Weak antiferromagnetism and superconductivity in pseudo-binary spinel compounds (Cu,Co)Co₂S₄investigated by⁵⁹Co and⁶³Cu magnetic resonance

Abstract

We have carried out a nuclear magnetic resonance study on pseudo-binary spinel compounds (Cu x Co 1−x )Co 2 S 4 with x = 0-1.0 to deduce the varied impact of variations in the antiferromagnetic (AF) spin correlation and density of states (DOS) that can clearly define the combination of magnetism and superconductivity in the system. The Curie-Weiss-type behaviour of the Knight shift K for both 59 Co and 63 Cu on the tetrahedral A site and the temperatureindependent K for 59 Co on the octahedral B site indicate that the magnetism of the system originates from 3d bands associated with the transition-metal elements on the A site. With the Cu substitution x for Co on the A site, the negative Weiss temperature θ deduced from the K data initially decreases, takes a deep minimum around x = 0.7 followed by a rapid increase for x → 1. For the compounds with x 0.3 and x > 0.8, the nuclear spin-lattice relaxation rate T −1 1 for the nuclei on both A and B sites has a T 1/2 dependence at high temperatures, which is characteristic of three-dimensional itinerant weak antiferromagnets. In the Cu-rich region (x > 0.7) the compounds transform into a superconducting state below T S with no long-range magnetic ordering. From a T −1 1 data analysis, we find that the increase of T S correlates not only with the development of the AF spin correlation but also with the significant increase in the DOS at the Fermi level of the 3d bands N d (E F ) associated with Co on the B site. T −1 1 in the superconducting state of CuCo 2 S 4 (T S = 4.4 K), measured for 59 Co on the B site utilizing the pure quadrupole resonance spectrum, has a coherence peak followed by an exponential decrease, indicating that CuCo 2 S 4 is an s-wave superconductor with an energy gap of 2 = 4.14k B T S . We conclude that the appearance of superconductivity for the compounds in the Cu-rich region originates from the large increase in N d (E F ) associated with Co

Superconducting Transition Temperatures of Cu(Co1-xMx)2S4

“…The existence of the CoS 2 phase was also noted in the previous works, 1,2) indicating that the synthesis of single phase samples of the present system is not easy. In their studies, the superconducting transition was not found for samples with the Cu/Co molar ratio of 1 : 2.…”

“…Due to this structural characteristic, electrons in these compounds are often frustrated. One of the spinel system CuCo 2 S 4 exhibits the superconducting transition at T c $ 4:4 K. 1,2) For this system, a significant peak has been reported in the T dependence of the nuclear spin relaxation rate divided by T, 1=T 1 T of 59 Co, indicating that it has an s-wave symmetry of the order parameter. 3) Na z CoO 2 .…”

Electronic environments in carrollite, CuCo2S4, determined by soft X-ray photoelectron and absorption spectroscopy

“…However, that d 9 character might explain the previously observed 63 Cu NMR data. Neither the Co L 2,3 absorption spectrum nor the Co 2p photoelectron spectrum indicated a definitive formal oxidation state for the Co, but the 59 Co NMR data were consistent with a non-magnetic state for that element (Wada et al, 2002), implying low-spin Co III rather than Cu I (Co 2 ) VII S -II 4 or Cu I (Co 2 ) V (S 4 ) -VI . The S 2p spectra provided no evidence for S in more than one formal oxidation state in the carrollite bulk, therefore the question arises as to whether the S formal oxidation state is 12.5% less negative than S -II and the mineral can be represented by Cu I Co III 2 (S 4 ) -VII .…”

“…While accepting the reasoning that led to the deduction of d 6.6 , as noted in Sections 3.1 and 3.3, we considered that neither the Co 2p spectrum nor the Co L 2,3 -edge spectrum provided definitive information on the oxidation state of Co in carrollite. The information that did seem definitive was the 59 Co NMR behaviour that indicated a non-magnetic state (d 6 ) for the Co atoms (Wada et al, 2002). This appeared to preclude not only a representation of Cu I (Co 2 ) V (S 4 ) -VI , close to that proposed by Pattrick et al (2008), but also the otherwise possible representation of Cu I (Co 2 ) VII S -II 4 , which would have incorporated the oxidation state of S -II typically allocated to S in metal sulfides.…”

“…The Cu II description has either been assumed or proposed on the basis of several different experimental measurements on synthetic and natural specimens, including magnetic behaviour (Miyatani et al, 1993;Furukawa et al, 1995), 63 Cu and 59 Co NMR spectroscopy (Sugita et al, 2000;Wada et al, 2002) and X-ray absorption spectroscopy (van der Laan et al, 1992). The X-ray absorption spectroscopic evidence consisted of a Cu L 3 -edge absorption peak at 930.5 eV, 0.7 eV lower than the corresponding peak for Cu II oxide (tenorite, CuO), and a clear indicator of a major contribution of Cu 3d 9 to the ground state.…”