Strain tuning: calibrated deformation and superconducting criteria

Scope

Strain is a directional deformation relative to a stated reference structure. A uniaxially loaded crystal, a clamped epitaxial film and a hydrostatically compressed sample follow different mechanical paths. Applied displacement is not necessarily sample strain: mounting adhesive, thermal contraction, bending and fracture can all intervene. The useful control variable is the calibrated deformation in the region probed, together with its uncertainty and history. In the ideal uniaxial-stress geometry, transverse Poisson strains remain nonzero: controlling one loading axis does not set every other strain component to zero. A displacement sensor also does not directly measure the strain throughout a specimen. R1 R3

Separate the early Sr2RuO4 experiments. Hicks's 2014 Science study reported different responses for loading along the crystallographic ⟨100⟩ and ⟨110⟩ directions, within an approximately 0.23% accessible range. The later Steppke study reached the much larger compressive strain associated with a transition near 3.4 K. These are distinct measurements, not interchangeable references for one apparatus or peak. The earlier result is checked here only through the authors' institutional abstract. R2 R3

Mathematical assumptions

For a lattice parameter relative to a relaxed reference, engineering strain is $\epsilon=(a-a_0)/a_0$. Nominal epitaxial mismatch uses $f=(a_{\rm sub}-a_{\rm ref})/a_{\rm ref}$ instead. Neither a substrate name nor this calculation establishes the strain remaining after relaxation. Report reference composition, phase, temperature and orientation; compare actual film lattice parameters with that same reference. R4

In small-deformation elasticity, $\epsilon_{ij}=(\partial_i u_j+\partial_j u_i)/2$ and $\sigma_{ij}=C_{ijkl}\epsilon_{kl}$. As an explicitly isotropic illustration, a film with equal in-plane strains $\epsilon_{xx}=\epsilon_{yy}=f$ and a traction-free surface has $\epsilon_{zz}=-2\nu f/(1-\nu)$. Positive Poisson ratio therefore permits out-of-plane expansion under in-plane compression. This illustration is not an anisotropic nickelate fit, and it does not convert epitaxial mismatch into an equivalent hydrostatic pressure. R1 R4

Hicks's simplified adhesive model gives a load-transfer length $\lambda=\sqrt{Etd/(nG_e)}$, with sample Young's modulus $E$, sample thickness $t$, adhesive thickness $d$, adhesive shear modulus $G_e$, and one or two bonded faces $n$. For symmetric ends, displacement is distributed over approximately $L+2\lambda$, not just the exposed length $L$. With $E=182$ GPa, $t=50$ micrometres, $d=10$ micrometres, $n=2$ and $G_e=6$ GPa, the model gives 87 micrometres, consistent with the paper's approximately 90 micrometres. For an illustrative 1 mm exposed length, dividing displacement by $L$ alone overestimates the model's central strain by about 17% relative to that strain. This is local arithmetic, not a recalibration of Steppke's samples. Rigid plates, much softer adhesive, long bonded regions and uniform strain across the sample cross-section are assumptions. R1

The same author manuscript's Appendix A contains a numerical discrepancy worth preserving. Using its printed Eq. A1 and the parameters above, a 0.2% sample strain gives a displacement-over-glue-thickness shear of approximately 1.74% and a shear stress of 105 MPa; the adjacent prose instead reports 1.3% and 80 MPa. The scalar shear here follows that equation's convention, not the half-sized off-diagonal strain-tensor component. This substitution check does not establish a formal correction or a failure threshold. The paper itself notes uncertain low-temperature adhesive properties; room-temperature tensile strength is not a measured cryogenic shear-strength limit. R1

Experimental criteria

Steppke measured magnetic transitions using mutual inductance, with approximately 0.2 Oe excitation at 1–20 kHz, and used a capacitive displacement sensor. Figure 4 defines the plotted transition temperatures at 50% of the magnetic response; the 20% and 80% levels show transition width. The reported peak strain is approximately −0.60% with a ±0.06 percentage-point uncertainty from four samples, while individual strain calibration was estimated to have roughly 20% uncertainty. These are different uncertainty statements. R3

The often-quoted 3.4 K scale is not a universal sample maximum: sample 5 reached approximately 3.5 K. Resistive transitions in the two contacted specimens did not exceed the highest magnetic transition by more than 0.08 K. That comparison does not redefine a magnetic midpoint as zero resistance. A flat feature near zero strain in sample 1 arose after the sample broke during cooling and needed compression to re-engage; it should not be fitted as an intrinsic phase boundary. Reported recovery of magnetic curves after unloading other checks supports reversible deformation under those conditions, not a guarantee that every mounting survives or that thermodynamic equilibrium has been established. The critical-field measurements on sample 4 use intersections of linear fits around a susceptibility cusp; that separate procedure must not replace the 50% criterion for the Figure 4 temperature-versus-strain data. R3

Bhatt's final Figure 1 labels nominal/measured mismatch pairs, respectively, as SLAO −2.0%/−1.6%, LAO −1.2%/−0.9%, NGO +0.7%/+0.6% and STO +1.9%/+1.8%. These final labels replace several different measured values in the inherited chapter. The SLAO film has a broad resistive transition, with onset near 42 K and zero near 2 K; those endpoints cannot be merged into a 42 K zero-resistance claim. R4

Ko's publicly archived author manuscript now supplies selected original methods and criteria. It reports sample A1 with onset at 42.4 K and zero resistance below 2.2 K; the latter is not an exact 2.2 K transition value. Its field-dependent analysis instead uses 90% of the normal-state resistance near onset. The methods describe approximately 100 kHz low-pass filtering; the text warns that observing zero resistance requires adequate radiation filtering. The reported diamagnetic response belongs to sample A12, so it must not be attached to A1 as a same-specimen measurement. The March 2026 correction fixes published resistivity-axis units to mΩ cm and the Figure 3d horizontal label to temperature $T$, not $T_c$. The archived manuscript already has mΩ cm but still labels that axis $T_c$; it is not a verified copy of the corrected journal version. Existing rounded film records retain Bhatt's final passage, while the added A1 onset record cites the author manuscript. No raw-data refit or full final-version comparison is claimed. R4 R5 R6

The comparisons below retain the measured quantity and specimen history; none supplies a universal conversion from applied strain to superconducting temperature.

Limits and disputes

Bhatt's films on SLAO used pulsed-laser deposition, whereas the other substrate films used molecular-beam epitaxy. Oxygen annealing and uncertain oxygen occupancy remain relevant covariates. Figure 3 supports out-of-plane expansion under in-plane compression, contrasting with hydrostatic contraction. It does not support the inherited universal superlative that the SLAO film has the largest c-axis: the plotted LAO values are comparable or higher. Avoid turning a qualitative structural contrast into an unsupported ranking. R4

A peak near a calculated Lifshitz transition constrains possible models without uniquely identifying pairing symmetry. Steppke's electronic-structure and order-parameter calculations are model-dependent; their discussion should not be read as a present-day determination of Sr2RuO4's order parameter. This scoped review checked selected experimental passages and a figure, not the full multiband calculation, all supplements or experimental raw data. R3

Discovery implications

Design a strain study around a measured mechanical path and a defined superconducting criterion. Retain loading direction, reference state, adhesive or substrate constraints, thermal history and reversibility. Check whether a broadened transition tracks strain inhomogeneity or a changed specimen, and compare transport with magnetic or thermodynamic evidence where feasible. Local arithmetic here verifies mismatch, ideal elastic identities and the adhesive model; it neither predicts a superconducting transition nor establishes a unique strain mechanism. R1 R3 R4

Related topics

Compare pressure tuning, chemical doping, electrostatic gating, nickelates and two-dimensional interfaces.

Material cases

Values describe individual reports, not a ranking. Unknown means the checked source did not supply the field at this review scope.