Bulk superconductivity in a four-layer-type Bi-based compound La2O2Bi3Ag0.6Sn0.4S5.7Se0.3

Recently, we reported the observation of superconductivity at ~0.5 K in a La2O2M4S6-type (M: metal) layered compound La2O2Bi3AgS6, which is a layered system related to the BiS2-based superconductor system but possesses a thicker Bi3AgS6-type conducting layer. In this study, we have developed the La2O2Bi3AgS6-type materials by element substitutions to increase the transition temperature (Tc) and to induce bulk nature of superconductivity. A resistivity anomaly observed at 180 K in La2O2Bi3AgS6 was systematically suppressed by Sn substitution for the Ag site. By the Sn substitution, Tc increased, and the shielding volume fraction estimated from magnetization measurements also increased. The highest Tc (=2.3 K) and the highest shielding volume fraction (~20%) was observed for La2O2Bi3Ag0.6Sn0.4S6. The superconducting properties were further improved by Se substitutions for the S site. By the combinational substitutions of Sn and Se, bulk-superconducting phase of La2O2Bi3Ag0.6Sn0.4S5.7Se0.3 with a Tc of 3.0 K (Tconset = 3.6 K) was obtained.

Layered superconductors have been extensively studied due to observations of unconventional superconductivity and high transition temperature (T c ) 1,2 . In addition, the great flexibility of constituent elements and layered (stacking) structure, the studies on layered superconductor system can be widely developed. The recent discovery of BiS 2 -based superconductors has also created remarkable attention in the superconductivity community: the typical materials are Bi 4 O 4 S 3 , REO 1−x F x BiS 2 (RE = La, Ce, Pr, Nd, Yb), and Sr 1−x RE x FBiS 2 [3][4][5][6][7][8][9][10][11][12][13][14] . For instance, the crystal structure of the typical system LaOBiS 2 is composed of alternate stacks of a La 2 O 2 blocking layer and two BiS 2 layers. Since the parent phase (LaOBiS 2 ) is an insulator with a band gap 15 , electron doping is needed to induce metallic and superconducting characteristics. Furthermore, local in-plane structure should be optimized to induce bulk superconductivity in the BiS 2 layer 16 . Thus, the T c of BiS 2 -based is sensitive to physical pressure 4,17-21 and chemical pressure [22][23][24] effects and reaches 11 K in LaO 0.5 F 0.5 BiS 2 . Therefore, an increase in the highest record of T c in the BiS 2 -based superconductor family can be expected by further material development and tuning of structural and electrical properties 14 .
Very recently, we reported on the superconductivity in oxychalcogenide La 2 O 2 Bi 3 AgS 6 with T c of 0.5 K 25 . The crystal structure of La 2 O 2 Bi 3 AgS 6 is similar to the typical BiS 2 -based superconductor LaOBiS 2 , but the conducting layer thickness is thicker than that of LaOBiS 2 . The crystal structure of the system can be described as La 2 O 2 M 4 S 6 , in which the M site can be Pb, Bi, and Ag [26][27][28] . There are two different metal sites [M1 and M2 displayed in Fig. 1(c)] in the M 4 S 6 -type conducting layer. For the outer layers with the M1 site, Bi selectively occupies, and the layers can be regarded as the BiS 2 -type layers. For the inner layer with the M2 site, NaCl-type (Bi,Ag)S 2 layers are inserted in between the BiS 2 -type layers.
Interestingly, an anomalous transport property was observed in the temperature dependence of resistivity in La 2 O 2 Bi 3 AgS 6 25 , which is similar to the charge-density-wave (CDW) transition in EuFBiS 2 29 . In this study, we have investigated the Sn substitution effect for the Ag site in La 2 O 2 Bi 3 Ag 1−x Sn x S 6 and found that the resistivity anomaly was suppressed by Sn substitutions. T c reached 2.3 K for x = 0.4 (La 2 O 2 Bi 3 Ag 0.6 Sn 0.4 S 6 ). Furthermore, by Se substitutions for the S site in La 2 O 2 Bi 3 Ag 0.6 Sn 0.4 S 5.7 Se 0.3 , the T c further increased to 3.5 K, and bulk nature of superconductivity was confirmed from magnetic shielding volume fraction.

Results
Sn substitution effect on structural and physical properties in La 2 o 2 Bi 3 Ag 1−x Sn x S 6 . Figure 1a displays the room temperature XRD patterns for La 2 O 2 Bi 3 Ag 1−x Sn x S 6 (x = 0-0.5). All the La 2 O 2 Bi 3 Ag 1−x Sn x S 6 samples are crystallized in the tetragonal structure with the space group of P4/nmm. An impurity phase of La 2 Sn 2 O 7 was observed for x = 0.2-0.5. Figure 1b shows the shift in the 103 peak position, which slightly shifts towards the low angle side for x = 0.1. As the Sn concentration increases from x = 0.2 to 0.5, the 103 peak shifts towards the higher angle side. The schematic image of the crystal structure of La 2 O 2 M 4 S 6 is shown in Fig. 1c. For the La 2 O 2 Bi 3 Ag 1−x Sn x S 6 phase, Bi selectively occupies the M1 site, and Sn is expected to substituted for Ag at the M2 site, which was qualitatively confirmed by Rietveld refinements. The evolutions of the lattice parameters by the Sn substitution are shown in Fig. 1d,e. The lattice parameters are a = 4.061(1) Å and c = 19.445(1) Å for x = 0 and a = 4.0648(1) Å and c = 19.48(1) Å for x = 0.1. The lattice parameter c tends to decrease by Sn substitution for x = 0-0.2, and then, it continuously increases with increasing x for x = 0.2-0.5. The lattice parameter a increases for the x = 0.1, but it tends to decrease with increasing x for higher x. However, the changes in those lattice parameters are small. On the basis of the lattice parameter evolutions, we consider that the Sn substitution does not largely affect the lattice volume, which may be due to smaller ionic radius of Sn 2+ (93 pm) than that of Ag + (113 pm). The La 2 O 2 Bi 3 AgS 6 -type structure can be considered as the stacking of a La 2 O 2 layer, two BiS 2 layers, and an NaCl-type AgBiS 2 layer. In such a layered structure, the unit cell is almost determined by the hardest layer due to the different ionic bonding nature. In this structure, La 2 O 2 layer structure is the hardest.
The actual ratio of the metals in the conducting layer (Bi, Ag, and Sn) in the La 2 O 2 Bi 3 Ag 1−x Sn x S 6 (x = 0-0.5) samples were examined by EDX. The nominal and analyzed values of x are plotted in the Fig. 1f. Although there is slight deviation from the nominal values, the analyzed values of x EDX linearly increases with increasing nominal x. We found that Bi is slightly excess for x = 0.1-0.3. Therefore, y parameter with a formula of La 2 O 2 Bi 3+y (Ag 1−x Sn x ) 1−y S 6 was introduced to analyze the Bi concentration. The analyzed values of y EDX are plotted in Fig. 1g. y EDX is higher for x = 0.1-0.3 but almost zero for x = 0, 0.4, and 0.5. Therefore, we consider that the Bi excess can be ignored in the discussion on the Sn substitution (and Se substitution) effect, and we use the formula La 2 O 2 Bi 3 Ag 1−x Sn x S 6 in this paper. Figure 2a-c show the temperature dependences of magnetic susceptibility (4πχ-T) under an applied magnetic field of 10 Oe for La 2 O 2 Bi 3 Ag 1−x Sn x S 6 (x = 0.3-0.5). The diamagnetic signals in the 4πχ curve were observed below 2.2, 2.8, and 2.6 K for x = 0.3, 0.4, and 0.5, respectively. A large diamagnetic signal was observed below 2.8 K in the ZFC curve for x = 0.4. The shielding value fractions estimated from 4πχ (ZFC) at 1.9 K is nearly 20% [See Fig. 2(d).] while it is still not saturated. From the susceptibility results, we consider that Sn substitution is effective to improve the superconducting properties of La 2 O 2 Bi 3 AgS 6 but not sufficient to induce bulk superconductivity. Figure 3 shows the temperature dependences of electrical resistivity from 300 to 0.1 K for La 2 O 2 Bi 3 Ag 1−x Sn x S 6 (x = 0-0.5). The electrical resistivity at 300 K decreases with increasing Sn concentration up to x = 0.3 and increases again for x = 0.4 and 0.5. The normal-state resistivity of the Sn-doped samples changes remarkably. For example, the pure sample (x = 0) shows a linear decrease in resistivity on cooling below the anomaly temperature T* = 180 K. A similar behavior was observed up to x = 0.2. The resistivity anomaly at T* appears for x ≤ 0.2, and the T* shifts towards the lower temperature side with increasing x. In contrast, the normal-state ρ(T) for x = 0.3-0.5 shows an upturn below ~50 K. The anomaly disappears for x ≥ 0.3. Figure 3g shows the zoomed view of the Figs. 3a-f near the superconducting transition. The T c clearly increases with increasing Sn concentration in La 2 O 2 Bi 3 Ag 1−x Sn x S 6 . The highest T c was achieved for x = 0.4, and T c decreases for a higher substitution with x = 0.5.
The room-temperature Seebeck coefficient (S) for La 2 O 2 Bi 3 Ag 1−x Sn x S 6 (x = 0-0.5) are shown in Fig. 4. The Seebeck coefficient is a good scale for the carrier concentration in BiS 2 -based compounds 30 . We observed a slight change in S by Sn substitution. The S in x = 0.2-0.4 are almost the same, but that for x = 0 and 0.5 are slightly large. This suggests that the carrier concentrations for x = 0.2-0.4 are higher than those for x = 0 and 0.5. This seems to be related to the evolution of T c . However, the large change in T c from 0.6 to 2.3 K between x = 0.1 and 0.4 cannot be simply understood by the carrier concentration only. Figure 5 shows the superconductivity phase diagram of La 2 O 2 Bi 3 Ag 1−x Sn x S 6 , which shows the interplay between the resistivity anomaly temperature (T * ) and the superconducting transition temperature (T c zero ). The T* is suppressed by the Sn substitution, and it disappears at x = 0.3. The T c gradually increases with increasing x in La 2 O 2 Bi 3 Ag 1−x Sn x S 6 . The highest T c zero = 2.3 K is achieved for x = 0.4. A lower T c zero = 1.9 K is observed for the highest (solubility-limit) Sn concentration of x = 0.5.
Here, we discuss about the possible influence of the presence of the La 2 Sn 2 O 7 impurity to the composition. Due to the change in the impurity amount for x = 0-0.5, the actual compositions may deviate from the nominal compositions. Although the Sn concentration to that of Ag was checked by EDX (Fig. 1f), oxygen deficiency in the blocking layer was not checked in this study. However, we consider that oxygen deficiency was not introduced because it is expected to make the structure unstable even if it has been introduced in the La 2 O 2 Bi 3 AgS 6 -type structure. We have tried to dope electrons by oxygen deficiency or fluorine substitution for the La 2 O 2 Bi 3 AgS 6 -type structure. However, in the La 2 O 2 Bi 3 AgS 6 system, such trials of fluorine substitutions (or oxygen deficiency) resulted in decomposing of the La 2 O 2 Bi 3 AgS 6 -type structure into LaO 1−x F x BiS 2 . This indicates that an electron-doped composition with the La 2 O 2 Bi 3 AgS 6 -type structure cannot be obtained easily due to the competition to the high stability of the REO 1−x F x BiS 2 -type phase. In addition, oxygen deficiency has not been observed in the REOBiS 2 systems.  www.nature.com/scientificreports www.nature.com/scientificreports/ from the EDX analyses for Bi, Ag, Sn, S, Se elements was La 2 O 2 Bi 3.09 Ag 0.65 Sn 0.26 S 5.73 Se 0.27 . Since the obtained composition is close to the nominal formula, we call the sample with the nominal value below. Figure 6 shows the XRD pattern and the Rietveld refinement result for La 2 O 2 Bi 3 Ag 0.6 Sn 0.4 S 5.7 Se 0.3 . Although two peaks related to the La 2 Sn 2 O 7 impurity phase were observed, other peaks could be refined using the tetragonal (P4/nmm) model with a reliability factor R wp of 13.4%. In the refinement, Se was assumed to be substituted for the S1 site. The lattice parameters were a = 4.0759(2) Å and 19.4824(11) Å, which are clearly larger than those of La 2 O 2 Bi 3 Ag 1−x Sn x S 6 due to the presence of Se. Figure 7 displays the superconducting properties of La 2 O 2 Bi 3 Ag 0.6 Sn 0.4 S 5.7 Se 0.3 . As shown in Fig. 7a, a large shielding volume fraction close to 100% was observed. From the resistivity measurements (Fig. 7b), zero resistivity was observed at 3.0 K, and the onset temperature (T c onset ) was 3.5 K; we estimated the temperature where the resistivity becomes almost 90% of normal-state resistivity. Although superconductivity was observed, the ρ(T) curve still shows a semiconducting-like localization at low temperatures. We have measured ρ(T) under magnetic fields up to 9 T. The obtained T c onset and T c zero were plotted in Fig. 7d to evaluate the upper critical field H c2 and the irreversible field H irr . The H c2 (0) was estimated as 2.15 T using the WHH model (Werthamer-Helfand-Hohenberg model) 31 . In addition, from rough estimation with a linear fitting of H irr , the H irr (0) was estimated as 1.0 T.

Discussion
Suppression of resistivity anomaly by the Sn substitution. Here, we discuss the possible origin of the increase in T c by the Sn substitution. As revealed in the crystal structure part, the lattice parameters were not largely affected by the Sn substitution. Therefore, in-plane chemical pressure amplitude in the Bi-S superconducting plane, which has been revealed as the essential parameter for the emergence of superconductivity in BiS 2 -based compounds 24 , should not be significantly changed. Therefore, we consider that the in-plane chemical pressure effect is not the origin for the increase in T c by the Sn substitution. On carrier concentration, the absolute value of the Seebeck coefficient slightly decreases by Sn substitution for x = 0.1-0.4, which can be corresponding to the slight increase in electron carriers by Sn substitution. However, the large increase in T c for x = 0.4 may not be understood from the increase in carrier concentration only because the difference in carrier concentration  www.nature.com/scientificreports www.nature.com/scientificreports/ between x = 0.1 (T c = 0.6 K) and x = 0.4 (T c = 2.3 K) is expected to be quite small. On the basis of these facts, we briefly mention about the possible relation to the CDW ordering and the possible scenario of the suppression of CDW by Sn substitution in this system. In the ρ-T plots, an anomaly was observed in the La 2 O 2 Bi 3 Ag 1−x Sn x S 6 system. A similar feature in the normal-state resistivity has been observed in the EuFBiS 2 superconductor (T c = 0.3 K). The origin of the hump was proposed as a CDW transition 29 . We assume that the suppression of the CDW ordering is the origin for the increased T c in La 2 O 2 Bi 3 Ag 1−x Sn x S 6 . In addition, the anomaly temperature T * was shifted to a lower temperature by Sn substitution, and the anomaly disappeared at x = 0.3. At around x = 0.3 and 0.4, T c is the maximum. These facts imply that T c increased by the suppression of T * . Although we have no  www.nature.com/scientificreports www.nature.com/scientificreports/ evidence for the CDW states in the La 2 O 2 Bi 3 Ag 1−x Sn x S 6 system and the suppression mechanism by the Sn substitution, introduction of randomness at the M2 site may be effective to suppress the charge ordering states.
Bulk superconductivity in La 2 o 2 Bi 3 Ag 0.6 Sn 0.4 S 5.7 Se 0.3 . As shown in the Result part, a partial Se substitution for S induced bulk superconductivity in La 2 O 2 Bi 3 Ag 0.6 Sn 0.4 S 5.7 Se 0.3 . Although the solubility limit of Se for the S site is very low (5%), the lattice parameters clearly changed by the partial Se substitution, and the superconducting properties were significantly improved. Although we refined three models with different Se site (assuming the substitution at the S1, S2, or S3 sites), we could not find the site selectivity of doped Se. However, we expect that the doped Se occupies the Ch1 site in the inset of Fig. 6. In previous reports on the Se substitution in BiS 2 -based compounds, the site selectivity of Se at the in-plane site was observed 24,32,33 . According to the relationship between in-plane disorder at the chalcogen site and superconductivity in BiS 2 -based systems 16,24 , we assume that Se substitution reduced the in-plane disorder at the S1 site and induced bulk superconductivity. Furthermore, the room-temperature Seebeck coefficient for the La 2 O 2 Bi 3 Ag 0.6 Sn 0.4 S 5.7 Se 0.3 sample was similar to those shown in Fig. 4 (S = −25 μV/K). This also suggests that the bulk nature of superconductivity was induced by local structural optimization but not due to changes in carrier concentration.
Since the phases contain Sn in their structure, we have to exclude the possibility of superconductivity of elemental Sn in the examined samples. From XRD, no peaks of Sn was observed. However, amorphous Sn may exist and show superconducting transition. We cannot directly exclude this possibility, but we consider that the bulk superconductivity observed here is not originated from elemental Sn. We have investigated similar Sn substitution effects in the La 2 O 2 Bi 2 Pb 2−x Sn x S 6 system. La 2 O 2 Bi 2 Pb 2 S 6 (LaOBiPbS 3 ) shows an insulating transport property 26,27 . The insulating behavior was not suppressed by Sn substitution, and superconducting transition was not observed for all the examined samples of La 2 O 2 Bi 2 Pb 2−x Sn x S 6 . This fact may indirectly suggest that the bulk superconductivity observed in the La 2 O 2 Bi 3 Ag 0.6 Sn 0.4 S 5.7 Se 0.3 sample is not caused by Sn amorphous impurities.
In conclusion, we have reported the emergence of bulk superconductivity in La 2 O 2 M 4 S 6 -type (four-layer-type) layered oxychalcogenide La 2 O 2 Bi 3 Ag 0.6 Sn 0.4 S 5.7 Se 0.3 . According to the material design strategy shown here, we can further develop related four-layer-type Bi-based layered superconductors. Recently, a new superconductor Bi 3 O 2 S 2 Cl with the one-layer-type Bi-Cl-S superconducting layer was discovered by Ruan et al. 34 . In Fig. 8, schematic images of typical one-layer-type (Fig. 8a), two-layer-type (Fig. 8b), and four-layer-type (Fig. 8c) are displayed for comparison. All the materials have the similar RE 2 O 2 or Bi 2 O 2 blocking layer. By changing the constituent elements in the superconducting layers, the thickness can be changed in this superconductor family. On the basis of these facts, we will be able to design various types of Bi-based layered superconductors with a higher T c as developed with various Cu-oxide layers in the cuprate family. With such material development with a different type of conducting layer, remarkable changes in superconducting properties can be expected.   34 . In this material, the conducting layer can be regarded as the one-layer-type BiS 2 Cl. (b) Schematic image of the crystal structure of the typical BiS 2 -based superconductor RE(O,F)BiS 2 . In the series, the conducting layer can be regarded as the two-layer-type Bi 2 S 4 bilayer. (c) Schematic image of the crystal structure of La 2 O 2 Bi 3 AgS 6 -type (La 2 O 2 M 4 S 6 -type) materials whose conducting layer can be regarded as the four-layertype M 4 S 6 layer. La 2 O 2 Bi 3 Ag 0.6 Sn 0.4 S 6−z Se z , heating the sample to 725 °C in 1 h was needed to suppress the generation of impurity phases. The obtained samples were reground for homogeneity, pelletized, and heated in the same procedure. The phase purity of the prepared samples and the optimal annealing conditions were examined using X-ray diffraction (XRD) with a Cu-K α radiation. The lattice parameters were determined using the Rietveld method with RIETAN-FP 35 . Schematic image of the crystal structure was drawn using VESTA 36 . The actual composition was analyzed by energy-dispersive X-ray spectroscopy (EDX) on scanning electron microscope TM3030 (Hitachi). The magnetic susceptibility measurements were carried out using a superconducting quantum interference device (SQUID) magnetometer (MPMS-3, Quantum Design). The susceptibility data were taken after both zero-field cooling (ZFC) and field cooling (FC). The temperature dependence of electrical resistivity [ρ(T)] was measured by four-terminal method on the Physical Property measurement system (PPMS, Quantum Design). The resistivity measurement down to 0.4 K was measured using a 3 He probe platform of PPMS. The ADR system on PPMS was used for resistivity measurements down to 0.1 K. For clarity, we labeled the examined samples with the nominal compositions. The Seebeck coefficient was measured by a four-probe method on ZEM-3 (Advance RIKO) at 300 K.

Data Availability
The datasets generated and analyzed during the current study are available from the corresponding author on reasonable request.