Two electronic orders respond differently to one tuning knob
The paper's central result is a split response, not a single collapse. Bilayer nickelates host a spin-density-wave (SDW) order, a periodic modulation of spins detected here by resonant elastic X-ray scattering (REXS), and a second order seen in the σ-polarization channel that the authors call a charge-like anisotropy order. Enhancing interlayer coupling weakens the first but erases the second.
That distinction matters because earlier work on these films found that oxygen annealing reduced the SDW peak intensity in REXS while leaving the SDW transition temperature largely unchanged. Here the transition temperature itself moves, and the two orders separate: the SDW survives at a reduced temperature scale, whereas the charge-like order has no resolvable signal in the superconducting films. The authors therefore argue the two orders are not governed by a single energy scale.
Compressing the c-axis and filling interlayer oxygen vacancies
The study uses two independent routes to the same physical parameter. The first is isovalent rare-earth A-site substitution: replacing La with smaller Pr or Sm compresses the c-axis, which the authors describe as effectively applying c-axis uniaxial pressure. The second is ozone-assisted annealing, which fills interlayer oxygen vacancies and, according to the polarization-resolved O K-edge X-ray absorption data, restores the 3d_z2–2p_z–3d_z2 hybridization channel between the two NiO2 layers.
Films of La2PrNi2O7−δ, La3Ni2O7−δ and La2SmNi2O7−δ were grown by pulsed laser deposition on LaAlO3 and SrLaAlO4 substrates, with thicknesses from about 5.5 nm to 28 nm. The thicker LPNO films on LAO remained non-superconducting even after ozone annealing, while the thinner LPNO films on SLAO became superconducting after annealing. That difference is what allows the paper to compare non-superconducting and superconducting films of the same nominal composition.
The SDW temperature falls from 150 K to 70 K
In the π-polarization channel, which the authors associate with the magnetic SDW, the transition temperature decreases systematically as interlayer coupling is strengthened. For LPNO films on LAO, ozone annealing shifted T_SDW from about 150 K to about 130 K. For the thinner LPNO films on SLAO, the shift was larger: from roughly 150 K in the non-superconducting film to about 110 K in the superconducting one. Across the superconducting LNO, LPNO and LSNO series, the SDW peak intensity decreased progressively from undoped LNO to Pr-doped LPNO and further to Sm-doped LSNO, and the transition temperature fell to 70 K in the most strongly tuned case.
The order does not vanish. Residual SDW intensity persists in the superconducting films, and the paper describes the SDW as weakened but remaining. Readers should keep the two quantities apart: peak intensity and transition temperature both fall, but neither measurement alone establishes that the SDW is gone.
The charge-like anisotropy order leaves no resolvable signal
The σ-polarization channel behaves qualitatively differently. In as-grown films it shows a non-monotonic temperature profile: intensity rises on cooling and then drops sharply below about 110 K. After ozone annealing this non-monotonic feature disappears, and in the superconducting LPNO films on SLAO no discernible peak appears at any measured temperature. In the A-site series, only the parent LNO film retains a weak but discernible σ-channel peak, while the Pr- and Sm-doped samples show no resolvable signal above background.
This is the sharper half of the selective suppression. The paper's interpretation is that the charge-like order is tied to a magnetic quadrupole on the spinless sublattice of a spin–spinless stripe state, and that removing this anisotropic local environment weakens the SDW correlations that the quadrupole helps lock in place. That mechanism is the authors' framework, not a directly imaged structure.
Fermi-surface reconstruction accounts for the SDW side only
Polarization-resolved O K-edge absorption shows increased spectral weight of the interlayer bonding γ band near the Fermi level after coupling is strengthened. Density functional theory calculations, performed with a reduced c-axis to model A-site substitution, show a reconstructed Fermi surface and a suppressed electronic susceptibility near the SDW wavevector. The authors offer two readings of the same spectral-weight transfer: in an itinerant picture, altered nesting between the α and β pockets weakens the SDW instability; in a strong-correlation picture, delocalization of Ni d_z2 orbitals undermines the local moments that stabilize static SDW order.
Both readings explain why the SDW weakens. Neither explains why the charge-like order disappears completely, which is why the paper concludes that band theory alone is insufficient for the σ-channel and that electronic correlations are essential there.
A constraint on density-wave origins, not a superconductivity recipe
The value of the result lies in the comparison it forces. Because one tuning parameter suppresses the two orders to different degrees, any account that treats them as a single density-wave instability faces a constraint. The paper states this directly: the selective suppression demonstrates that the two orders are not governed by a single energy scale and provides a new constraint on their origins and on how they relate to superconductivity in bilayer nickelates.
The scope is equally specific. The superconducting comparisons come from thin films on SLAO after ozone annealing, while the thicker LAO-grown films stayed non-superconducting; the A-site series covers LNO, LPNO and LSNO. The SDW magnetic structure itself remains under debate in the literature cited by the paper, and the spin–spinless stripe picture is the framework the authors adopt to interpret their data rather than a settled structure. The paper reports consistency with that ground state, not a direct determination of it.
Read the two polarization channels as separate verdicts
For anyone assessing these materials, the practical lesson is to treat the π and σ channels as answering different questions. A surviving π-channel SDW at reduced T_SDW does not imply that the charge-like order also persists; in the superconducting films reported here, the σ-channel signal is absent while the SDW remains. Conversely, the disappearance of the charge-like order is not evidence that the SDW has been eliminated.
The paper's own priority follows from that split: interlayer coupling is identified as the central parameter controlling both orders, with the SDW side explicable through Fermi-surface reconstruction and the charge-like side requiring correlation physics beyond band theory. The most consequential number for comparison across samples is the SDW transition temperature, which the authors move from 150 K down to 70 K while keeping the order present.
Sources used in this answer
Selective suppression of electronic orders via interlayer coupling in superconducting bilayer nickelate thin films
Combining rare-earth A-site substitution with ozone annealing strengthens interlayer coupling in bilayer nickelate films, lowering the spin-density-wave transition temperature from 150 K to 70 K while the SDW persists and completely eliminating a second charge-like anisotropy order in the superconducting state.