic Order in Er 2O2bi

Coexistence of superconductivity and antiferromagentic order in Er2O2Bi with anti-ThCr2Si2 structure

Lei Qiao1, Ning-hua Wu1, Tianhao Li1, Siqi Wu1, Zhuyi Zhang1, Miaocong Li1, Jiang Ma1, Baijiang 2 1 1 1 3 1,4 1,4 Lv , Yupeng Li , Chenchao Xu , Qian Tao , Chao Cao , Guang-Han Cao , and Zhu-An Xu ∗ 1 Zhejiang Province Key Laboratory of Quantum Technology and DeviceDepartment of Physics, Zhejiang University, Hangzhou 310027, China; 2 Key Laboratory of Neutron Physics and Institute of Nuclear Physics and Chemistry, China Academy of Engineering Physics, Mianyang 621999, China; 3 Department of Physics, Hangzhou Normal University, Hangzhou 310036, China; and 4 State Key Laboratory of Silicon Materials, Zhejiang University, Hangzhou 310027, China; (Dated: May 18, 2021) We investigated the coexistence of superconductivity and antiferromagnetic order in the compound Er2O2Bi with anti-ThCr2Si2-type structure through resistivity, magnetization, specific heat measurements and first-principle calculations. The superconducting transition temperature Tc of 1.23 K and antiferromagnetic transition temperature TN of 3 K are observed in the sample with the best nominal composition. The superconducting upper critical field Hc2(0) and electron-phonon coupling constant λe ph in Er2O2Bi are similar to those in the previously reported non-magnetic − superconductor Y2O2Bi with the same structure, indicating that the superconductivity in Er2O2Bi may have the same origin as in Y2O2Bi. The first-principle calculations of Er2O2Bi show that the Fermi surface is mainly composed of the Bi 6p orbitals both in the paramagnetic and antiferromagnetic state, implying minor effect of the 4f electrons on the Fermi surface. Besides, upon increasing the oxygen incorporation in Er2OxBi, Tc increases from 1 to 1.23 K and TN decreases slightly from 3 to 2.96 K, revealing that superconductivity and antiferromagnetic order may compete with each other. The Hall effect measurements indicate that hole-type carrier density indeed increases with increasing oxygen content, which may account for the variations of Tc and TN with different oxygen content.

I. INTRODUCTION ﬂuctuations is proposed to have a spin-triplet pairing s- tate [16] , and the coexistence of magnetism and superconductivity in the RPdBi systems can serve as a unique The interplay of superconductivity and magnetism has platform to investigate topological orders with multisym- been a research focus for decades. Usually, superconduc- metry breaking states[11]. tivity and magnetic order will compete with each other, and thus superconductivity seldom occurs in strong The R2O2Bi (R = rare earth) family are a series of magnetic materials [1]. In conventional s-wave supercon- layered compounds with anti-ThCr2Si2 structure. Super- ductors, local magnetic moments would break up spin conductivity has been found by excess oxygen incorpora- singlet Cooper pairs and hence strongly suppress super- tion in some of R2O2Bi members, irrespective of the pres- conductivity [2]. In the case of rare-earth superconduc- ence of magnetic ordering [17]. Superconductivity in non- tors, 4f electrons may have a localized magnetic momen- magnetic Y2O2Bi was found dependent on the oxygen t which could weakly couple with conduction electrons, content[18]. The oxygen incorporation will cause the ex- and the coexistence of superconductivity and magnetism pansion of inter-Bi-layer distance, which is proposed to be is possible to realize. There have been some examples crucial to the occurrence of superconductivity. The obvi- of such materials, like RMo6S8 [3, 4], RRh4B4 [3, 4], ous change in the carrier density companied with oxygen RBa2Cu3O7 δ [5–8], RNi2B2C[9, 10], RPdBi [11](R incorporation was not found. Meanwhile, the search for − = some of the rare earth elements), and iron-based su- superconductivity in the magnetic R2O2Bi system has perconductors containing rare earth elements [12]. In produced various outcomes. The Ce2O2Bi was found to ferromagnetic ErRh4B4 [13] and HoMo6S8 [14], and an- be an AFM Kondo lattice compound without observa- tiferromagntic (AFM) GdMo6S8[15], the competition be- tion of superconductivity for temperature down to 0.3 tween superconductivity and magnetic long-range order K[19]. Although Er2O2Bi, Tb2O2Bi and Dy2O2Bi have has resulted in the reentrant superconductivity behav- been reported to be superconducting, the detailed inves- ior. In the RNi2B2C and RPdBi systems, the compet- tigation of their properties is lacked[17, 20]. Whether itive relationship between AFM transition temperature the 4f electron magnetism has remarkable inﬂuence on TN and superconducting (SC) transition temperature Tc superconductivity is an interesting issue and it is wor- can be described by the Gennes scaling[10, 11]. More- thy of careful study. Up to now, among the magnetic over, the coexistence of magnetism and superconductiv- R2O2Bi members, only the Er2O2Bi is found to be super- ity may lead to some exotic phases. For example, UTe2 conducting without the CaO incorporation which serves where superconductivity coexists with ferromagnetic spin as an oxidant to obtain excess oxygen in R2O2Bi [17]. 2

However, the Ca ion impurity would also be incorporat- l effect were measured by a standard four-probe method ed in the samples and it may affect the superconducting on a physical property measurement system(PPMS-9). properties. Thus it is better to tune the oxygen content Specific heat measurements were also carried out on the without CaO incorporation in studying the interplay of same PPMS-9 by a heat-pulse relaxation method. We magnetism and superconductivity in a series of R2O2Bi. also used an Oxford superconducting magnet system e- 3 Here, we systematically studied the Er2O2Bi polycrys- quipped with a He cryostat to measure the supercon- talline compounds with various oxygen content by per- ducting transitions. forming magnetization, specific heat and resistivity mea- The first-principle calculations were performed with surements. The antiferromagnetic (AFM) transition as- the Vienna Ab Initio Simulation Package (VASP) code 3+ 2 sociated with the Er2 O2− layer occurs at TN of about 3 [22, 23]. Throughout the calculations, the Perdew, K, which indicates that there is a coexistence of supercon- Burke, and Ernzerhof (PBE) parameterization of gen- ductivity and AFM order in Er2O2Bi. The first-principle eralized gradient approximation (GGA) to the exchange calculations suggest that the Fermi surface of Er2O2Bi is correlation function was used [24]. The spin-orbit cou- similar to that of Y2O2Bi, implying superconductivity in pling (SOC) is included using the second variational both compounds could share the same mechanism and method. The energy cutoff of plane-wave basis was cho- 2 the conducting monatomic Bi − square net layer may be sen to be 600 eV, which was sufficient to converge the crucial to the occurrence of superconductivity. Above total energy to 1 meV/atom. A Γ-centered 18 18 18 × × TN, there seems a kind of short-range magnetic correla- Monkhorst-Pack [25] k-point mesh was chosen to sample tion which enhances the specific heat significantly and the Brillouin zone in the calculations. Considering that reduces the magnetic entropy above TN. Furthermore, the Er-4f states are fully localized at high temperature, we have synthesized four samples with different nomi- the paramagnetic (PM) band structure was calculated nal oxygen content x in the Er2OxBi compounds (x = by assuming the Er-4f open-core electron. While for the 1.8, 1.9, 2.0 and 2.1). Upon increasing x, the actual AFM state, the states of f electrons were treated as semi- oxygen content indeed increases, and TN is slightly sup- core valence states by performing LDA+U with U = 6 pressed. Meanwhile the hole-type carrier density and Tc eV [26–29]. both increases, indicating that the AFM order may compete with superconductivity. The variation of hole-type carrier density may account for the change of Tc and TN. III. RESULTS AND DISCUSSION This work provides a new example to study the interplay of rare earth magnetism and superconductivity in such Figure 1(a) shows the XRD patterns of Er2OxBi with an anti-ThCr2Si2-type structure. different nominal oxygen content (x). It is obvious that all the four samples have the same main phase of anti- ThCr2Si2 type structure with a few minor impurity phas- II. EXPERIMENTAL DETAILS es depending on the oxygen content. From the XRD patterns, the sample with x = 2.0 has the best quality with The Er2OxBi (x = 1.8, 1.9, 2.0, 2.1) polycrystalline least amount of impurity phases, therefore the most of compounds were synthesized by solid-state reactions. superconducting property’ characterization is performed High-purity Er2O3 (Alfa, 99.99%), Er (Alfa 99.8%) and based on this sample. We used the Rietveld refinements Bi (Alfa 99.999%) were used as starting materials. The employing program GSAS with EXPGUI [21] to obtain mixed powders were pressed into pellets and covered with lattice parameters of the four samples. Figure 1(b) shows Ta foils. An excess amount of Bi was added because of the Rietveld refinement of the XRD pattern of the sample its high vapor pressure. The pellets was sealed in an e- Er2O2Bi. All diffraction peaks can be indexed with the valuated silica tube, and heated at 1000 ◦C for 20 h in crystalline structure of Er2O2Bi, and their (h k l) indices furnace. After it had cooled down to room temperature are also indicated for each peak. The fitting parameters naturally, the pellets was reground and pressed again, are a = 3.8459(1) A˚ and c = 13.1585(5) A˚ with the pa- 2 then annealed at 1000 ◦C to improve the sample homo- rameters Rwp = 7.65% and χ = 1.93 which indicate a geneity. All the operations were taken in an Ar-filled good convergence. The inset of Figure 1(a) displays the glove box except the heating processes. variations of the lattice constants a and c with nominal The powder x-ray diffraction (XRD) patterns were oxygen content. As the oxygen content x increases, there recorded on a PANalytical X-ray diffractometer with Cu is a remarkable increase in the lattice constant c, and Kα radiation at room temperature. We used the Rietveld the lattice constant a remains almost constant. This c- refinements employing program GSAS with EXPGUI[21] axis expansion due to excess oxygen incorporation is also to obtain lattice parameters. The DC magnetization was observed in Y2O2Bi [18]. The inset of Figure 1(b) dis- measured on two magnetic property measurement sys- plays the anti-ThCr2Si2-type structure of Er2O2Bi which tems (MPMS-XL5 for T > 1.8 K and MPMS3 with 3He is composed of a Bi square net layer and two ErO layers. refrigerator for T < 1.8 K ). Electrical resistivity and Hal- Figure 2(a) shows the temperature-dependent resistiv- 3

FIG. 1. (Color online) (a) XRD patterns of Er2OxBi with diﬀerent value of x. The grey circle, the red rectangle, the blue triangle points represent the peak of impurity Er2O3, ErBi and Bi, respectively. The inset shows the lattice constants a and c of these four samples. (b) Rietveld reﬁnement of the XRD pattern of the sample Er2O2Bi. Inset shows the anti-ThCr2Si2-type structure of Er2O2Bi.

2 1 ity for the sample Er2O2Bi from 300 to 0.4 K. From the a2 = 1-(λ11-λ22)/λ0, λ0 = [(λ11-λ22) +4λ12λ21] 2 , h = inset of Figure 2(a), there are two obvious transitions in Hc2D1/2φ0T , η = D1/D2, U(x) = Ψ(x+1/2)-Ψ(x), Ψ(x) resistivity which should correspond to Tc and TN, respec- is the digamma function. λ11 and λ22 are the intraband tively. The χ T curves around T are also plotted in BCS coupling constants, λ and λ are the interband − N 12 21 the inset for comparison, which indicates the good co- BCS coupling constants, and D1 and D2 are the in-plane incidence between the transitions in both resistivity and diffusivity of each band. The fitting parameters are listed magnetic susceptibility at TN . The increased magnet- in Table 1. ic field does not depress the transition in magnetic sus- Table 1 The fit parameters of two-band model. ceptibility around T of 3 K, excluding a possible super- λ λ λ λ η H (0)(T) conducting transition. Figure 2(b) shows the resistivity 11 22 12 21 c2 0.95 0.51 0.29 0.84 9 0.061 of Er2O2Bi near the Tc under different magnetic fields. The resistivity decreases to zero abruptly below the SC Based on this two-band model, we can obtain Hc2(0) transition temperature Tc = 1.23 K under zero magnet- = 0.061 T which is lower than the BCS Pauli limit, i.e., ic filed. Here we define T as the temperature at which µ HBCS(0) = 1.84T 2.3 T. Compared to the upper c 0 P c ∼ the resistivity drops to 50% of normal-state resistivity critical field of Y2O2Bi which is about 0.0574 T [18], zero just above the SC transition. The Tc (zero resistivity Hc2(0) of Er2O2Bi does not show much difference, in criterion) is 1.17 K. With the magnetic field increasing, spite of the substitution of non-magnetic Y3+ ion by the 3+ Tc is suppressed continuously. The reduced temperature magnetic Er ion. It indicates that superconductivi- (T/Tc) dependence of upper critical magnetic field Hc2 ty in both compounds could share the same mechanis- 2 is presented in Figure 2(c). We tried to use the one-band m and the conducting monatomic Bi − square net layer Ginzburg-Landau theory, H = H (0)(1 t2)/(1 + t2) may be crucial to the occurrence of superconductivity. c2 c2 − (where t = T/Tc), to fit the upper critical field and the Then, we estimate Ginzburg-Landau coherence length by 2 fitting result is not good. It is found that there seems to Hc2 = Φ0/2πξ and ξ(0) of about 730 A˚ is obtained. Ap- be a positive curvature close to Tc, which could be a char- parently ξ(0) is much larger than the lattice constants, acteristic of multi-band superconductor [30–33]. There- while the length of magnetic unit cell is usually compara- fore, we introduce the two-band Ginzburg-Landau theory ble to the lattice constants. The larger ξ(0) value (than [34], the length of magnetic unit cell) could account for the coexistence of superconductivity and antiferromagnetism a0[lnt + U(h/t)][lnt + U(ηh/t)] + [10]. a1[lnt + U(h/t)] + a2[lnt + U(ηh/t)] = 0 (1) Figure 2(d) shows the temperature dependence of mag- . Here a0 = 2(λ11λ22-λ12λ21), a1 = 1+(λ11-λ22)/λ0, netic susceptibility χ (left axis) and inverse magnetic sus- 4

FIG. 2. (Color online) (a) Resistivity of Er2O2Bi from 300 to 0.4 K. Inset: temperature dependence of resistivity around two transition temperatures and the temperature dependence of magnetic susceptibility under the magnetic field H = 100 and 1000 Oe around TN . (b) Resistivity of Er2O2Bi near the Tc under different magnetic fields. (c) Normalized temperature dependence of upper critical field of Er2O2Bi. The blue dotted line and the red solid line represent the fitting of one-band and two-band Ginzburg-Landau theory, respectively. (d) Temperature dependence of magnetic susceptibility, χ (left axis) and inverse magnetic susceptibility, 1/χ (right axis) measured under the magnetic field H = 100 Oe. The red solid line shows the Curie-Weiss fit over the temperature range 7 - 300 K. (e) Temperature dependence of magnetic susceptibility χ from T = 0.4 to 1.8 K in the zero-field cooling (ZFC) process (black solid circles) and field cooling (FC) process (red hollow circles) under the magnetic field H = 10 Oe, and temperature dependence of magnetic susceptibility χ from T = 2 to 5 K under the magnetic field H = 100 Oe (blue solid squares). Inset: field dependence of magnetization M(H) at T = 0.5 K. (f) Magnetic susceptibility χ measured in the temperature T -sweep mode containing both T -up and T -down processes between temperature 2 - 4 K. ceptibility, 1/χ (right axis) measured under the magnet- ing (ZFC) processes under the magnetic field H = 10 ic field H = 100 Oe. The χ(T ) satisfies the modified Oe around Tc, respectively. The data represented by the Curie-Weiss law χ = χ0 + /(T θW) where χ0 is a blue solid squares is measured under H = 100 Oe to show temperature independent susceptibility,C − is the Curie the change around T . To avoid demagnetization effec- C N constant and θW is the Weiss temperature. The red solid t, the sample is shaped into a long and narrow cuboid line shows the Curie-Weiss fit in the temperature range (about 3 0.66 0.4 mm3) and the magnetic field is between 7 - 300 K and the derived parameters = 11.77 applied along× the× long axis of the sample. The SC vol- C emu/mol-Er K, χ0 = 0.00372 emu/mol-Er and θW = - 22 ume fraction is estimated to be about 133% based on the K. The negative Weiss temperature indicates an AFM ex- ZFC data, confirming the bulk nature of superconductiv- change interaction between the magnetic moments. The ity in Er2O2Bi. Although the demagnetization factor is calculated effective magnetic moment based on the Curie close to 0, the overestimated SC volume fraction which constant is 9.7 µB/Er which is very close to that of the is beyond 100% may be caused by the error in measuring 3+ free Er ion (9.8 µB/Er). It implies that the 4f elec- the sample volume. The inset of Figure 2(e) is the field trons of Er3+ ion are very localized in the temperature dependence of magnetization M(H) at T = 0.5 K which range between 7 - 300 K. Figure 2(e) displays the tem- is a typical behavior of type II superconductor. Figure perature dependence of magnetic susceptibility χ from T 2(f) shows the magnetic susceptibility χ measured in the = 0.4 to 5 K. It is clear that there exist two transitions temperature T -sweep mode containing both T -up and T - which correspond to the AFM order and SC transition down processes in the temperature range between 2 - 4 respectively. The red hollow circles and black solid cir- K. To avoid the error of T , the speed of T sweep is as cles stand for the field cooling (FC) and zero-field cool- slow as 0.15 K/min. There is a clear magnetic hystere- 5 sis around T of 3 K under 100 Oe, suggesting a possible ions. Based on ΘD = 146 K, the electron-phonon cou- first-order transition. pling constant λe ph is estimated to be 0.53 from the − Figure 3(a) shows the temperature dependence of spe- McMillan equation[36]: cific heat C of Er2O2Bi. The peak at TN is so sharp, ΘD indicating a possible first-order transition. Such a sharp µ∗ln( 1.45T c ) + 1.04 λe ph = (3) peak in specific heat has been observed in the magnet- − ΘD ln( 1.45T c )(1 0.62µ∗) 1.04 ic first order phase transitions in other compounds, i.e., − − DyCu2Si2 [35]. With the magnetic field increasing, the where µ∗ is the Coulomb pseudopotential, and an empir- peak is suppressed gradually, consistent with the AFM ical value of 0.13 is used. This value of λe ph in Er2O2Bi − order. As shown in the inset of Figure 3(a), TN decreas- is as the same as the value λe ph = 0.53 in Y2O2Bi, both − es monotonously with increasing magnetic field. Figure pointing to the typical weak coupling conventional BCS 3(b) displays the temperature dependence of specific heat superconductors. C of Er O Bi among the temperature range 0 - 2 K under The Sommerfeld coefficient γ 324 mJ/mol K2 (un- 2 2 ∼ magnetic fields µ0H = 0, 0.01, 0.05 and 0.1 T, respec- der zero magnetic field) is quite large. Usually, the large tively. There is no obvious peak of specific heat observed γ value may result from Kondo interactions or strong at the SC transition temperature Tc = 1.23 K, which is magnetic fluctuations, which is difficult to distinguish probably hidden in the shoulder of the large AFM peak. them only based on the specific heat data. After careful- However, from the inset of Figure 3(b), the different be- ly investigating the behaviors of resistivity and magnetic haviors of C/T at the lowest temperature may be induced susceptibility, we argue that it should be caused by the by SC transition. The C/T under zero magnetic field has magnetic fluctuations rather than the Kondo interaction- a drop around 0.6 K, while the drop of C/T disappeared s. There is no obvious feature of Kondo effect in the tem- under µ0H = 0.01 T, and C/T become coincident under perature dependence of resistivity, such as a negative log- µ0H = 0.05 and 0.1 T. Concerning the C/T behavior in arithmic temperature dependence of resistivity induced 3+ Y2O2Bi [18], there is also an upturn below about 0.6 K by Kondo scattering. The magnetic moments of Er which is proposed to be caused by a Schottky anomaly o- are localized from 300 to 7 K derived from magnetic sus- riginating from the 209Bi nuclei (I = 9/2). Therefore, we ceptibility, which is not consistent with the Kondo effect may ascribe the upturn below 0.6 K in C/T of Er2O2Bi model (magnetic moments will be partially screened by to the Bi nuclear contributions and the drop below 0.6 K the conduction electrons at low temperature). Instead, under 0 T to the SC transition. With increasing magnet- it seems that the magnetic fluctuations could account for ic field, the SC transition is suppressed and the upturn the large γ, which is supported by the calculated magnet- behavior gradually appears. When the magnetic field is ic entropy (to be shown below). Considering the contri- beyond Hc2(T ), i.e., µ0H = 0.05 and 0.1 T, the drop in bution to specific heat from spin fluctuations in Er2O2Bi C/T induced by the SC transition disappears and the do not include a T 3lnT term which is a character of fer- C/T would remain almost the same with further increas- romagnetic fluctuations, the spin fluctuations here may ing magnetic field. be antiferromagnetic type [37]. Figure 3(c) shows the specific heat divided by temper- Figure 3(d) shows the magnetic entropy calculated by 2 ature, C/T , versus T under the magnetic field µ0H = integrating Ce/T over temperature. The Ce denotes the 0 and 9 T. The following equation is used to fit the C/T electronic contribution to specific heat, which is separat- data, ed from the phonon contribution. The phonon contributions are calculated by the Debye model with ΘD = 146 2 C/T = γ + βT , (2) K. The magnetic entropy does not recover Rln2 at TN and keeps increasing, and finally saturates as tempera- where γ and β denote the coefficient of electron and ture is close to 18 K. Assuming that the ground state of phonon contributions for Er2O2Bi, respectively. The de- Er2O2Bi is Kramers two-fold degenerate, the free mag- rived parameters are γ = 324 (425) mJ/mol K2 and β netic entropy should be Rln2. The reduced magnetic 4 = 3.1 (3.7) mJ/mol K under the magnetic field µ0H = entropy at TN may be interpreted by a kind of short- 0 (9) T. The derived Debye temperatures ΘD is 146 K range magnetic correlation above the AFM long-range and 138 K under µ0H = 0 and 9 T respectively, by using order temperature (TN ). The saturation value of mag- 4 1/3 the formula ΘD = (12π NR/5β) . Since the magnetic netic entropy at T = 18 K is a little smaller than Rln2, field usually should not change the phonon contribution which is probably introduced by the error in calculating to specific heat, the very close values of Debye temper- the phonon contributions based on the Debye model. atures ΘD for µ0H = 0 T and 9 T are reasonable. The Figure 4(a) shows the temperature dependence of Hal- Debye temperature of Er2O2Bi (146 K) is a little smaller l resistivity for the sample Er2O2Bi. The magnetic than that of La2O2Bi, which is about 150 K [19]. The field dependence of Hall resistivity at high temperature difference between their Debye temperatures may result is nearly linear which can be fitted by the usual one- from a little larger mass of Er ions compared to the La band model. However, when the temperature decreases, 6

FIG. 3. (Color online) (a) Temperature dependence of the specific heat C of Er2O2Bi under magnetic fields µ0H = 0, 2, 4, 6 and 9 T, respectively. Inset: magnetic field dependence of TN which is determined by the peak of dC/dT . (b) Temperature dependence of the specific heat C of Er2O2Bi among the temperature range 0 - 2 K under the magnetic fields µ0H = 0, 0.01, 0.05 and 0.1 T, respectively. Inset: C/T versus T under different magnetic field µ0H = 0, 0.01, 0.05 and 0.1 T, respectively. 2 (c) Specific heat divided by temperature C/T versus T under magnetic fields µ0H = 0 and 9 T, respectively. The red solid line represents the fitting by equation 2. (d) Magnetic entropy calculated by integrating C/T (exclude phonon contribution) and T .

the magnetic field dependence becomes nonlinear, which impurity phases like Er2O3. Thus the sample with x = could be induced by magnetic skew-scattering and the 2.1 does not have good quality and its Hall coefficient effects of two types of charge carriers. Figure 4(b) dis- is not represented for discussions. Figure 4(c) shows the plays the temperature dependence of Hall coefficient of x dependence of carrier density nH and mobility µ cal- Er2OxBi with x = 1.8, 1.9 and 2.0, respectively. With culated based the Hall coefficient and resistivity at 100 increasing nominal oxygen content (x), the positive Hal- K. The carrier density is calculated by a single band ap- l coefficient is decreasing, implying the increase in the proximation. It is obvious that there is an increase in the hole-type carrier density. Since the samples with x = 2.1 carrier density as x increases from 1.8 to 2.0, which means and x = 2.0 have almost the same parameters of c-axis, the oxygen incorporation indeed induce more hole-type and meanwhile their Tc and TN are also the same, we charge carriers. believe that both sample should have the same oxygen Figure 5(a) shows temperature dependence of resistiv- content, and the excess oxygen content in the sample ity of the samples Er2OxBi with different oxygen content with nominal x = 2.1 may lead to the formation of a few x. As x increases, Tc increases from 1 K to 1.23 K. We use 7

FIG. 4. (Color online) (a) Magnetic field dependence of Hall resistivity of the sample Er2O2Bi at different temperatures. (b) Temperature dependence of the Hall coefficient RH of Er2OxBi with x = 1.8, 1.9 and 2.0, respectively. The RH is determined from the initial slope of ρyx(B) at B 0. (c) x dependence of carrier density (left axis) and mobility (right axis) at 100 K. The connected lines are the guides to→ eyes.

FIG. 5. (Color online) (a) Temperature dependence of resistivity of Er2OxBi. (b) Temperature dependence of the derivative susceptibility T dχ/dT of samples with diﬀerent x. (c) x dependence of nH (left axis) at T = 100 K, Tc (left axis) and TN (right axis).

the peak of T dχ/dT displayed in Figure 5(b) to define electron-phonon coupling of Er2O2Bi and Y2O2Bi. By the TN. In contrast to the SC transition, TN decreases integrating the 4f contribution below EF, each Er ion with increasing x. The parameters, nH, Tc and TN of the has a spin polarization of 2.9 µB in the z direction, cor- samples are presented in the Figure 5(c). The increase in responding to a spin configuration of S = 3/2. Therefore, the carrier density can account for the increase of Tc and the effective magnetic moment can be obtained with µeff the decrease of TN, which reveals that the AFM order = gj (J(J + 1))µB. In this calculation, it is found gj = may compete with superconductivity. 1.2, resulting in µeff = 9.58 µB, in good agreement with above experimental data, implying the fully localized na- The primitive unit cell (with the simplest AFM con- ture of 4f electrons in Er2O2Bi in the AFM ground state. figuration) and Brillouin zone of Er2O2Bi are shown in Figure 6(a) and (b). The band structures of PM phase (black line) and AFM phase (red line) are shown in 6(c). In the AFM phase, the Er-4f orbitals are away from the IV. CONCLUSION Fermi level, located around 6 eV (1 eV) below (above) Fermi energy EF. Thus the electronic structure of AFM In summary, we have systematically investigated phase resembles that of PM phase near Fermi surface the superconducting properties and AFM order of the with two bands derived from Bi-p orbits crossing EF, Er2O2Bi compounds with different oxygen content by which indicates that the magnetic moments of Er3+ ions performing resistivity, magnetization, specific heat mea- probably have little effect on the electronic structure near surements and first-principle calculations. There may ex- the Fermi level. This result may account for the similar ist strong magnetic fluctuations or short-range AFM cor- 8

FIG. 6. (Color online) (a) Primitive unit cell (with the simplest AFM configuration) of Er2O2Bi. The a, b and c represent the directions of the lattice vectors. The directions of magnetic moments of different Er atoms in the simplest AFM configuration are also displayed. (b) Brillouin zone of Er2O2Bi. The b1, b2 and b3 are the reciprocal lattice vectors and the red points are high symmetry points. (c) Band structures of paramagnetic phase (black line) and AFM phase (red line). The hollow circles with different colours represents the different orbits of Bi.

relation above TN and they contributes to the significant- [1] L. N. Bulaevskii, A. I. Buzdin, M. L. Kulić, and S. V. ly enhanced Sommerfeld coefficient of specific heat. The Panjukov, Coexistence of superconductivity and magnetism theoretical predictions and experimental results, magnitude of Hc2(0) and electron-phonon coupling constants in both Er O Bi and Y O Bi are very close, imply- Adv. Phys. 34, 175 (1985). 2 2 2 2 [2] P. W. Anderson, Theory of dirty superconductors, J. ing that superconductivity in both magnetic or nonmag- Phys. Chem. Solids 11, 26 (1959). netic R2O2Bi (R = Y and Er) may share the same origin. [3] O. Fischer and L. B. Maple, Superconductivity in Ternary The band structures from the first-principle calculation- Compounds I: Structural, Electronic, and Lattice Proper- s support the view that the charge carriers in Er2O2Bi ties, Vol. 32 (Springer Science & Business Media, 2012). mainly come from the Bi square net layer. Therefore, the [4] M. B. Maple and Ø. Fischer, in 2 Superconductivity in Ternary Compounds II , Vol. 34 conducting monatomic Bi − square net layer may be crucial to the occurrence of superconductivity in R O Bi. (Springer, Boston, MA, 1982) pp. 1–10. 2 2 [5] D. M. Paul, H. A. Mook, A. W. Hewat, B. C. Sales, By adjusting the nominal oxygen content in Er2OxBi, L. A. Boatner, J. R. Thompson, and M. Mostoller, Mag- we obtained several samples with different carrier den- netic ordering in the high-temperature superconductor sity. The Er2O2Bi compound is a typical example with GdBa2Cu3O7, Phys. Rev. B 37, 2341 (1988). the coexistence of superconductivity and AFM order in [6] B. D. Dunlap, M. Slaski, D. G. Hinks, L. Soderholm, the series of R2O2Bi (R = rare earth or Y) materials. M. Beno, K. Zhang, C. Segre, G. W. Crabtree, W. K. This study suggested that there may be a competition Kwok, S. K. Malik, et al., Electronic and magnetic properties of rare-earth ions in REBa2Cu3O7 x (RE= Dy, between superconductivity and AFM order in Er O Bi. − 2 2 Ho, Er), J. Magn. Magn. Mater. 68, L139 (1987). It is worth further studying to understand the interplay [7] B. D. Dunlap, M. Slaski, Z. Sungaila, D. G. Hinks, of rare earth magnetism and superconductivity in such K. Zhang, C. Segre, S. K. Malik, and E. E. Alp, Magnet- an anti-ThCr2Si2-type structure. ic ordering of Gd and Cu in superconducting and non- superconducting GdBa2Cu3O7 δ, Phys. Rev. B 37, 592 (1988). − ACKNOWLEDGMENTS [8] Z. Zou, J. Ye, K. Oka, and Y. Nishihara, Superconducting PrBa2Cu3Ox, Phys. Rev. Lett. 80, 1074 (1998). [9] H. Eisaki, H. Takagi, R. J. Cava, B. Batlogg, J. J. Kra- The authors would like to thank Yifeng Yang, and jewski, W. F. Peck Jr, K. Mizuhashi, J. O. Lee, and Jianhui Dai for insightful discussions. This work was S. Uchida, Competition between magnetism and super- supported by the National Key R&D Projects of China conductivity in rare-earth nickel boride carbides, Phys. (Grant Nos. 2019YFA0308602 and 2016YFA0300402), Rev. B 50, 647 (1994). [10] K. H. Müller and V. N. Narozhnyi, Interaction of su- the National Science Foundation of China (Grant No. perconductivity and magnetism in borocarbide supercon- 11774305), and the Fundamental Research Funds for the ductors, Rep. Prog. Phys. 64, 943 (2001). Central Universities of China. [11] Y. Nakajima, R. Hu, K. Kirshenbaum, A. Hughes, P. Syers, X. Wang, K. Wang, R. Wang, S. R. Saha, D. Pratt, et al., Topological RPdBi half-Heusler semimet- als: A new family of noncentrosymmetric magnetic superconductors, Sci. Adv. 1, e1500242 (2015). ∗ [email protected] 9

[12] For example, Y. Luo, H. Han, S. Jiang, X. Lin, Y. Li, [24] J. P. Perdew, K. Burke, and M. Ernzerhof, Generalized J. Dai, G. Cao, and Z.-a. Xu, Interplay of superconduc- gradient approximation made simple, Phys. Rev. Lett. tivity and Ce 4f magnetism in CeFeAs1 xPxO0.95F0.05, 77, 3865 (1996). − Phys. Rev. B 83, 054501 (2011). [25] H. J. Monkhorst and J. D. Pack, Special points for [13] W. A. Fertig, D. C. Johnston, L. E. DeLong, R. W. Mc- Brillouin-zone integrations, Phys. Rev. B 13, 5188 Callum, M. B. Maple, and B. T. Matthias, Destruction of (1976). Superconductivity at the Onset of Long-Range Magnetic [26] V. I. Anisimov, J. Zaanen, and O. K. Andersen, Band Order in the Compound ErRh4B4, Phys. Rev. Lett. 38, theory and Mott insulators: Hubbard U instead of Stoner 987 (1977). I, Phys. Rev. B 44, 943 (1991). [14] M. Ishikawa and Ø. Fischer, Destruction of superconduc- [27] V. I. Anisimov and O. Gunnarsson, Density-functional tivity by magnetic ordering in Ho1.2Mo6S8, Solid State calculation of effective Coulomb interactions in metals, Commun. 23, 37 (1977). Phys. Rev. B 43, 7570 (1991). [15] M. Ishikawa, Ø. Fischer, and J. Muller, in Superconduc- [28] V. N. Antonov, B. N. Harmon, and A. N. Yaresko, Elec- tivity in Ternary Compounds II (Springer, 1982) pp. 143– tronic structure of mixed-valence and charge-ordered Sm 165. and Eu pnictides and chalcogenides, Phys. Rev. B 72, [16] L. Jiao, S. Howard, S. Ran, Z. Wang, J. O. Rodriguez, 085119 (2005). M. Sigrist, Z. Wang, N. P. Butch, and V. Madhavan, [29] C. Xu, Q. Chen, and C. Cao, Unique crystal field s- Chiral superconductivity in heavy-fermion metal UTe2, plitting and multiband RKKY interactions in Ni-doped Nature 579, 523 (2020). EuRbFe4As4, Commun. Phys. 2, 1 (2019). [17] R. Sei, H. Kawasoko, K. Matsumoto, M. Arimitsu, [30] H. Wang, C. Dong, Q. Mao, R. Khan, X. Zhou, C. Li, K. Terakado, D. Oka, S. Fukuda, N. Kimura, H. Kasai, B. Chen, J. Yang, Q. Su, and M. Fang, Multiband su- E. Nishibori, et al., Tetragonality induced superconduc- perconductivity of heavy electrons in a TlNi2Se2 single tivity in anti-ThCr2Si2-type RE2O2Bi (RE= rare earth) crystal, Phys. Rev. Lett. 111, 207001 (2013). with Bi square nets, Dalton Trans. 49, 3321 (2020). [31] S. V. Shulga, S.-L. Drechsler, G. Fuchs, K.-H. Müller, [18] R. Sei, S. Kitani, T. Fukumura, H. Kawaji, and K. Winzer, M. Heinecke, and K. Krug, Upper criti- T. Hasegawa, Two-Dimensional Superconductivity E- cal field peculiarities of superconducting YNi2B2C and 2 merged at Monatomic Bi − Square Net in Layered LuNi2B2C, Phys. Rev. Lett. 80, 1730 (1998). Y2O2Bi via Oxygen Incorporation, J. Am. Chem. Soc. [32] F. Hunte, J. Jaroszynski, A. Gurevich, D. C. Larbalestier, 138, 11085 (2016). R. Jin, A. S. Sefat, M. A. McGuire, B. C. Sales, D. K. [19] L. Qiao, J. Chen, B. Lv, X. Yang, J. Wu, Y. Cui, H. Bai, Christen, and D. Mandrus, Two-band superconductivity M. Li, Y. Li, Z. Ren, et al., Antiferromagnetic Kondo lat- in LaFeAsO0.89F0.11 at very high magnetic fields, Nature tice compound Ce2O2Bi with anti-ThCr2Si2-type struc- 453, 903 (2008). ture, J. Alloys Compd. 836, 155229 (2020). [33] H. Bai, X. Yang, Y. Liu, M. Zhang, M. Wang, Y. Li, [20] K. Terakado, R. Sei, H. Kawasoko, T. Koretsune, D. Oka, J. Ma, Q. Tao, Y. Xie, G.-H. Cao, et al., Superconduc- T. Hasegawa, and T. Fukumura, Superconductivity in tivity in a misfit layered compound (SnSe)1.16(NbSe2), J. anti-ThCr2Si2-type Er2O2Bi induced by incorporation of Phys.: Condens. Matter 30, 355701 (2018). excess oxygen with CaO oxidant, Inorg. Chem. 57, 10587 [34] A. Gurevich, Enhancement of the upper critical field by (2018). nonmagnetic impurities in dirty two-gap superconduc- [21] B. H. Toby, EXPGUI, a graphical user interface for tors, Phys. Rev. B 67, 184515 (2003). GSAS, J. Appl. Cryst. 34, 210 (2001). [35] Y. Takeda, N. Duc Dung, Y. Nakano, T. Ishikura, S. Ike- [22] G. Kresse and J. Hafner, Ab initio molecular dynamics da, T. D. Matsuda, E. Yamamoto, Y. Haga, T. Takeuchi, for liquid metals, Phys. Rev. B 47, 558 (1993). R. Settai, et al., Calorimetric Study in Single Crystalline [23] G. Kresse and J. Furthmüller, Efficient iterative schemes RCu2Si2 (R: Rare Earth), J. Phys. Soc. Jpn. 77, 104710 for ab initio total-energy calculations using a plane-wave (2008). basis set, Phys. Rev. B 54, 11169 (1996). [36] W. L. McMillan, Transition temperature of strong- coupled superconductors, Phys. Rev. 167, 331 (1968). [37] T. Moriya, Spin fluctuations in nearly antiferromagnetic metals, Phys. Rev. Lett. 24, 1433 (1970).