The in-plane resistivity r a ͑T ͒ and the out-of-plane resistivity r c ͑T ͒ have been systematically measured for Bi 2 Sr 2 CaCu 2 O 81d single crystals with their oxygen contents precisely controlled. In the underdoped region, deviation from T-linear in-plane resistivity, which evidences the opening of the "spin gap," is clearly observed, while the out-of-plane resistivity is well reproduced by the activationtype phenomenological formula r c ͑T ͒ ͑a͞T ͒ exp͑D͞T͒ 1 c. In contrast to the YBa 2 Cu 3 O 72d system, we find that the onset of the semiconducting r c ͑T ͒ does not coincide with the opening of the spin gap seen in the r a ͑T ͒ in this Bi 2 Sr 2 CaCu 2 O 81d system. [S0031-9007(97)04015-5] PACS numbers: 74.25.Fy, 74.62.Dh, 74.72.Hs It is well recognized that an understanding of the characteristic electronic phase evolution with carrier doping is important to elucidate the high-T c superconductivity. One striking phenomena is a "spin gap" state typically observed in underdoped YBa 2 Cu 3 O 72d by various techniques [1] including resistivity measurements [2]. Recently, angle-resolved photoemission spectroscopy (ARPES) [3,4] has directly revealed the normal state pseudogap in an underdoped Bi 2 Sr 2 CaCu 2 O 81d system and pointed out its intimate relation with superconductivity (i.e., d x22y2 symmetry). Up to now, however, there have been few transport studies [5,6] of the underdoped state (and thus a spin gap effect) in the Bi 2 Sr 2 CaCu 2 O 81d system. Another striking feature is that a semiconductive out-of-plane resistivity ͑dr c ͞dT , 0͒ coexists with metallic in-plane resistivity r a over a wide temperature and carrier doping range [7]. Among the many ideas on the origin of the semiconductive r c , there is a theoretical suggestion [8] that it is caused by the opening of the spin gap in the context of spin charge separation. Experimental results on YBa 2 Cu 3 O 72d [9] seem to be consistent with the theory. However, the mechanism for the out-of-plane conduction is still in dispute [10].The Bi 2 Sr 2 CaCu 2 O 81d system is a good choice for studying such an electronic phase diagram, because its doping level can be controlled over a wide range by varying the oxygen content. But we encounter several difficulties in controlling the doping level, especially when we go into the underdoped side of the system. One is oxygen inhomogeneity, introduced when a sample is cooled down after an anneal, which causes a broadening of a superconducting phase transition and obscures the intrinsic properties of the system. If samples are rapidly cooled to avoid this problem, high-temperature disorder is frozen. Herein lies the second difficulty. The disorder causes electron localization at low temperatures. The third difficulty is related to the stability of the material. Annealing at high temperatures ͑$600 ± C͒ and low oxygen pressures ͑#10 25 atm͒ is needed for realizing the under-doped state of the Bi 2 Sr 2 CaCu 2 O 81d system. However, these conditions sit close to the decomposing line [11]. We solved all of the a...

No abstract

We report a systematic high-resolution angle-resolved photoemission spectroscopy on high-T(c) superconductors Bi(2)Sr(2)Ca(n-1)Cu(n)O(2n+4) (n=1-3) to study the origin of many-body interactions responsible for superconductivity. For n=2 and 3, a sudden change in the energy dispersion, so called "kink", becomes pronounced on approaching (pi,0) in the superconducting state, while a kink appears only around the nodal direction in the normal state. For n=1, the kink shows no significant temperature dependence even across T(c). This could suggest that the coupling of electrons with Q=(pi,pi) magnetic mode is dominant in the superconducting state for multilayered cuprates, while the interactions at the normal state and that of single-layered cuprates have a different origin.

Tunneling spectroscopy using a very thin stack of intrinsic Josephson junctions has revealed that the superconducting gap is definitely different from the pseudogap in the Bi(2)Sr(2)CaCu(2)O(8+delta) system. In the underdoped region, the conductance peak arising from the superconducting gap is independently observed in the dI/dV-V curve and its position is much lower than that of the pseudogap. Near the optimum doping level and in the overdoped region, both peaks are located in close proximity. These findings are in conflict with a previous understanding of the pseudogap.