Magnetism in Parent Iron Chalcogenides: Quantum Fluctuations Select Plaquette Order

Abstract

We analyze magnetic order in Fe-chalcogenide Fe1+yTe -the parent compound of hightemperature superconductor Fe1+yTe1−xSex. Experiments show that magnetic order in this material contains components with momentum Q1 = (π/2, π/2) and Q2 = (π/2, −π/2) in Fe-only Brillouin zone. The actual spin order depends on the interplay between these two components. Previous works assumed that the ordered state has a single-Q (either Q1 or Q2). In such a state, spins form double stripes along one of diagonals breaking the rotational C4 symmetry. We show that quantum fluctuations actually select another order -a double Q plaquette state with equal weight of Q1 and Q2 components, which preserves C4 symmetry. We argue that the order in Fe1+yTe is determined by the competition between quantum fluctuations and magnetoelastic coupling.Introduction. The analysis of magnetism in parent compounds of iron-based superconductors (FeSCs) is an integral part of the program to understand the origin of superconductivity in these materials [1][2][3][4][5][6][7][8][9][10][11][12]. Parent compounds of Fe-pnictides are moderately correlated metals, [5,13] whose magnetic order can be reasonably well understood within itinerant scenario [7-9, 14, 16] The locations and the shapes of the Fermi surfaces (FSs) select two possible ordered state with momenta (0, π) and (π, 0)-in the Fe-only Brillouin zone (BZ) [9].In each of these two states spins are ordered in a stripe fashion -ferromagnetically along one direction in 2D Feplane and antiferromagnetically in the other. Such an order breaks C 4 lattice rotational symmetry and causes pre-emptive spin-nematic order [15]. The same magnetic order is selected in the strong coupling approach, based on J 1 − J 2 model of localized spins with nearest and second-nearest neighbor spin exchange [17,18]. The actual coupling in Fe-pnictides is neither truly small nor strong enough to cause Mott insulating behavior [13], which makes it extremely useful that the two descriptions agree.There is one family of FeSCs -11 Fe-chalcogenides Fe 1+y Te 1−x Se x , in which smooth evolution between parent and optimally doped compounds does not hold. Magnetism in these materials changes considerably between x = 0 and x ∼ 0.5, where the T c is the largest. Near optimal doping magnetic fluctuations are peaked at or near (0, π) and (π, 0), as in Fe-pnictides, while magnetic order in a parent compound Fe 1+y Te has very different momenta ±(π/2, ±π/2) [19][20][21][22][23]. Upon doping, the spectral weight at ±(π/2, ±π/2) decreases, and the spectral weight at (0, π) and (π, 0) increases [20]. The transport properties of Fe 1+y Te are also quite different from those of parent compounds of Fe-pnictides: the resistivity, ρ(T ), of Fe 1+y Te does not show a prominent increase with increasing T , but instead remains flat and even shows a small increase as T decreases [24]. ARPES studies of Fe 1+y Te show that low-energy spectra are very broad [25], consistent with the notion that electrons are not propagating. These observations lead several

Order on Command

“…In contrast, our SPSTM images with a spin-polarized Cr tip show (above a small bias threshold [∼30 meV, [8].…”

“…DOI: 10.1103/PhysRevLett.119.227001 Iron-based superconductors (FeSCs) have shown intriguing phenomena related to the coexistence of magnetism and superconductivity below the superconducting transition temperature (T c ) [1][2][3]. Although an understanding of their detailed interplay is still under debate, certain magnetic orders seem to be very crucial in realizing coexistent superconductivity [3][4][5][6][7][8][9][10][11][12][13][14][15]. Recent studies have shown new reentrant C 4 symmetric antiferromagnetic phases (C 4 magnetism from now on) coexisting with superconductivity and have reported that the superconducting T c is suppressed by C 4 magnetic order [16][17][18][19].…”

“…4(a)] with a bias condition exceeding the threshold V N th , simulating thermal annealing in this area. Figure 4 [8], shown together with the ARPES-based Fermi surfaces (dark curves) [34]. PRL …”

Emergent Ising degrees of freedom above a double-stripe magnetic ground state

“…We can see that fluctuations about the Q α cost energy via J 2 and J 3 , as expected, while J 1 drives the system away from these states (towards a spiral state, as it turns out) [29]. In the following, we set J 1 = 0.…”

“…Since the exchange fields due to both J 1 and J 2 cancel out at each site, the four sublattices are decoupled in the classical ground state. J 1 drives the classical ground state into a spiral state and away from DS magnetism [29], so we neglect J 1 in this paper, which is valid for sufficiently large four-spin interactions. As in the SS case, four spin interactions are required to couple together the four sublattices, which requires not only K 1 but also K 2 , the NNN biquadratic coupling.…”