Asymmetry of collective excitations in electron- and hole-doped cuprate superconductors

Abstract

High-temperature superconductivity (HTSC) mysteriously emerges upon doping holes 1 or electrons 2 into insulating copper oxides with antiferromagnetic (AFM) order. It has been thought that the large energy scale of magnetic excitations, compared to phonon energies for example, lies at the heart of an electronically-driven superconducting phase with high transition temperatures (T c ) 3-5 . Comparison of high-energy magnetic excitations of hole-and electron-doped superconductors in connection with the respective T c provides an exceptional, yet un-capitalized opportunity to test this hypothesis 6-9 . Here, we use resonant inelastic x-ray scattering (RIXS) at the Cu L 3 -edge 10,11 to reveal high-energy collective excitations in the archetypical electron-doped cuprate Nd 2-x Ce x CuO 4 (NCCO) 2 . Surprisingly, despite the fact that the AFM correlations are short-ranged 12 , magnetic excitations harden significantly across the AFM-HTSC phase boundary, in stark contrast with the hole-doped cuprates 6,7 . Furthermore, we find an unexpected and highly dispersive branch of collective modes in superconducting NCCO that are absent in hole-doped compounds. These modes emanate from zone center and weaken with increasing temperature, which signal a quantum phase distinct from superconductivity. The asymmetry uncovered between electron-and hole-doped cuprates provides new, unexpected dimensions to collective excitations that are generally important to the mechanism of superconductivity in these materials. Hole-doped cuprates display compelling evidence for the surprising persistence of magnetic excitations beyond the AFM phase boundary 6,7 , as well as the existence of symmetrybroken phases, such as charge density waves 13,14,15 and orbital loop currents 16 , distinct from superconductivity. Whether these are universal and exist on the other side of the cuprate phase diagram, i.e. with electron-doping, remains an important open question. To address this issue,

Spin and charge excitations in artificial hole- and electron-doped infinite layer cuprate superconductors

“…The charge mode irradiates from the Γ point and moves rapidly towards higher energy with increasing q . As mentioned above, similar excitations have been observed in RIXS measurements on the prototypical e-doped cuprate NCCO 11,12 , with the same steep dispersion we find in SLCO, as shown in Figure 5 (b).…”

“…11,12 , at lower q (< 0.18 r.l.u.) the non-elastic spectral weight was decomposed into two non-resolution-limited components: a magnetic peak and a charge feature.…”

“…It has thus been shown that paramagnons persist throughout the phase diagram, from superconducting underand optimally-doped [7][8][9] to non-superconducting highly-overdoped 10 samples. Paramagnons have also been observed in the prototypical electron-doped (e-doped) cuprate compound Nd 2−x Ce x CuO 4 (NCCO) 11,12 . These findings have triggered several works 7,[13][14][15] , aiming to show that paramagnons are the driving excitations in Cooper pair formation.…”

“…These findings have triggered several works 7,[13][14][15] , aiming to show that paramagnons are the driving excitations in Cooper pair formation. However there is no obvious correlation between the critical temperature (T c ) and the paramagnon energy, and this may indicate that also other low energy excitations play a role, such as phonons (via electron-phonon coupling) and charge modes 11,12,16 . It is then worth studying the low energy excitations in cuprate compounds with different structure, in order to identify their correlations with high temperature superconductivity.…”

“…Experiments show that also the paramagnon excitations have a different doping evolution upon hole and electron doping: for h-doped compounds the dispersion is almost unaffected by doping, while for e-doped ones the dispersion is doping dependent, with hardening of the paramagnon mode up to 50% 11,12 . This asymmetry can be correctly captured by a relatively simple model including three-site exchange 19 , even though there is no perfect agreement with the experimental dispersion data 11,12 .…”