Spin pseudogap in the S=12 chain material Sr2CuO3 with impurities

Abstract

The low energy magnetic excitation spectrum of the Heisenberg antiferromagnetic S = 1/2 chain system Sr2CuO3 with Ni-and Ca-impurities is studied by neutron spectroscopy. In all cases, a defect-induced spectral pseudogap is observed and shown to scale proportionately to the number of scattering centers in the spin chains.

Dynamic structure factor in impurity-doped spin chains

“…Both E 1 and E 2 become exponentially small for energies below the averagelength gap ω ≪ πv/ L ≡ pπv. The suppression of bulk spectral weight with E 1 due to the finite size gaps was discussed and observed experimentally [24,25], but we find that the additional redistribution of spectral weight becomes very important, which can be traced to the effect of boundary correlations. The rescaled average ω 2−2K Sp from Eq.…”

“…However, surprisingly a systematic analysis of the doping effects on the energy and momentum resolved dynamic structure factor is still missing so far. Previous research has taken into account the discrete spectrum of finite chains [24,25,36], which leads to an exponential suppression at low energies [24,25]. The understanding of the momentum dependence is more involved, however, since characteristic correlations near the impurities play an important role and lead to a strong redistribution of spectral weight to higher momenta outside the dispersion relation as shown in this paper.…”

“…in describing exact form factors [2][3][4][5][6], exact correlations [7], nonequilibrium states [8,9], and dynamic correlations in the regime of a non-linear spectrum [10][11][12][13][14][15][16][17][18]. Recently, there has been renewed experimental interest in intentionally doped spin chain systems [19,20] with new results on the Knight shift [21,22], magnetic ordering [23], and the dynamic structure factor [24][25][26][27]. Doped spin chains are known to acquire characteristic boundary correlation functions [28], which lead to impurity induced changes in the Knight shift [29,30], the susceptibility [31][32][33], the static structure factor [30], and the ordering temperature [34,35].…”

“…(67)upon using the substitution L = πvm ν and dL = −dν mπv ν 2 . HereE 1 (y) = m my 2 e −my = y 2 e y (e y − 1) 2 (68)is the Einstein function of the scaling variable y = pπv/ω which measures the "average-length" gap vπ/ L compared to ω[24,25]. For the average impurity correction we use the same substitution L = πvm ν and dL = −dν mπv /ν S imp (ω)δ(ω − ν) /ω S imp (ω) (71)= pE 2 (πvp/ω)S imp (ω).…”