WorksPhysics and Astronomy
Controllable suppression of the unconventional superconductivity in bulk and thin-film SrRuO via high-energy electron irradiation
Abstract
In bulk SrRuO, the strong sensitivity of the superconducting transition temperature to nonmagnetic impurities provides robust evidence for a superconducting order parameter that changes sign around the Fermi surface. In superconducting epitaxial thin-film SrRuO, the relationship between and the residual resistivity , which in bulk samples is taken to be a proxy for the low-temperature elastic scattering rate, is far less clear. Using high-energy electron irradiation to controllably introduce point disorder into bulk single-crystal and thin-film SrRuO, we show that is suppressed in both systems at nearly identical rates. This suggests that part of in films comes from defects that do not contribute to superconducting pairbreaking, and establishes a quantitative link between the superconductivity of bulk and thin-film samples.
Introduction
Conclusive identifications of the order parameter and of the interactions that induce superconductivity in have remained elusive despite receiving intense research interest for more than 25 years [1, 2]. Among the first evidence for the unconventional nature of superconductivity in was the observation that the transition temperature could be completely suppressed with a minute concentration of nonmagnetic impurities [3], indicating that the phase of the superconducting order parameter changes sign in momentum space and, with sufficient scattering, averages to zero [1]. Subsequent studies confirmed this extreme sensitivity to disorder, either by introducing variable defect densities via intentional chemical substitutions during growth [4, 5, 6] or by taking advantage of native impurities and structural defects that exist in all nominally stoichiometric crystals [7, 8].
While the basic fact of the unconventionality is well established, determining the specific symmetry of the order parameter in continues to be challenging, not least because of the intrinsically low energy scales of the problem, with K. In principle, using epitaxial thin films of as a materials platform would offer a new angle of attack with numerous potential advantages: films can be patterned using standard optical lithography into devices and junctions for phase-sensitive magnetization and electrical transport measurements [9, 10], and are naturally compatible with sophisticated scanned-probe microscopies [11, 12] and spectroscopic techniques [13, 14]. Indeed, in the cuprate high-temperature superconductors, scanning SQUID microscopy of thin-film samples provided conclusive evidence of the pairing symmetry [15].
The strong sensitivity of to disorder meant that it required many years of effort following the initial discovery of superconductivity in single crystals [16] for the materials science community to realize epitaxial thin films of that were electronically clean enough to exhibit superconductivity [17]. Now that superconducting thin-film samples can be reproducibly synthesized [18, 19, 20, 21], there is an urgent need to establish a quantitative link between the physical properties of single crystals and the best thin films. The essential challenge in doing so is that the correlation between and the residual resistivity —taken in bulk samples to be a proxy for the low-temperature elastic scattering rate—of as-grown films is not nearly as robust as that observed in single crystals.
One hypothesis to explain the difference between the two forms of is that epitaxial thin films contain extended in-plane defects, such as small-angle grain boundaries, that increase the electrical resistivity but have a much smaller effect on the intrinsic intra-grain quasiparticle scattering rate and hence on . To test this hypothesis would require a means of introducing point disorder to such films while, hopefully, generating as small a change as possible to the resistivity of the extended defects.
Here, we use high-energy electron irradiation to controllably introduce point disorder into bulk single-crystal and thin-film SrRuO. We establish that the disorder from irradiation suppresses in bulk SrRuO in the same way as chemical substitution or native point defects, but with far greater control over the disorder axis. We then demonstrate that thin films respond to irradiation-induced disorder in essentially the same way as bulk samples, with being suppressed in crystals and films at nearly identical rates.
Methods
FIB-sculpted single crystals of SrRuO
To prepare samples of bulk, single-crystal SrRuO that were compatible with high-energy electron irradiations and that enabled reliable measurements of the low-temperature resistivity, we used a modified version of the epoxy-free method of mounting and preparing microstructures described in Ref. [22]. Procedures for the growth of the SrRuO crystals are described in Ref. [23].
From an oriented slice of the parent crystal, we cleaved a piece of SrRuO having -oriented edges to expose a flat, (001)-oriented surface which later became the top surface of the finished microstructure. To make electrical contact, we sputtered 200 nm of gold onto the cleaved surface and annealed the crystal for 5 to 6 minutes at 500°C in air. We then used a focused ion beam (FIB) to extract lamellae (typical dimensions about 200 m 160 m 6 m), working from the uncleaved side inwards, and transferred these lamellae ex situ onto mica substrates with the Au-coated surface facing upwards. After depositing 10 nm of Ti and another 150 nm of Au onto each lamella, we deposited Pt bridges in a xenon plasma-FIB to join the top of each lamella to the mica substrate. We then deposited a final 10 nm Ti/150 nm Au to ensure good electrical contact between the lamella and the gold film on the mica substrate.
After mounting and contacting the lamellas, we used a gallium FIB to pattern each into a meandering microstructure suitable for four-point resistance measurements, as shown in Figure 1(a). From the dimensions of each microstructured SrRuO bar [24], we calculated the appropriate geometrical factors to convert the measured resistances into resistivities. Such FIB-sculpted bars of SrRuO have a well-defined shape and the positions of the voltage-sensing contacts on the crystal are independent of the wires and epoxy used for connecting the sample to external electronics equipment in subsequent resistance measurements. The latter point is important because the wires attached to the mica substrate must be removed and remade every time the sample is transferred between the cryostat used for low-temperature electrical transport experiments and that used for electron irradiation experiments. Having a stable, reproducible geometry for the resistance measurement is essential for detecting small, irradiation-induced changes in the absolute resistivity.
Figure 1. Effects of high-energy electron irradiation on the electrical resistivity and superconducting of . (a) Scanning electron microscope image of a single crystal of bulk that was microstructured with a focused ion beam. (b) Optical microscope image of many resistivity bridges lithographically patterned on an epitaxial thin-film sample of . (c) Resistivity versus temperature at K for the crystal shown in (a), before and after the sample received a 0.45 C/cm dose of 2.5 MeV electron irradiation. The irradiation causes an increase in from 0.13 -cm to 0.45 -cm and a decrease in from 1.35 K to 0.97 K. (d) Resistivity versus temperature data measured for a representative thin-film bridge, before and after a 0.45 C/cm dose of 2.5 MeV electron irradiation. Inset shows a zoomed-in view of for K: irradiation induces an increase in from 1.14 -cm to 1.56 -cm and a decrease in from 1.40 K to 0.79 K.
Epitaxial thin films of SrRuO
For our experiments on thin-film SrRuO, we synthesized a (001)-oriented film of SrRuO on a (110)-oriented NdGaO substrate (Crystec, GmbH) by molecular beam epitaxy following the substrate preparation and adsorption-controlled film growth procedures detailed in Refs. [19, 25]. After preliminary structural characterization of the SrRuO/NdGaO(110) wafer via x-ray diffraction [24], we used standard photolithography, sputter deposition, and ion milling techniques to pattern the SrRuO film into many standalone devices for electrical resistivity measurements, as described in Ref. [10]. As shown in Figure 1(b), the region of the wafer patterned in this way is about 6 mm 4 mm in lateral extent, and each device occupies a total areal footprint of 1 mm 1 mm, thus allowing many bridges to be fabricated from the same film.
Each resistivity bridge consists of two Pt (25 nm)/Ti (5 nm) contact pads at its ends for current injection and removal, and two Pt/Ti pads in the middle for attaching voltage-sensing wires, separated by a center-to-center distance of m. Mismatch between the lattices of SrRuO and the (110) surface of NdGaO imposes in-plane, uniaxial and biaxial strains on the SrRuO film; at 295 K, the two Ru-O-Ru bonding directions along and are compressed by and relative to unstrained, bulk single crystals of SrRuO ( Å) [26, 27], corresponding to in-plane and strains of and . The longest dimension of every resistivity bridge is aligned with the slightly longer of the two in-plane Ru-O-Ru bond directions. Since the NdGaO substrate is a wide-band-gap electrical insulator, the cross-sectional area through which the current is constricted to flow through is the SrRuO film thickness, nm, multiplied by the width of the resistivity bridges as determined by the lithography, m. All of the physical dimensions of the thin-film resistivity bridges are significantly larger than the characteristic mean free paths of the charge carriers in SrRuO along the corresponding crystallographic directions, so we expect Ohmic conduction to dominate the charge transport response at all temperatures.
We diced the patterned wafer into individual pieces containing one or two devices and mechanically polished each of these smaller pieces down from the backside until the NdGaO substrate was reduced to a thickness of 100 to 200 m. Each piece could then be handled and irradiated separately from the others; the substrate-thinning step allowed the irradiation to penetrate the full thickness of the film and substrate without causing excessive attenuation of the beam. Throughout the main text and Supplemental Material [24], we refer to individual thin-film resistivity bridges separated out in this way from the same original wafer as distinct as-grown samples. Comprehensive electrical characterization of these as-grown samples is described in the Supplemental Material [24].
Electrical transport measurements
We measured four-point resistances of the epitaxial thin films and FIB-sculpted single crystals using standard ac techniques at excitation frequencies less than 200 Hz, employing a multichannel lock-in amplifier (Synktek) and dual-ended current sources with active common mode rejection [28]. Unless otherwise noted, we used bias currents of A for resistance measurements.
surements.
We controlled the sample temperature and external magnetic field using a Physical Properties Measurement System (Quantum Design) equipped with a helium-3 insert and a 14 Tesla superconducting magnet. To ensure uniform thermalization of the samples upon traversing the superconducting transitions, we incremented the temperature gradually at low temperatures: typical cooling/warming rates for thermal cycles between 0.4 K and 2 K were about 1 K/hour. Detailed explanations of the definitions and analysis methods that we used to extract and from resistivity versus temperature data can be found in [24].
High-energy electron irradiation
We performed high-energy electron irradiations ( MeV) at the SIRIUS Pelletron linear accelerator in Palaiseau, France, following the procedures described in [22]. The FIB-sculpted single-crystal and epitaxial thin-film samples all had the -axis of the crystal structure aligned with the out-of-plane direction, along which the electron beam propagated. The samples were immersed in a bath of liquid hydrogen at a temperature of 22 K during irradiation in order to produce a random distribution of vacancy-interstitial Frenkel pairs [29]. Maintaining low temperatures during electron irradiation promotes the formation of immobile, pointlike defects, without appreciable defect clustering, cascades, columnar defects, or other extended spatial correlations [30]. Measurements of the upper critical field of films before and after irradiation (Appendix B) confirm that the irradiation-induced defects are pointlike.
Using NIST tables [31] to calculate the electron stopping power of in the continuous-slowing-down approximation, we estimate the stopping length of the 2.5 MeV electrons to be 2.9 mm in , meaning that the irradiation-induced defects should be homogeneously distributed throughout the entire thickness of all samples studied here. According to the standard electron-scattering formalism, we expect that 2.5 MeV electrons transfer sufficient energy to the atoms in the crystal structure to create vacancy-interstitial Frenkel pairs on all four unique sublattices in (i.e., strontium, ruthenium, equatorial oxygen, and apical oxygen), albeit with higher production rates for the cations than for either oxygen site based on likely values for the interaction cross-sections (Appendix A).
The electron irradiation facility in Palaiseau allows for in situ electrical resistance measurements to be conducted during irradiation, with the sample held at constant K. Data of this kind are shown in Appendix A for representative epitaxial thin-film and FIB-sculpted single-crystal samples of , and allow for perhaps the most straightforward comparison with theories that attempt to model cross sections for Frenkel pair production, based on tabulated cross sections for electron scattering from different nuclei and the displacement energies of the nuclei away from the unique chemical bonding sites in question.
All low-temperature ( K) electrical transport data on irradiated samples shown in this work, however, were acquired after ex situ transfers of the samples from the cryostat at the irradiation facility to different cryostats at the Max Planck Institute for Chemical Physics of Solids in Dresden, Germany, which required warming up the samples to room temperature. This initial “room-temperature annealing” step causes recombination of some of the irradiation-induced Frenkel pairs, as well as migration of some of the interstitials created during irradiation to various other sinks in the materials, such as surfaces, grain boundaries, and/or dislocations. Nevertheless, quasi-equilibrium populations of irradiation-induced vacancies and interstitials remain after the initial warmup to room temperature that are robust against further thermal cycles between 0 and 300 K. It is the cumulative effect of the latter point-like scattering centers on the charge transport and superconductivity in SrRuO that we attempt to quantify when investigating the dose dependencies of and .
Results
In Figure 1(c,Figure 1d), we show that a moderate dose of 2.5 MeV electron irradiation ( C/cm = /cm) subtly perturbs the electrical resistivity of SrRuO. At low temperatures, where other, presumably inelastic, charge-carrier scattering mechanisms become frozen out by phase-space constraints, the effects of irradiation on become proportionally more relevant and noticeable. The irradiation-induced modification of the elastic transport scattering rate, as measured by the change in residual resistivity , reduces the superconducting while largely preserving the sharpness of the superconducting transitions, indicating that the irradiation-induced defects are distributed homogeneously throughout the samples over the characteristic coherence length scales of the Cooper pairs. The initial rates of increase and of suppression with dose are similar for the FIB-sculpted single crystal of SrRuO and for the 28-nm SrRuO/NdGaO(110) epitaxial thin-film sample.
Magnetoresistance measurements of the upper critical fields for superconductivity and of Shubnikov-de Haas oscillations for electron-irradiated SrRuO thin films (Appendix B) provide additional evidence for the spatial uniformity and point-like nature of the irradiation-induced defects, and show that the primary effect of this additional disorder is to boost the rate at which the momentum of the charge-current-carrying excitations is relaxed, while preserving the itinerant charge-carrier densities (i.e., band occupancies) as well as other salient features of the normal-state electronic structure near the Fermi level, such as the quasiparticle velocities.
Having established the qualitative effects of high-energy electron irradiation on some of the basic properties of the superconducting and correlated-metal states in SrRuO, we now turn our attention to a detailed investigation of -dose-dependent changes of and . We leave direct microscopic visualization of the irradiation-induced disorder in SrRuO to future studies, because detecting dilute concentrations of point-like defects, on the order of one part per thousand, is a notoriously challenging task, well beyond the current capabilities of state-of-the-art microstructural probes such as scanning transmission electron microscopy [32, 33].
Given that substitutional disorder on the Ru site has been shown to increase at approximately the rate of substitutional disorder on the Sr site [4, 5, 6], we infer that Ru vacancies (produced at a rate of to displacements per Ru atom per C/cm of irradiation) account for most of the observed modifications to . We provide a detailed discussion of the conversions between and Frenkel defect densities, and an analysis of the likely contributions of these defects to , in Appendix A.
In Figure 2(a), we plot the relative irradiation-induced changes in measured across four distinct FIB-sculpted single-crystal samples of SrRuO as a function of the accumulated dose of MeV electron irradiation, and in Figure 2(c) we plot the resistively measured versus for all of these samples, before and after irradiation. In Figure 2(b,Figure 2d), we plot the same quantities as in Figure 2(a,Figure 2c), except now for nine distinct -nm SrRuO/NdGaO(110) thin-film samples. Raw curves for all thin-film samples that underlie the data points in Figure 2(b,Figure 2d) are presented in the Supplemental Material [24]. For both types of SrRuO samples, there are clear systematic effects of on both and , that further corroborate the trends suggested by the data and samples highlighted in Figure 1: precisely controlled increases in cause nearly monotonic increases in that are accompanied by precisely controlled, smooth decreases in .

Figure 2. Irradiation dose dependencies of and in SrRuO epitaxial thin films and single crystals. (a,b) Relative change in residual resistivity induced per unit dose of 2.5 MeV electron irradiation in (a) microstructured bulk single crystals and (b) thin films after a single round of irradiation. Each data point represents a physically distinct sample. Linear fits of all available data points are shown as solid gray lines; the shaded region bounded by dashed lines indicates uncertainties in the fitted slope of the best-fit line for each data set. (c,d) Resistively-measured superconducting versus accumulated dose for the irradiated samples displayed in panels (a,b) (teal markers), as well as the as-grown s for the same samples (dark blue markers clustered at ). Here is defined as the temperature at which crosses the 50% threshold of , and vertical error bars on each thin-film square in (d) indicate the temperatures at which crosses the 10% and 80% thresholds of ; error bars defined analogously for the single-crystal in (c) are smaller than the heights of the corresponding midpoint squares. In panels (c,d), irradiated samples having s below the base temperatures of the measurement cryostats (horizontal black dashed lines) are drawn as open squares, and were not included in the linear fits of drawn as solid gray lines [dashed gray lines have the same meaning as in panels (a,b)]. The vertical and horizontal scales of the single-crystal graphs in (a,c) are identical to the corresponding thin-film graphs in (b,d) to facilitate visual comparisons between the data sets.
For the FIB-sculpted SrRuO single-crystal sample that received the highest irradiation dose in Figure 2(a,Figure 2c) ( C/cm), such a dose was sufficient to reduce from its as-grown value of K to a post-irradiation value that was less than the base temperature of the cryostat (in this case, an adiabatic demagnetization insert for the PPMS), K. Accordingly, we schematically display this sample in Figure 2(c) as open square markers consistent with a range of possible s between and K. This sample was not included in fits to described below, but it provides an illustrative example that highlights the stringent requirements on sample purity for observing superconductivity in SrRuO: an extremely small (in absolute terms) increase of the quasiparticle scattering rate by -cm is sufficient to suppress from near its maximally-achievable value to almost zero. Likewise, for the SrRuO thin-film samples that accumulated the highest irradiation doses in Figure 2(b,Figure 2d) ( C/cm and C/cm), we did not observe superconductivity in data collected down to the base temperature of the helium-3 cryostat, K. We do not include these points in subsequent fits to , and we leave a more comprehensive exploration of the regime of very low s to future work.
To provide the simplest-possible quantitative parameterizations of these dose dependencies, we assume that and should initially vary linearly with over the investigated ranges of doses in Figure 2, because all here correspond to dilute concentrations of irradiation-induced defects, of the order of (Appendix A). Fitting the data points in Figure 2 to lines, and taking systematic error bars on the fit coefficients to be surrounding the best-fit values, we obtain the curves summarized in Table I. Alternatively, the explicit dependencies of and on can be eliminated by taking the ratios of the slopes of the best-fit lines, which yields K/-cm for bulk crystals and K/-cm for the thin films.
In other words, although the individual magnitudes of both of the rates and for the SrRuO/NdGaO(110) epitaxial thin-film samples are greater, in a statistically significant sense, than the corresponding rates for the FIB-sculpted single crystal SrRuO samples, these differences essentially compensate each other when taking the ratio of these quantities, resulting in initial rates of suppression with increasing residual resistivity, , for the two types of SrRuO samples that are indistinguishable within experimental precision.
Discussion
In Figure 3, we plot data points extracted from the literature as gray circles for bulk single crystals of SrRuO containing natively occurring (as-grown) defects [3], deliberate substitutions of Ir or O on the Ru site [5], as well as deliberate substitutions of La on the Sr site [6]. Results for electron-irradiated FIB-sculpted single crystals of SrRuO from the present work are overlaid as dark purple squares on these reference data points. For the -irradiated crystals, several repetitions of the cycle of allowed us to use the same physical samples to acquire multiple data points—hence the larger number of data points shown in Figure 3 ( in total) than in Figure 2(b,Figure 2d)—which, taken together, span nearly the entire range of disorder over which superconductivity is observed in unstrained bulk single crystals of SrRuO. The overall trend of the previously reported dependence is reproduced well, with the more precisely clustered data from this work suggesting a slightly steeper slope of the best-fit line than previously, as seen by comparing the purple and gray solid lines in Figure 3.
Figure 3. Universal impact of elastic, momentum-relaxing scattering on superconductivity in SrRuO. Gray circles represent previously published data for bulk single-crystal samples of SrRuO containing native defects [3], as well as intentionally-substituted chemical impurities on the Ru and Sr sites—viz., Sr(RuTi)O [5], Sr(RuIr)O [5], and (SrLa)RuO [6]. Dark purple squares represent results from this work for -irradiated SrRuO single crystals containing native defects plus irradiation-induced vacancy-interstitial Frenkel pairs on all atomic sublattices. For all single-crystal data shown here, the measured residual resistivity in absolute units, [-cm], is utilized as a proxy assumed to be directly proportional to the Cooper-pairbreaking energy scale, , with no adjustments of the common zero point of these scales (bottom axis). On the other hand, results from this work for -irradiated 28-nm SrRuO/NdGaO(110) epitaxial thin-film samples are plotted as red squares against the irradiation-induced change in residual resistivity, [-cm] (top axis). As explained in the main text, a range of rigid shifts of the zero point of the top axis relative to the bottom axis are, in principle, compatible with the results of this work; here we chose an offset of -cm for visual clarity, indicated by the red arrow in the upper left corner. Color-coded straight lines are linear fits to each family of SrRuO samples and the systematic uncertainties in the best-fit rates of suppression () are indicated by the color-coded shaded regions between dashed lines.
The data points for all SrRuO single-crystal samples in Figure 3 are plotted versus the absolute residual resistivity , without any shift in the zero point of this scale. Implicit in this choice is the assumption that all sources of quasiparticle scattering that contribute to , both intrinsically present from the original crystal growth and induced by irradiation, also contribute proportionally to the Cooper-pair-breaking energy scale , and thus to the suppression of . This assumption is justified by the limited scatter in the behavior observed across different studies, and by the inferred extrapolation of the experimental data to a common range of zero-disorder to K, which are furthermore consistent with the highest values of found in the literature for actual crystals under ambient conditions.
| Sample type | Range of fit in | Range of fit in | ||
| Bulk single crystal | ||||
| 28 nm films |
Table I. Linear fits to the dose dependencies of and of plotted in Figure 2(a,Figure 2b) and Figure 2(c,Figure 2d), respectively. Error bars on the fit coefficients are . Units: [-cm], [K], [C/cm].
In contrast, the relatively large scatter in the initial absolute values of across nine distinct as-grown thin films (1.1 to 1.5 -cm [24]) is incommensurate with the narrow distribution of as-grown s for these same samples (1.35 to 1.41 K). An assumption of direct proportionality between and is therefore not appropriate for the thin-film samples. Nevertheless, while there is no discernible correlation between absolute values of and , the changes in and after being subjected to different doses of 2.5 MeV electron irradiation, and , respond in a highly correlated fashion, in a way that is similar to the dose-dependent trends that we observed on FIB-sculpted single crystals. The most intuitive way to reconcile these observations is to assume that many of the natively occurring extended defects in as-grown thin films [10, 18, 19, 20, 21, 33, 34, 35, 36] contribute to but do not contribute appreciably to suppression, whereas any pointlike defects, including those created by electron irradiation, cause both and to rapidly increase in lockstep, just as in bulk crystals.
The natural interpretation of the thin-film data as reflecting a separation of independent sources of scattering that couple differently to and therefore suggests that our irradiation experiments should still accurately probe the Cooper-pair-breaking rate of suppression, , in thin films, provided that the relative change in is measured pairwise for the same thin-film sample before and after each sample accumulates a given dose of electron irradiation (Appendix § 2.3 and Figure 7). Following this line of reasoning, we combine the data for the as-grown and electron-irradiated 28-nm thin-film samples previously shown in Figure 2(b,Figure 2d) and replot these 18 data points as red squares against the top axis of Figure 3, , thereby eliminating the explicit dependence on the dose of 2.5 MeV electron irradiation. By employing the same units of [-cm] for the quasiparticle scattering rates on the top and bottom axes in Figure 3, we do not allow for any arbitrary scale factors in comparing the thin-film and single-crystal results.
Figure 7. versus absolute for SrRuO single crystals and epitaxial thin films. For SrRuO single crystals, purple square markers from the electron-irradiation experiments detailed in this work and gray circular markers extracted from impurity-scattering studies in the literature [3, 5, 6] are reproduced from Figure 3 of the main text. For SrRuO thin films, the red square markers plotted here are obtained directly from , without the shift in the horizontal axis used in Figure 3. Red lines connect pairs of points measured before and after 2.5 MeV electron irradiation for the same physical thin-film sample: solid lines for irradiated samples having measurable , and dashed lines terminating at K for irradiated samples having below the base temperature of our cryostat, 0.4 K. The two-point slopes suggested by this pairwise grouping of the data are remarkably consistent, despite differing “initial conditions” (i.e., differing absolute values); moreover, the Cooper-pair-breaking effect induced by irradiation is at least 10 times larger in magnitude than whatever subtle trend may be discernible in the behavior observed across distinct as-grown thin-film samples, which are all encircled by a red oval.
Applying this procedure reveals a striking similarity between the best-fit slopes of the single-crystal and thin-film data sets, as depicted by the color-coded solid lines in Figure 3 and summarized in Table II. This result strongly suggests that momentum-relaxing elastic scattering from pointlike disorder plays the same microscopic role in Cooper-pair breaking in thin-film and
Table II. Error bars on the fit coefficients are .
single-crystal , a conclusion which is further supported by a comparison of the versus behavior observed here with previous impurity-scattering studies of bulk single crystals [3, 4, 5, 6, 8]. Despite those studies encompassing numerous distinct realizations of “disorder”—i.e., variable densities of chemically different pointlike scattering centers, some of which couple strongly and some of which couple strongly to the charged quasiparticle excitations near the Fermi level—there is remarkable consistency across all studies when these microscopic details of the scattering processes are subsumed into their combined impact on the residual resistivity, and is treated as the relevant independent variable that ultimately controls the dependent variable, .
Quantitatively determining the appropriate rigid shift between the zero points of the top and bottom axes in Figure 3 would require more information than we have obtained in these experiments, since we do not know the clean-limit of these thin-film samples with certainty because the films are subject to biaxial () and uniaxial () in-plane strains imparted by lattice matching to the substrate. Indeed, the simplest-possible empirical estimates, described in Appendix § 2.3, place a maximum range on for of 1.4 to 2.7 K; the illustrative choice of the 0.5 -cm horizontal offset employed in Figure 3 corresponds to K, and is furthermore consistent with a comparison to expectations from Abrikosov-Gor'kov pairbreaking theory (Appendix § 2.4).
Regardless of this remaining quantitative uncertainty in determining the exact of on a true plot of , the results displayed in Figure 3 offer compelling evidence that quantitatively very similar mechanisms of superconductivity are operative in epitaxial thin films and single crystals of . We finish with a brief discussion of what insights are offered by the robust connection between and revealed by this work regarding open questions surrounding the materials science and the physics of superconducting thin films.
Although the in-plane orthorhombic strain imparted by the NdGaO(110) substrates on SrRuO likely boosts of these thin films slightly above that of unstrained bulk tetragonal SrRuO [19, 37], our data indicate that any strain-induced enhancement of the clean-limit pairing scale in our thin films (and concomitant symmetry-imposed anisotropies in the superconducting gaps caused by the reduction of the crystal point-group symmetry from tetragonal () to orthorhombic ()) are modest effects. They likely result from subtle changes in the superconducting order parameter that become irrelevant in the experimentally accessible range of quasiparticle scattering rates, because samples of either variety having as-grown s of about 1.4 K exhibit nearly identical rates of suppression with increasing when additional point-like defects are introduced via high-energy electron irradiation.
Turning this observation around, the relatively “high” s achieved for the as-grown films, which approach the clean-limit value for bulk SrRuO of 1.50 to 1.55 K, indicate minuscule concentrations of native in-plane defects that are known to be strongly pair-breaking (for example, of order 0.01% Ru vacancies). This attests to the remarkable degree of vacancy control that has been achieved in the best SrRuO thin films, which sets a benchmark for modern oxide molecular-beam epitaxy. This result is also encouraging from a physics point of view, particularly for ongoing research efforts using SrRuO thin films to gain new perspectives on the outstanding puzzles regarding the order parameter and mechanism of the enigmatic unconventional superconductivity in this material. Finally, we note that the smooth progression from “high” to “low” s in all cases is consistent with the presence of a sign-changing gap on all bands (likely with symmetry-enforced, non-accidental nodes) if intraband scattering dominates the pair breaking [3], or with the absence of large differences between the average magnitudes of on distinct bands if interband scattering dominates the pair breaking (see also Appendix § 2.4, Figure 8 & Figure 9).

Figure 8. Abrikosov-Gor’kov pair-breaking theory simulations of for bulk SrRuO. Purple square markers are data collected in this work for as-grown and electron-irradiated FIB-sculpted SrRuO single crystals, reproduced from Figure 3 of the main text. The orange curves and shaded region are simulations obtained by solving the implicit AG-like equation for as a function of , varying while keeping fixed; for quantitative values of the parameters employed here, see the text of Appendix D
Figure 9. Abrikosov–Gor’kov pair-breaking theory simulations of for thin films. Square markers are data collected in this work for as-grown and electron-irradiated 28-nm thin-film samples. Differently-colored sets of markers are the same underlying versus data set reproduced from Figure 3 of the main text, plotted with different ad hoc rigid shifts of the horizontal axis, , as indicated in the inset legend. Solid color-coded lines are simulations obtained by solving the implicit AG-like equation for as a function of the Cooper-pair-breaking energy scale , varying while keeping both eV and fixed.
In conclusion, the results displayed in Figure 1-Figure 3 show the insights that can be gained by precisely tuning the point defect concentrations in SrRuO using high-energy electron irradiation. These experiments can be performed equally well on epitaxial thin films and single crystals of SrRuO, and have the extremely high precision demonstrated here. Other methods commonly used to locate samples along a horizontal “disorder” axis, such as x-ray spectroscopy in electron microprobes [3, 38], require large sample volumes to accurately quantify dilute impurity concentrations, and would not have the sensitivity to detect such minute deviations from ideal stoichiometry. The precision available in low-volume stoichiometry techniques based on transmission electron microscopy [22, 33], Raman scattering [32], or Rutherford backscattering spectrometry, is lower still, so the high-energy electron irradiation enables investigation of a part of parameter space that cannot currently be reached by any other technique. Moreover, the ability in irradiation studies to directly compare measurements performed on the same samples, before and after irradiation, renders moot concerns about inevitable sample-specific peculiarities that could obscure the effects of the pairbreaking point disorder that we wish to study.
With a controlled and reproducible knob now available for finely adjusting the quasiparticle scattering rate and in SrRuO, it may be possible to definitively resolve some of the longstanding controversies regarding the superconductivity, such as the interpretation of heat capacity data for [39, 40, 41] in light of other thermodynamic probes that evidence a multi-component order parameter [42, 43, 44], as well as the interpretation of signals in muon spin relaxation [45, 46, 47] and polar Kerr effect data [48], which evidence some kind of time-reversal symmetry breaking onsetting near . Looking forward, it will also be interesting to see what new experiments thin films enable in this field; for example, performing irradiations using a collimated beam inside an electron microscope [49, 50] may provide a straightforward way to spatially pattern the disorder landscape so as to create superconductor/normal-metal-all- tunnel junctions for use in phase-sensitive measurements.
Data plotted in Figure 1–Figure 9 of the main text and Figure S1–Figure S5 of the Supplemental Material will be available at [51]. Additional data available upon reasonable request.

Figure S1. Scanning electron microscope images of three of the four microstructures of bulk, single-crystal that were studied in this work. Sample A is shown in Figure 1(a) of the main text.
Figure S5. Determining and from low-temperature resistivity data acquired for SrRuO single crystals. The data set shown here was also plotted in Figure 1(c) of the main text. (a) Entire dataset and fit to normal-state data for finding , the residual resistivity at . (b,c) Same dataset as in (a), now showing fits used to determine , with (b) warming and (c) cooling transitions excluded for clarity.
Acknowledgments
We thank Cyrus Dreyer for useful discussions. J.P.R., H.M.L.N., E.Z., P.H.M., D.S., and A.P.M. thank the Max Planck Society for financial support. H.M.L.N. was supported by the Alexander von Humboldt Foundation Research Fellowship for Postdoctoral Researchers. N.K. is supported by JSPS KAKENHI (No. JP18K04715, No. JP21H01033, and No. JP22K19093). Research in Dresden benefits from the environment provided by the DFG Cluster of Excellence ct.qmat (EXC 2147, project ID 390858490). Electron irradiation experiments performed at the SIRIUS beamline were supported by the EMIR&A French network (FR CNRS 3618). This work was supported by the National Science Foundation Platform for the Accelerated Realization, Analysis, and Discovery of Interface Materials (PARADIM) under Cooperative Agreement No. DMR-2039380. This research was funded in part by the Gordon and Betty Moore Foundation’s EPiQS Initiative through Grant Nos. GBMF3850 and GBMF9073 to Cornell University. Sample preparation was facilitated in part by the Cornell NanoScale Facility, a member of the National Nanotechnology Coordinated Infrastructure (NNCI), which is supported by the National Science Foundation (Grant No. NNCI2025233).
J.P.R. and H.M.L.N. contributed equally to this work.
Conversions between accumulated dose and Frenkel defect density
In situ resistivity experiments
In Figure 4(a,Figure 4b), we show the resistivity measured in situ (during 2.5 MeV electron irradiations) at K as a function of accumulated dose, for representative (a) FIB-sculpted single-crystal samples and (b) epitaxial thin-film samples of , respectively. Increasing causes monotonic, nearly-linear increases in . The best-fit lines to data collected over the range : [0, 0.60] C/cm are:
Figure 4. Dependence of on irradiation dose, cross sections for creating Frenkel defects, and the effects of these defects on the residual resistivity in . (a,b) measured in situ during 2.5 MeV irradiations, for a representative (a) FIB-sculpted single-crystal and (b) epitaxial thin-film sample of . (c-e) Calculated cross sections for creating vacancy-interstitial Frenkel pairs from (c) strontium, (d) ruthenium, and (e) oxygen atoms via electron irradiation. Each panel considers a range of possible displacement energies for creating each type of Frenkel pair (light to dark green color scale). (f) Data for FIB-sculpted single crystals showing the experimentally measured increments in residual resistivity after accumulating variable doses of MeV irradiation (black markers). These data are to be compared with the expected changes in that are attributable to irradiation-induced Sr vacancies (blue dashed lines), Ru vacancies (red dashed lines), and the sum of Ru + Sr vacancies (purple dashed lines), given the Frenkel pair production rates displayed in (c,d). The region between each pair of dashed lines represents how the systematic uncertainties in convert into ranges of possible versus behavior; the physically relevant parameter ranges are likely ranging from 7.5 to 15 eV for all sublattices in .
After both samples accumulated doses of 0.60 C/cm, we thermally cycled the samples to 300 K and back to 22 K without removing the samples from the irradiation cryostat, and remeasured the resistivities indicated by the blue markers in Figure 4(a,Figure 4b). This initial “room-temperature annealing” of the irradiated samples caused to decrease from 3.3 to 3.0 -cm (5.89 to 5.62 -cm) for the FIB-sculpted single-crystal (thin-film) SrRuO samples, respectively, or to about 91% (95%) of the starting values. Such reductions of are commonly observed upon initially warming up electron-irradiated crystals, including materials that have similar solid-state chemistries to SrRuO (e.g. SrRuO [52, 53] and high- cuprates [30, 54]), and are typically attributed to some fractions of the original concentrations of irradiation-induced defects becoming mobile at elevated temperatures. Finally, we note that due to technical difficulties with conducting in situ transport measurements during some irradiation beamtimes, the thin-film sample measured in Figure 4(b) is a resistivity bridge patterned on a 24-nm SrRuO/LSAT(100) sample, rather than the 28-nm SrRuO/NdGaO(110) samples characterized elsewhere in the manuscript. Although this likely affects certain quantitative details of the measurements, such as the rate of resistivity increase , we believe that all qualitative features of the observed behavior in Figure 4(b), including the annealing of some defects upon the first thermal cycle to 300 K, would also be observed for irradiated SrRuO/NdGaO(110) samples measured in situ.
Conversions between accumulated electron irradiation dose and Frenkel defect densities
We used the methods outlined in Ref. [22] to estimate the rates at which Frenkel defects (i.e., vacancy-interstitial pairs) are produced in SrRuO by high-energy electron irradiation. Creating a defect requires a certain amount of energy, known as the displacement energy , which is specific to a given atomic site and type of defect. The probability that such a defect will be formed during irradiation is given by the relativistic cross section for the transfer of from an electron having kinetic energy to a nucleus of that atom. In our irradiation experiments, the probability that a high-energy electron () interacts with any atom in the crystal is extremely low; we are operating in the “long mean free path” limit. Therefore, we calculate irradiation-induced defect concentrations for each atomic site independently, using the same global, measured irradiation dose in each case, without self-consistently accounting for energy losses and/or multiple scattering events that reduce below as the electron beam propagates through the samples.
In Figure 4(c-Figure 4e), we plot the calculated cross sections for creating Frenkel pairs on the strontium, ruthenium, and oxygen sites, respectively. The dependence of on the electron kinetic energy (horizontal axis) is modeled using values for the relativistic Mott scattering cross section tabulated in the literature [22, 55, 56], and the simulations are performed considering a range of displacement energies (individual traces, green color-scale) for the formation of each type of Frenkel pair. Density-functional-based calculations of Frenkel defect energetics suggest that in SrRuO, eV and eV (C. E. Dreyer, private communication). Meanwhile, previous electron irradiation studies of high- copper oxides (specifically, YBaCuO) that are structurally similar to SrRuO have found that eV [54] or 8.4 eV [49] for creating Frenkel defects on the planar oxygen site. Accordingly, we propose that displacement energies in the range of to 15 eV are the physically relevant parameter regimes for producing Frenkel defects on each of the strontium, ruthenium, and oxygen sublattices; for simplicity, we neglect any distinction between for equatorial and apical oxygen sites. Over these ranges of , the cross sections for Frenkel pair production at MeV can be read off from Figure 4(c-Figure 4e) as to 159 barn, to 193 barn, and to 37 barn.
Note that 1 barn = cm; thus, to convert these cross sections to the number of displacements per atom (dpa) for a specific dose of 2.5 MeV electron irradiation, we simply need to multiply by given in units of [number of electrons per unit area]. It follows that Frenkel defect densities () are produced on the unique sublattices per unit dose at rates of
As mentioned in the main text, direct visualization of such dilute point-like defect concentrations, = one part per thousand, is a formidable challenge, well beyond the current capabilities of state-of-the-art microstructural probes such as scanning transmission electron microscopy [22, 33]. Nonetheless, by making a few more assumptions, it is possible to indirectly confirm the general consistency of these estimated defect densities by relating to one of the observables measured in our current study—namely, the residual resistivity in the absence of superconductivity, .
In Figure 2(a) of the main text, we displayed the results of ex situ measurements of the irradiation-induced increment in residual resistivity for four distinct FIB-sculpted single crystals of SrRuO that accumulated variable doses of 2.5 MeV irradiation. As discussed above, some fraction of the defects created by irradiation are lost when a sample is first warmed to room temperature following irradiation. We model this loss as a simple multiplicative reduction of the dose received by each sample, calculating the reduction factor using the 22 K in-situ data [red points in Figure 4(a)] and the resistivity measured at 22 K after thermally cycling the sample to room temperature and back [blue point in Figure 4(a)]. In Figure 4(f), we re-plot the data points from Figure 2(a) of the main text versus the effective irradiation dose received, , where 0.62 is the calculated reduction factor. We used the reduction factor determined from sample C for all four samples because most of the other samples had incomplete in situ resistivity datasets. Although this adjustment procedure implicitly appeals to Matthiessen’s-Rule-type reasoning, which is known to not be strictly valid in perovskite-based ruthenates [14, 52], the errors associated with this approximation are small compared with other systematic uncertainties in these estimates of and in how efficiently the equilibrium for Sr, Ru, and O are transduced into corresponding increases in .
Previous studies of SrRuO single crystals that were intentionally doped with chemical impurities on the Ru site, namely, SrRu(Ti, Ir)O, demonstrated that such defects act as nearly unitary-limit scatterers that increase the residual resistivity at a rate of -cm, where is the fractional dopant density per unit cell [4, 5]. Likewise, previous studies of SrRuO single crystals that were intentionally doped with chemical impurities on the Sr site, namely, SrLaRuO, demonstrated that such defects act as nearly Born-limit scatterers that increase the residual resistivity at a much smaller rate of -cm, where is the fractional dopant density per unit cell [6]. Assuming that Sr and Ru vacancies scatter the charge-carrying quasiparticles in an identical fashion to chemically substituted impurities on the respective atomic sites, we can then multiply the above estimates of () by the empirical value of () to obtain:
These simulated rates of residual resistivity increase per unit of 2.5 MeV irradiation dose are drawn as dashed lines in Figure 4(f). The experimentally measured versus data points for FIB-sculpted single crystals of SrRuO agree reasonably well with these simulations, which suggests that the estimated cross sections for Frenkel defect production are realistic. We note that in arriving at these simulated rates of , we neglected any possible contributions of Ru and Sr interstitials to increases in , and we ignored entirely all irradiation-induced defects on the O site. Although these choices were made out of convenience—namely, because there are not any available data in the literature to indicate how much such defects might contribute to in SrRuO—Figure 4(f) indicates that considering these additional types of irradiation-induced disorder is not strictly necessary to account for the experimentally observed rate .
It remains an open question to determine whether the apparent level of agreement between simulation and experiment in Figure 4(f) is somewhat artificial, in the sense that the true displacement energies for Sr and Ru are actually towards the higher end of our estimated range (15 eV) and that the contributions of interstitials and oxygen defects to are not negligible, or whether it is suggestive that the effects on of the explicitly neglected defects are already implicitly accounted for by the “room-temperature annealing” conversion of the experimental data. Along these lines, we note that the extent to which Frenkel defects on the oxygen sites affect the resistivity and superconductivity in SrRuO might be addressable in a more direct fashion in future irradiation studies that utilize lower incident electron kinetic energies, which would transfer sufficient energy to create defects on the O sublattices, but not to create defects on the Sr or Ru sites (cf. the low MeV regions of Figure 4(c–Figure 4e)); a similar approach has been employed previously for high- cuprates [49].
Magnetoresistance of thin films
Superconducting upper critical field
In Figure 5, we show that a moderate dose ( C/cm) of 2.5 MeV electron irradiation substantially reduces the upper critical magnetic fields required to suppress superconductivity in thin-film SrRuO. For these electrical transport measurements, the external magnetic field is directed along the crystallographic -axis of SrRuO, which is also the out-of-plane direction of epitaxial thin-film samples studied here [see inset to Figure 5(c)]. Thus, the critical field probed is in the terminology of Refs. [57, 58], which is essentially equivalent to (with ) for a bulk specimen of a type-II superconductor such as SrRuO. Hereafter we refer to and interchangeably.
Figure 5. Effect of electron irradiation on the superconducting upper critical fields of SrRuO thin films. (a) Normalized resistivity versus temperature data for an as-grown SrRuO thin-film resistivity bridge, acquired at external magnetic fields ranging from 0 to 1.1 kOe in discrete steps of 0.1 kOe. (b) Same measurements as in panel (a), but after the sample had received a 0.45 C/cm dose of 2.5 MeV electron irradiation. Note that the external magnetic field steps in (b) are half of those displayed in (a), and that the maximum applied field is now merely 0.4 kOe, which is sufficient to suppress to substantially below the base temperature of the helium-3 cryostat, K. Currents of 1 A, rather than the usual 10 A, were used in the acquisition of the data shown in (a,b). (c) Upper critical magnetic fields for superconductivity (), before and after irradiation, extracted from the resistivity data in panels (a,b), along with linear fits (grey solid lines) used for estimating . The inset schematically depicts the measurement geometry and orientation of with respect to the SrRuO crystal structure. Solid markers indicate the s where crosses the 50% threshold of [bold dashed line in (a,b)], and the horizontal error bars on these s represent where crosses the 10% and 80% thresholds of [dashed lines in (a,b)].
The simultaneous decreases of both and upon high-energy electron irradiation provides further evidence that the irradiation-induced defects are point-like scattering centers that uniformly suppress the characteristic intra-granular condensation energy scale of the superconductivity in SrRuO. By contrast, if the electron irradiation created columnar and/or other types of extended defects, we would generally expect to remain unchanged, or perhaps even to be enhanced, due to increased pinning of vortices at the extended defects, while properties that depend on the details of the inter-granular Josephson network of superconducting grains [30], such as critical currents, would change.
Without a clear understanding of the effective pairing interactions that cause superconductivity to emerge in SrRuO, it is difficult to quantitatively interpret the observed magnitudes of and reduction upon irradiation, since both of these quantities depend on some appropriately weighted averages of the superconducting order parameter over the quasiparticle states near the Fermi level . Nevertheless, making the standard assumption that is orbitally limited, we can convert the measured upper critical fields into phenomenological in-plane superconducting coherence lengths () using the Ginzburg–Landau relation:
where is the superconducting flux quantum.
To account for the temperature dependence of the quantities in this expression, we follow previous studies of epitaxial thin films of SrRuO, which have shown that the nearly linear scaling of versus —observed in Figure 5 for down to 0.4 K, i.e., down to for the as-grown sample—in fact persists over at least another decade in , down to the lowest temperatures accessible in a dilution refrigerator, [59]. Accordingly, we fit all available versus data points for the as-grown and irradiated 28-nm SrRuO/NdGaO(110) samples to separate lines, and then extrapolate the fit results to zero temperature, as shown by the gray lines in Figure 5(c), to obtain Å for as-grown films, and 790 30 Å after a dose of 0.45 C/cm.
The error bars quoted here on refer to the spread of values obtained by employing 10% or 80% thresholds of to define , rather than a 50% threshold, and propagating these systematic uncertainties through the linear fits and extrapolations of the measured phase boundary [cf. gray-shaded regions in Figure 5(c)]. While the sensitivity of to percolating, but otherwise filamentary, superconducting pathways implies that electrical transport measurements inherently tend to overestimate upper critical fields compared with bulk-sensitive measurements of superconductivity [57, 60], such considerations only affect the absolute values of that are sensed by different probes; by contrast, the relative change observed here in SrRuO upon electron irradiation is unambiguous: increasing the disorder scattering rate suppresses (enhances ). This result is qualitatively consistent with previous studies of SrRuO that deduced such a correlation by analyzing data across a collection of different as-grown samples, in both bulk [8] and thin-film form [59].
Effect of electron irradiation on Shubnikov-de Haas oscillations
To place the irradiation-induced modifications to , , and in proper context, it is important to note that certain features of the normal-state near- electronic structure in SrRuO, such as the sizes and shapes of the Fermi surfaces, are not appreciably affected by the doses of high-energy electron irradiation administered in this work. To investigate the Fermiology in detail, in Figure 6 we plot low-temperature magnetotransport data acquired for an as-grown 28-nm SrRuO/NdGaO(110) sample, and for the same sample after accumulating a moderate dose ( C/cm) of 2.5 MeV electron irradiation. The external magnetic field is again directed out-of-plane, along the crystallographic axis of SrRuO, and the temperature is held constant at a nominal value of K. We used rms excitation currents of A for transport measurements performed during these field sweeps, to further enhance the signal-to-noise ratio.
Figure 6. Effect of electron irradiation on the magnetoresistance of SrRuO. (a) Normalized magnetoresistance (MR) plotted versus the externally applied magnetic field for an as-grown SrRuO thin-film resistivity bridge (blue), and the same sample after receiving a 0.60 C/cm dose of 2.5 MeV electron irradiation (teal). Here we define MR [%] , where is the extrapolated resistivity as in the absence of superconductivity. Currents of 100 A, rather than the usual 10 A, were used in the acquisition of the data shown in this figure. (b,c) Zoomed-in views of the resistivity measured over the high-field range indicated by the dashed box in panel (a), after subtracting and normalizing to empirically determined quadratic polynomial backgrounds for both data sets: . Superimposed on these slowly and smoothly varying backgrounds are rapidly oscillating Shubnikov-de Haas contributions to the magnetoresistance, that are periodic in inverse field and have dominant frequencies characteristic of charge carriers completing phase-coherent cyclotron motion around the -sheet Fermi surface of SrRuO [cf. upper left inset to panel (a)].
The overall magnitude of the normalized magnetoresistance (MR) in the normal state (i.e., for ) is somewhat reduced upon increasing [Figure 6(a)]. This behavior is generically expected whenever more than one decay rate enters the charge-carrier momentum relaxation dynamics (see, e.g., Ref. [22] and references contained therein), and indeed time-domain THz spectroscopy measurements on these SrRuO epitaxial thin films have directly shown that the low-energy optical conductivity at low temperatures contains contributions from more than one Drude-like oscillator [14]. Although detailed accounts of how electron irradiation couples to the multiple transport lifetimes that determine the dc MR (as well as the ac for small ) are beyond the scope of the present work, there is another feature of the data that is simpler to interpret quantitatively: namely, rapidly oscillating Shubnikov-de Haas (SdH) contributions to the MR.
To better isolate and visualize the SdH signal, we fit the raw MR data within the 12 to 14 Tesla field range [dashed box in Figure 6(a)] to the locally quadratic form , and then compute the background-subtracted and background-normalized MR, . We plot the results thus obtained for for the as-grown and electron-irradiated SrRuO samples in panels (b) and (c) of Figure 6, respectively. Both traces clearly exhibit SdH oscillations that are periodic in inverse field.
The dominant oscillation frequency in the data collected at the temperatures and shown here is , which results from charge carriers completing phase-coherent cyclotron orbits around the -Fermi surface of SrRuO. Counting peaks and troughs in Figure 6(b,Figure 6c), we measure complete oscillation periods over the field range to 13.964 Tesla for both the as-grown and the electron-irradiated samples, corresponding to oscillation frequencies of Tesla for both samples. In other words, moderate doses do not modify , within the experimental resolution of approximately 1%. We estimate the systematic error bars quoted here on by our ability to detect of an additional period of the -sheet-derived SdH signal over the (inverse) magnetic field range probed in Figure 6(b,Figure 6c).
In the inset to Figure 6(a), we schematically sketch the well-known Fermi surfaces of SrRuO, projected onto the two-dimensional first Brillouin zone, which is bounded by solid black lines. The area enclosed by each Fermi surface in k-space follows directly from the corresponding primary quantum oscillation frequency (), according to the Onsager relation . Therefore, the invariance of noted above implies that electron irradiation does not alter the k-space area enclosed by the -sheet Fermi surface, i.e., the gray-shaded region drawn in the inset to Figure 6(a). Invoking Luttinger’s theorem, this observation also implies that moderate doses of electron irradiation do not alter the itinerant carrier density in SrRuO.
Estimating clean-limit for thin films
In the discussion surrounding Figure 3 of the main text, we noted that in SrRuO thin films, responds sensitively to relative changes in the residual resistivity induced by electron irradiation, , yet appears to depend much less regularly on when all residual resistivities for as-grown and irradiated thin-film samples are plotted in absolute units. Figure 7 illustrates both of these observations graphically. All 18 red square markers for SrRuO thin films plotted in Figure 7 are extracted in absolute units from the corresponding data traces plotted in Figure S4 of the Supplemental Material [24]. To serve as side-by-side references, the same gray points and purple squares for SrRuO bulk single crystals from Figure 3 of the main text are also included in Figure 7.
Figure S4. Figure S4. Electrical resistivity versus temperature of thin films, before and after irradiation. Panels (a-f) are sorted from top to bottom in order of increasing dose of 2.5 MeV irradiation, ranging from (a) to (f) . The vertical and horizontal scales of all graphs are identical, to facilitate visualization of -dose-dependent changes in and . Gray arrows within each panel are guides to the eye, schematically indicating the irradiation-induced changes in and for each sample.
Viewed merely as a collection of independent points, the versus behavior of the thin-film data in Figure 7 appears rather scattered, and it is difficult to establish any clear correlation between these two variables. By contrast, when one draws lines between each pair of points that correspond to the same physical sample, measured before and after irradiation, visual inspection of the solid lines in Figure 7 reveals that electron irradiation has very similar, proportional effects on and across all distinct SrRuO thin-film samples. Furthermore, as emphasized in the main text, collecting these independent measurements of into a single number for the rate of suppression with increasing quasiparticle scattering rate results in a number that is quantitatively very similar between these electron-irradiated SrRuO thin-film samples and single-crystal SrRuO samples containing various types of defects (Table II).
In Figure 3 of the main text, we proposed that the most natural interpretation and framing of the raw data for thin films can be summarized as follows: (i.) the quantitative similarity of across all as-grown thin-film samples (1.35 to 1.41 K) implies that their values are nearly identical; (ii.) based on the detailed comparisons between SrRuO thin films and single crystals presented in the main text, using electron irradiation is a valid way to produce pure (quasi-elastic and quasi-isotropic) Cooper-pair-breaking scattering, and thereby suppress at an initial rate of ≈ 1.2 K/-cm; thus, (iii.) the remaining outstanding question is to determine the relative rigid offset of these as-grown thin-film samples relative to the zero point of an appropriately-defined Cooper-pair-breaking energy scale. Equivalently, since our electron irradiation studies have now precisely established the magnitude of suppression for a given change in that also proportionally increments , determining requires knowledge of the clean-limit of SrRuO/NdGaO(110).
At present, the simplest-possible empirical estimates constrain to be within the following extreme limits: (i.) , in which case none of the experimentally-measured in as-grown thin-film samples actually contributes to (i.e., of SrRuO/NdGaO(110) is equal to the values measured in this work on as-grown samples, ≈ 1.4 K), and (ii.) -cm, which is equal to the experimentally-measured for as-grown K thin-film samples exhibiting the lowest absolute s in the present work [24]. In this case, the hypothetical of SrRuO/NdGaO(110) would equal 1.4 K + (1.1 -cm 1.2 K/-cm) K. In Figure 3 of the main text, we arbitrarily chose a rigid shift of the top axis by -cm relative to the bottom axis, about halfway between these two extremes, corresponding to a SrRuO/NdGaO(110) K.
Based on the method proposed here of interpreting the thin-film data, it would be possible to further narrow down the range of potential values for SrRuO/NdGaO(110) in the future either by synthesizing such SrRuO thin-film samples with appreciably higher as-grown K, or by maintaining “high” K in as-grown samples with appreciably lower as-grown -cm. While some limited evidence for the former exists in the literature, e.g. Ref. [19] reported a broad (about 0.4 K-wide) resistively-measured superconducting transition centered at a midpoint K in a 55-nm-thick SrRuO/NdGaO(110) sample, to the best of our knowledge, there are no examples in the literature of SrRuO thin-film samples synthesized on any perovskite-based substrate that exhibit the latter property. The latter observation suggests that the levels of currently achieved are pushing up against a fundamental length scale for charge-carrier momentum-relaxing scattering from extended structural defects in heteroepitaxial thin film/substrate material platforms [10].
Comparison to Abrikosov-Gor’kov theory
In Figure 8, we show that the results of our electron irradiation experiments on FIB-sculpted bulk single crystals of SrRuO can be described well by theoretical curves calculated according to the standard Abrikosov-Gor’kov (AG) Cooper-pair-breaking theory of suppression [61], applied to the case of non-magnetic point-defect scattering in an effectively single-band unconventional superconductor [62, 63, 64, 65, 66].
In such theories, is found by solving an implicit equation of the general form (see, e.g., Eqn. 16 of Ref. [63] or Eqn. 35 of Ref. [66]):
where is the digamma function, is Planck’s constant, is the quasiparticle scattering rate, is Boltzmann’s constant, is the so-called “clean-limit” superconducting transition temperature in the disorder-free limit , and
is a constant that describes the anisotropy of the superconducting order parameter (SCOP), averaged over the Fermi surface (FS). For specific limiting examples of the latter, a purely sign-changing (e.g., “-wave”) SCOP that averages to zero over the Brillouin zone () has , whereas a purely uniform/isotropic ( constant, “-wave”) SCOP has .
Although more complicated relationships are often observed in real materials, in theory the single-particle scattering rate that parameterizes the Cooper-pair-breaking energy scale in Eq. (1) is typically assumed to be directly proportional to the experimentally measured planar residual resistivity , , where is a suitably averaged characteristic plasma frequency describing in-plane motion of the charge carriers [64, 66].
The quantitative parameters that describe the range of theoretical curves for bulk SrRuO drawn in Figure 8 are , K to 1.60 K, and eV to 1.34 eV. Taken together, these parameter choices determine critical scattering rates -cm to 1.00 -cm. For , only the trivial solution exists to the above AG-like equation with .
We consider the agreement observed here between theory and experiment to be essentially phenomenological in nature; there are too many outstanding unknowns about the superconductivity of SrRuO for this type of effective description to be interpreted too literally. In the regime of “small” where many of the experimental data points we have collected exist, there are redundant dependencies of the initial rate of suppression [62, 63, 65, 66], both in how increases in
transduce into increases in (i.e., on the effective ) [64, 66], and in how the fully k-resolved SCOP of a multi-orbital system such as reduces to an effectively single-band SCOP anisotropy [67, 68]. Nevertheless, while the solutions presented here are by no means uniquely determined, the overall phenomenological curves shown in Figure 8 may still prove useful as initial guesses for modeling strain-induced changes to across different thin-film variants of , as discussed in more detail below.
In principle, the systematic deviations from purely linear behavior expected in AG pair-breaking theory should allow one to infer the proximity of a given set of data points to the critical at which superconductivity is suppressed completely. This method of discerning in turn allows for an absolute determination of , because increases in cause linear increases in in simple Cooper-pair-breaking theories that are effectively single-band, with one dominant superconducting pairing scale, .
To apply this type of reasoning to our thin-film electron-irradiation data, we started with a phenomenological AG-like curve similar to those that describe well the data points we collected for electron-irradiated FIB-sculpted single crystals in this work, Figure 8. To optimize the general agreement of these calculated curves with the experimental thin-film data points, we adjusted the effective plasma frequency eV used for the bulk AG simulations in Figure 8 to eV for the thin-film AG simulations in Figure 9. The effective SCOP anisotropy was held at the same fixed value as in the case of bulk samples, although as we stressed earlier, within AG-like pair-breaking theories there are redundant dependencies of the calculated behavior on , so other solutions would be possible that would fit the existing data equally well. Finally, we assumed that any epitaxial-strain-induced increase (or decrease) in for relative to the of bulk results from a -independent multi-plicative enhancement (or suppression) of , along with negligible strain-dependent changes in the Fermi surface (cf. Figure 6). Under this assumption, we uniformly scaled up (or down) the phenomenological AG-like curve based on various hypothetical values, keeping eV constant, to produce simulated curves against which to compare the thin-film data points.
Figure 9 illustrates the results of this type of analysis. For hypothetical curves towards the high end of our allowed range of , e.g., for the blue simulated curve with K ( -cm), the two irradiated thin-film samples with K and K (which both received a 2.5 MeV e dose of C/cm) would be small enough fractions of (i.e., ) that they would be well within the range of a given AG-like pair-breaking-theory curve where non-linearities in the behavior would become obvious. The absence of any such non-linear features in the experimental data suggests that for is likely less than 2.2 K, provided that our assumption of is, in fact, realized.
On the other hand, the red simulated curve having K (which combined with results in -cm) matches the red experimental data squares (plotted with an ad hoc -cm) quite well. The purple simulated curve having K ( -cm) matches the purple experimental data squares (plotted with a much reduced ad hoc -cm) almost equally well. Down to the low end of the range of that we believe are plausible ( K), the three most heavily irradiated thin-film samples, with K, never quite provide strong enough upper bounds on to exclude any hypothetical curves. Apparently all such reductions in can be compensated for, at least in principle, by postulating smaller values of . Overall, a more precise and conclusive realization of the basic ideas presented in this section would require collecting more experimental data points in the region near , where is supposed to rapidly evolve with changes in .
Supplemental Material for “Controllable suppression of the unconventional superconductivity in via high-energy electron irradiation”
Jacob P. Ruf; Hilary M. L. Noad; Romain Grasset; Ludi Miao; Elina Zhakina; Philippa H. McGuinness; Hari P. Nair; Nathaniel J. Schreiber; Naoki Kikugawa; Dmitry Sokolov; Marcin Konczykowski; Darrell G. Schlom; Kyle M. Shen; Andrew P. Mackenzie
FIB-SCULPTED SINGLE CRYSTALS
| Sample | Length () | Width () | Thickness () | 1st dose (C/) | 2nd dose (C/) |
| A | 109.8 | 9.7 | 6.5 | 0.45 | 0.60 |
| B | 80.0 | 9.7 | 5.5 | 1.177 | – |
| C | 101.8 | 8.7 | 6.5 | 0.60 | 0.30 |
| D | 71.2 | 8.5 | 4.1 | 0.30 | 0.45 |
Table S1. Dimensions of the FIB-sculpted single-crystal Hall bars, determined from scanning electron microscope images, and irradiation doses.
GROWTH AND STRUCTURAL CHARACTERIZATION OF EPITAXIAL THIN FILMS
The /NdGaO(110) film was grown by codepositing an elemental strontium flux of atoms and an elemental ruthenium flux of atoms in a background partial pressure of Torr of distilled ozone ( O + O).
To determine the thickness of the /NdGaO(110) film, we used a lab-based x-ray diffractometer (Rigaku) and Cu-K radiation to measure x-ray diffraction (XRD) data at room temperature along the specular crystal truncation rod of the thin film/NdGaO substrate, as shown in Figure S2. Coherent scattering from film crystallites having finite sizes along the out-of-plane direction produces characteristic interference fringes—i.e., secondary maxima and minima—in the XRD intensity surrounding each primary film Bragg peak. The spacing between these fringes scales inversely with the total number of layers in the crystallites, ; meanwhile, the average interplanar spacing between the layers in each crystallite, , determines where each primary film Bragg peak is centered along the axis (here and elsewhere, we use to denote the out-of-plane lattice constant of the conventional unit cell of Sr2RuO4). Together these two quantities determine the total film thickness according to , which is later needed to convert measured film resistances to resistivities.
Figure S2. Sr2RuO4 film thickness determination. (a) XRD data along the specular crystal truncation rod of the Sr2RuO4(001) thin film/NdGaO3(110) substrate, acquired with Cu-Kα radiation at room temperature. Substrate Bragg peaks are marked by asterisks, and the family of 00L peaks (L even) arising from scattering off the Sr2RuO4 film are labeled accordingly. (b) Zoomed-in view of the experimental data (blue) near the Sr2RuO4 006 and NdGaO3 220 Bragg peaks, with the best-fit simulation overlaid (dashed red).
We utilized the GenX software package [69] to directly fit the experimental XRD data over a region of momentum-transfer space surrounding the Sr2RuO4 006 and NdGaO3 220 Bragg peaks, (, – NdGaO3 r.l.u. – 1/Å). From these simulations, displayed as a dashed red line in Figure S2, we determined that the Sr2RuO4 film crystallites consisted of conventional unit cells—i.e., RuO2 planes/layers—stacked along the out-of-plane direction, having an average -axis lattice constant of Å, and thus a total film thickness of Å Å. The confidence intervals listed here for refer to the estimated scale of film roughnesses, which were incorporated as a parameter in the XRD simulations; thickness variations are likely the largest systematic uncertainty in determining the electrical resistivity of the film in absolute units.
We note that Å for a thin film of Sr2RuO4 corresponds to expansion of the room-temperature bulk single-crystal value ( Å [26]) by . Using the relevant Poisson ratio for Sr2RuO4 at 300 K, [70], this measured -axis expansion is consistent with an elastically strained thin film subject to an average in-plane compressive strain of , which agrees well with the nominal in-plane lattice mismatch of Sr2RuO4 and NdGaO3(110) at room temperature, [26, 27], as we discuss in more detail below.
The (110) surface of the NdGaO3 crystal structure (described using the non-standard setting of space group #62) is spanned by the mutually orthogonal [1] and [110] lattice translation vectors, which impart modest amounts of biaxial () in-plane compressive strain and uniaxial () in-plane strain on the Sr2RuO4 thin film. Specifically, at 295 K, we expect the two in-plane Ru-O-Ru bonding directions along Sr2RuO4[100] and Sr2RuO4[10] to be compressed by NdGaO3[1] and NdGaO3[] relative to unstrained bulk single crystals of Sr2RuO4 ( Å) [26, 27]. These deformations correspond to in-plane and strains of and . At 4 K, the in-plane and components of substrate-imposed misfit strains on Sr2RuO4 become and , respectively, accounting for differences in the thermal expansion of NdGaO3 and Sr2RuO4 [71, 72].
Electrical Characterization of Epitaxial Thin Films
As-grown films
In Figure S3, we summarize the electrical resistivity versus temperature behavior of 15 distinct resistivity bridges patterned on the same SrRuO/NdGaO(110) mother wafer, measured between 0.4 K and 300 K. Hereafter we denote the resistivity by , which is understood to be shorthand for the component of a resistivity tensor that is defined with respect to a crystallographic coordinate system having SrRuO[100] NdGaO[1], SrRuO[10] NdGaO[110], and SrRuO[1] NdGaO(110). We acquired all data displayed in Figure S3 with the SrRuO thin-film bridges in their as-grown states—i.e., before subjecting the samples to any doses of high-energy electron irradiation. The graded color scale of the data traces visually represents and corresponds to the ordering of the samples when they are sorted by their residual resistivity ratios (RRRs), defined here as , which range from about 70 (yellow-orange) to 110 (dark purple).
Figure S3. Electrical transport properties across 15 distinct as-grown SrRuO thin-film resistivity bridges. Data for a given sample are plotted using the same color in all panels; this color visually encodes the measured for that sample, as indicated by the inset legend in (a). (a) Absolute resistivities versus temperature. (b) Log-log plot of the scaled resistivities versus temperature. Data are omitted for temperatures between about where helium-3 circulation starts/stops in the cryostat used for these measurements; the according rapid changes in temperature prohibit proper thermalization of the samples in this range of . (c) Scaled resistivities versus temperature squared. All traces deviate in a subtle, but apparently systematic, fashion at low temperatures from the oft-prescribed power-series expansion, which precludes extrapolating the resistivity below based on fits to such a power law above . (d) Scaled resistivities versus temperature, zoomed in for to visualize the superconducting transitions across distinct samples. Inset displays the resistively-measured s [criterion: where crosses 50% of ], plotted against , which is employed as an approximation of the residual resistivity. Error bars on are the temperatures where crosses the 10% and 80% thresholds of . Neither nor the superconducting transition widths exhibit any discernible correlation with .
Following previous work [3], we assume that the variation in absolute observed across distinct as-grown resistivity bridges in Figure S3, with values ranging from 120 to 131 -cm, is primarily caused by systematic uncertainties in the sample preparation and in the measurements, rather than by any intrinsic differences in the scattering mechanisms relevant to SrRuO that relax the charge-carrier momentum in this highly incoherent regime of electrical transport. Likely sources of such errors include variations in the effective cross-sectional thickness that the electrical current is constricted to flow through, as well as small misalignments and discrepancies in the lithographically patterned dimensions of the electrical contacts to different bridges. More microscopically, such variations in the effective cross-sectional thicknesses of the resistance bridges may result from total film thickness variations, out-of-phase boundaries in the film that preferentially nucleate in the vicinity of substrate step edges [10, 33, 34, 73], and/or locally higher- Ruddlesden-Popper inclusions/stacking faults that form to accommodate non-stoichiometry in Ru-rich (nominally SrRuO) epitaxial thin films [18, 19, 21, 33].
Therefore, for all other curves plotted in Figure S3-Figure S3), and throughout the main text and the remainder of the Supplemental Material, we plot so-called ‘scaled ’ curves: we first normalize the measured resistance versus temperature data for a given bridge by its resistance measured at 300 K, and then we uniformly multiply this normalized data by the arithmetic mean of measured across all as-grown bridges, -cm. We note that the value of -cm (standard deviation about the mean -cm) obtained for the SrRuO thin-film resistivity bridges studied in this work agrees well with the established literature value appropriate to bulk SrRuO single crystals, -cm (standard deviation about the mean -cm) [3]. This quantitative level of agreement indicates that these SrRuO films do not contain significant volume fractions of higher- Ruddlesden-Popper intergrowths (SrO)(SrRuO), which for all compounds have characteristic values of -cm [14, 74]. Additionally, we note that the above procedure of computing appropriately scaled values is important only if one wishes to interpret in absolute units—for example, as a proxy for computing the transport lifetime (or transport scattering length) of charged quasiparticle excitations in SrRuO. By contrast, many of the results contained in this manuscript for irradiation-induced changes in quantities like are extracted from relative changes in the measured resistance as a function of temperature, which are independent of any remaining systematic errors in the deduced temperature-independent scale factors that enter into conversions of to .
From the data in Figure S3, we determined (arithmetic mean one standard deviation) values of -cm, -cm, and . Here is reported as the temperature at which the resistance of each SrRuO thin-film bridge measured at a current of —or equivalently, at a nominal current density of —crosses 50% of its value measured at 2 K. Section IV of the Supplemental Material contains further discussion of the definitions and analysis methods used to extract the values of shown in this work. The uniformity of the resistively measured s across the wafer (0.024/1.359 = 1.8%) justifies our use of different bridges as nearly-identical ‘initial conditions’ for gauging the effects of different electron irradiation doses on . On the other hand, the comparatively larger spread in across the wafer (0.17/1.33 = 13%), the lack of any discernible correlation between and for this set of as-grown samples [see inset to Figure S3], and the large absolute values of at which superconductivity remains robustly observed (“large” relative to the phenomenological dependence of versus established for bulk SrRuO single crystals, such as in Figure 3 of the main text), all suggest that scattering from most of the native defects existing in these SrRuO epitaxial thin films does not appreciably contribute to suppression.
Films before and after irradiation
In Figure S4, we show all of the low-temperature, zero-field electrical transport data for SrRuO thin films that underlie the data points included in Figure 2(b,Figure 2d) and Figure 3 of the main text. Each panel of Figure S4 focuses on one—or sometimes two, to check reproducibility—distinct SrRuO thin-film resistivity bridge(s), and plots the data trace(s) measured before and after incrementing the dose of 2.5 MeV electron irradiation received by the sample(s). We used excitation currents of () to measure all data displayed in Figure S4. No discernible shifts of the measured s occur for in either the as-grown or irradiated samples; on the other hand, increasing the excitation current to causes all measured s to decrease by about 0.02 K [cf. Fig. S3(d)].
All data displayed in Figure S4 are scaled resistivities, computed according to the procedure described in Figure S3 and Sec. III A above. By normalizing all raw values (experimentally measured before and after irradiation) to a common value of -cm, we intend to disentangle and to isolate the irradiation-induced modifications in from potential irradiation-induced modifications to the effective cross-sectional area of the conducting channel specific to each thin-film resistivity bridge. We believe that the former effects are the physically relevant changes needed to understand and to interpret the response of superconductivity in SrRuO to high-energy electron irradiation, which is the focus of this manuscript.
We do (somewhat irregularly) observe hints suggestive of the latter type of effects, however, such as nearly-temperature-independent scaling of the pre-irradiation curve for a given bridge by a factor of 1.02 − 1.03 after irradiation. Although other explanations of this are possible and should be investigated in more detail in future work, we note that one of the simplest scenarios—as suggested, e.g., for SrRuO thin films in Ref. [53]—would involve migration of irradiation-induced interstitial defects to the surfaces of the thin-film samples in a way that reduces the conducting film thickness by about one RuO monolayer. This would result in a fractional increase in measured resistances of about for a thin film that originally consisted of about 44 monolayers, in addition to the irradiation-induced changes in the temperature-dependent resistivity of the remaining intact 43 layers.
Independent of the above considerations about how seriously to interpret the plotted values of in absolute units for the epitaxial thin-film samples, the low and intermediate irradiation doses () displayed in Figure S4 clearly cause modest reductions of by several tenths of a Kelvin, and the highest irradiation doses () are sufficient to reduce from (as grown) to values after irradiation that are well below the base temperature of the helium-3 cryostat used for these measurements, 0.4 K.
Determining and from resistivity
FIB-sculpted single crystals
For the FIB-sculpted single crystals of SrRuO, we determined the extrapolated zero-temperature normal-state (ns) residual resistivities, , and superconducting s from resistivity data collected upon cooling and warming between 2 K and 0.4 K. We found by fitting the cooling and warming data together to the quadratic functional form (Figure S5, black), following previous work in the literature [3]. Across four distinct as-grown samples, we obtained fitted coefficients ranging from 6.9 to 8.2 n-cm/K, in agreement with the values to 12 n-cm/K reported in the literature [75, 76, 77, 78, 79].
To extract , we initially treated the superconducting transitions measured upon warming and cooling separately (red and blue data points in Figure S5, respectively), in order to check for hysteresis. We fitted data in the transition region to a linear function , and data in the superconducting state (Figure S5, gray) to a constant function . From the fitted coefficients , we then calculated the temperatures where intersects and where intersects , to find the onset and completion transition temperatures, respectively. Since the experimental behavior is described well by straight lines for all samples we have measured, we took to be the midpoint between the onset and completion transition temperatures. Because the thermal hysteresis between cooling and warming data is quite small—at most 0.017 K and typically < 0.010 K—throughout the manuscript we report as the arithmetic mean of the values obtained upon cooling and warming.
Epitaxial thin films
In epitaxial thin films, the measured normal-state resistivities (Figure S3) do not follow the oft-prescribed power-series expansion over a region that is wide enough in temperature to justify fitting the data to this functional form, which precludes using the results of such fits to extrapolate below . Accordingly, we adopted slightly more empirical approaches to extract values of and to use in plots and analyses throughout the main text and Supplemental Material: we set (i.e., simply equal to the measured resistivity at 2 K); we define to be the temperature at which crosses the 50% threshold of ; finally, we quote “error bars” on as the temperatures at which crosses the 80% and 10% thresholds of , respectively. Any effects of thermal hysteresis on the resistively measured superconducting transitions for thin films are negligibly small compared to other, likely extrinsic, factors that broaden these transitions, such as static material inhomogeneities; therefore, all plotted and analyzed data for were acquired upon gradually warming up the thin-film samples, starting from near the base temperature of the helium-3 cryostat.
We note in passing that if one insists on fitting the thin-film transport data to , e.g., over the interval , the resulting fit coefficients range from to and to across the 15 as-grown samples shown in Figure S3. Extending the fitting range to higher temperatures yields systematically smaller fitted coefficients and larger fitted values of . Additionally, we note that the relative contribution of temperature-dependent scattering mechanisms to in can be estimated using published results of fits to , which find that to [75, 76, 77, 78, 79], and hence to . On the other hand, the as-grown and irradiated thin films studied in this work have measured to (Figure S3 and Figure S4). Thus, our ad hoc assignment of will generally overestimate the hypothetical “true” values of by 1 to 5%.
References
- [1]A. P. Mackenzie and Y. Maeno, The superconductivity of Sr₂RuO₄ and the physics of spin-triplet pairing, Rev. Mod. Phys. 75, 657 (2003).DOI
- [2]A. P. Mackenzie, T. Scaffidi, C. W. Hicks, and Y. Maeno, Even older after twenty-three years: the superconducting order parameter puzzle of Sr₂RuO₄, npj Quantum Mater. 2, 1 (2017).arxiv.org/abs/1706.01942
- [3]A. P. Mackenzie, R. K. W. Haselwimmer, A. W. Tyler, G. G. Lonžarich, Y. Mori, S. Nishizaki, and Y. Maeno, Extremely Strong Dependence of Superconductivity on Disorder in Sr₂RuO₄, Phys. Rev. Lett. 80, 161 (1998).DOI
- [4]N. Kikugawa, S. Saita, M. Minakata, and Y. Maeno, Effect of Ti substitution on the residual resistivity in the spin-triplet superconductor Sr₂RuO₄, Physica B 312-313, 803 (2002).DOI
- [5]N. Kikugawa, A. Peter Mackenzie, and Y. Maeno, Effects of In-Plane Impurity Substitution in Sr₂RuO₄, J. Phys. Soc. Jpn. 72, 237 (2003).DOI
- [6]N. Kikugawa, A. P. Mackenzie, C. Bergemann, R. A. Borzi, S. A. Grigera, and Y. Maeno, Rigid-band shift of the Fermi level in the strongly correlated metal: Sr₂−yLayRuO₄, Phys. Rev. B 70, 060508 (2004).DOI
- [7]H. Suderow, J. P. Brison, J. Flouquet, A. W. Tyler, and Y. Maeno, Very low temperature thermal conductivity in the layered perovskite superconductor Sr₂RuO₄, J. Phys.: Condens. Matter 10, L597 (1998).
- [8]Z. Q. Mao, Y. Mori, and Y. Maeno, Suppression of superconductivity in Sr₂RuO₄ caused by defects, Phys. Rev. B 60, 610 (1999).
- [9]M. Uchida, I. Sakuraba, M. Kawamura, M. Ide, K. S. Takahashi, Y. Tokura, and M. Kawasaki, Characterization of Sr₂RuO₄ Josephson junctions made of epitaxial films, Phys. Rev. B 101, 035107 (2020).
- [10]Y. Fang, H. P. Nair, L. Miao, B. Goodge, N. J. Schreiber, J. P. Ruf, L. F. Kourkoutis, K. M. Shen, D. G. Schlom, and B. J. Ramshaw, Quantum oscillations and quasiparticle properties of thin film Sr₂RuO₄, Phys. Rev. B 104, 045152 (2021).DOI
- [11]H. Noad, E. M. Spanton, K. C. Nowack, H. Inoue, M. Kim, T. A. Merz, C. Bell, Y. Hikita, R. Xu, W. Liu, A. Vailionis, H. Y. Hwang, and K. A. Moler, Variation in superconducting transition temperature due to tetragonal domains in two-dimensionally doped SrTiO₃, Phys. Rev. B 94, 174516 (2016).DOI
- [12]C. A. Watson, A. S. Gibbs, A. P. Mackenzie, C. W. Hicks, and K. A. Moler, Micron-scale measurements of low anisotropic strain response of local T_c in Sr₂RuO₄, Phys. Rev. B 98, 094521 (2018).DOI
- [13]B. B. Burganov, C. Adamo, A. Mulder, M. Uchida, P. D. C. King, J. W. Harter, D. E. Shai, A. S. Gibbs, A. P. Mackenzie, R. Uecker, M. Bruetzam, M. R. Beasley, C. J. Fennie, D. G. Schlom, and K. M. Shen, Strain Control of Fermiology and Many-Body Interactions in Two-Dimensional Ruthenates, Phys. Rev. Lett. 116, 197003 (2016).
- [14]Y. Wang, H. P. Nair, N. J. Schreiber, J. P. Ruf, B. Cheng, D. G. Schlom, K. M. Shen, and N. P. Armitage, Separated transport relaxation scales and interband scattering in thin films of SrRuO₃, CaRuO₃, and Sr₂RuO₄, Phys. Rev. B 103, 205109 (2021).DOI
- [15]C. C. Tsuei and J. R. Kirtley, Pair-ing symmetry in cuprate superconductors, Rev. Mod. Phys. 72, 969 (2000).
- [16]Y. Maeno, H. Hashimoto, K. Yoshida, S. Nishizaki, T. Fujita, J. G. Bednorz, and F. Lichtenberg, Superconductivity in a layered perovskite without copper, Nature 372, 532 (1994).DOI
- [17]Y. Krockenberger, M. Uchida, K. S. Takahashi, M. Nakamura, M. Kawasaki, and Y. Tokura, Growth of superconducting Sr₂RuO₄ thin films, Appl. Phys. Lett. 97, 082502 (2010).DOI
- [18]M. Uchida, M. Ide, H. Watanabe, K. S. Takahashi, Y. Tokura, and M. Kawasaki, Molecular beam epitaxy growth of superconducting Sr₂RuO₄ films, APL Mater. 5, 106108 (2017).
- [19]H. P. Nair, J. P. Ruf, N. J. Schreiber, L. Miao, M. L. Grandon, D. J. Baek, B. H. Goodge, J. P. C. Ruff, L. F. Kourkoutis, K. M. Shen, and D. G. Schlom, Demystifying the growth of superconducting Sr₂RuO₄ thin films, APL Mater. 6, 101108 (2018).DOI
- [20]C. M. P. Garcia, A. Di Bernardo, G. Kimbell, M. E. Vickers, F. C.-P. Massabuau, S. Komori, G. Divitini, Y. Yasui, H. G. Lee, J. Kim, B. Kim, M. G. Blamire, A. Vecchione, R. Fittipald, Y. Maeno, T. W. Noh, and J. W. A. Robinson, Pair suppression caused by mosaic-twist defects in superconducting Sr₂RuO₄ thin-films prepared using pulsed laser deposition, Commun. Mater. 1, 23 (2020).
- [21]J. Kim, J. Mun, C. M. Palomares García, B. Kim, R. S. Perry, Y. Jo, H. Im, H. G. Lee, E. K. Ko, S. H. Chang, S. B. Chung, M. Kim, J. W. A. Robinson, S. Yonezawa, Y. Maeno, L. Wang, and T. W. Noh, Superconducting Sr₂RuO₄ Thin Films without Out-of-Phase Boundaries by Higher-Order Ruddlesden–Popper Intergrowth, Nano Lett. 21, 4185 (2021).DOI
- [22]V. Sunko, P. H. McGuinness, C. S. Chang, E. Zhakina, S. Khim, C. E. Dreyer, M. Konczykowski, H. Borrmann, P. J. W. Moll, M. König, D. A. Muller, and A. P. Mackenzie, Controlled Introduction of Defects to Delafossite Metals by Electron Irradiation, Phys. Rev. X 10, 021018 (2020).DOI
- [23]J. S. Bobowski, N. Kikugawa, T. Miyoshi, H. Suwa, H.-s. Xu, S. Yonezawa, D. A. Sokolov, A. P. Mackenzie, and Y. Maeno, Improved Single-Crystal Growth of Sr₂RuO₄, J. Condens. Matter 4, 6 (2019).DOI
- [24]See Supplemental Material at [URL will be inserted by publisher] for further details on samples, electrical characterization of films, and a discussion of the determination of ρ₀ and T_c. Includes references [69–79].
- [25]H. P. Nair, Y. Liu, J. P. Ruf, N. J. Schreiber, S.-L. Shang, D. J. Baek, B. H. Goodge, L. F. Kourkoutis, Z.-K. Liu, K. M. Shen, and D. G. Schlom, Synthesis science of SrRuO₃ and CaRuO₃ epitaxial films with high residual resistivity ratios, APL Mater. 6, 046101 (2018).
- [26]L. Walz and F. Lichtenberg, Refinement of the structure of Sr₂RuO₄ with 100 and 295 K X-ray data, Acta Crystallogr. Sect. C 49, 1268 (1993).DOI
- [27]M. Schmidbauer, A. Kwasniewski, and J. Schwarzkopf, High-precision absolute lattice parameter determination of SrTiO₃, DyScO₃ and NdGaO₃ single crystals, Acta Crystallogr. Sect. B 68, 8 (2012).DOI
- [28]M. E. Barber, Uniaxial stress technique and investiga tions into correlated electron systems, Ph.D. thesis, University of St. Andrews (2017).
- [29]G. W. Iseler, H. I. Dawson, A. S. Mehner, and J. W. Kauffman, Production Rates of Electrical Resistivity in Copper and Aluminum Induced by Electron Irradiation, Phys. Rev. 146, 468 (1966).DOI
- [30]M. Konczykowski and J. Gilchrist, Electron irradiation of YBa₂Cu₃O₇ ceramics, J. Phys. (Paris) III 1, 1765 (1991).DOI
- [31]M. J. Berger, J. S. Coursey, M. A. Zucker, and J. Chang, NIST Standard Reference Database 124 (National Institute of Standards and Technology, 2017).
- [32]G. Kim, Y. E. Soukolc, J. Herrero-Martin, D. Putzky, H. P. Nair, J. P. Ruf, N. J. Schreiber, C. Dietl, G. Christiani, G. Logvenov, M. Minola, P. A. van Aken, K. M. Shen, D. G. Schlom, and B. Keimer, Electronic and vibrational signatures of ruthenium vacancies in Sr₂RuO₄ thin films, Phys. Rev. Materials 3, 094802 (2019).DOI
- [33]B. H. Goodge, H. P. Nair, D. J. Baek, N. J. Schreiber, L. Miao, J. P. Ruf, E. N. Waite, P. M. Carubia, K. M. Shen, D. G. Schlom, and L. F. Kourkoutis, Disentangling types of lattice disorder impacting superconductivity in Sr₂RuO₄ by quantitative local probes, APL Mater. 10, 041114 (2022).DOI
- [34]M. A. Zurbuchen, Y. Jia, S. Knapp, A. H. Carim, D. G. Schlom, L.-N. Zou, and Y. Liu, Suppression of superconductivity by crystallographic defects in epitaxial Sr₂RuO₄ films, Appl. Phys. Lett. 78, 2351 (2001).DOI
- [35]M. A. Zurbuchen, Y. Jia, S. Knapp, A. H. Carim, D. G. Schlom, and X. Q. Pan, Defect generation by preferred nucleation in epitaxial Sr₂RuO₄/LaAlO₃, Appl. Phys. Lett. 83, 3891 (2003).DOI
- [36]M. A. Zurbuchen, W. Tian, X. Q. Pan, D. Fong, S. K. Streiffer, M. E. Hawley, J. Lettieri, Y. Jia, G. Asayama, S. J. Fulk, D. J. Comstock, S. Knapp, A. H. Carim, and D. G. Schlom, Morphology, structure, and nucleation of out-of-phase boundaries (OPBs) in epitaxial films of layered oxides, J. Mater. Res. 22, 1439 (2007).DOI
- [37]A. Steppke, L. Zhao, M. E. Barber, T. Scaffidi, F. Jerzembeck, H. Rosner, A. S. Gibbs, Y. Maeno, S. H. Simon, A. P. Mackenzie, and C. W. Hicks, Strong peak in T_c of Sr₂RuO₄ under uniaxial pressure, Science 355, eaaf9398 (2017).
- [38]A. P. Mackenzie, Recent progress in electron probe microanalysis, Rep. Prog. Phys. 56, 557 (1993).DOI
- [39]S. Nishizaki, Y. Maeno, and Z. Mao, Effect of Impurities on the Specific Heat of the Spin-Triplet Superconductor Sr₂RuO₄, J. Low Temp. Phys. 117, 1581 (1999).DOI
- [40]Y.-S. Li, N. Kikugawa, D. A. Sokolov, F. Jerzembeck, A. S. Gibbs, Y. Maeno, C. W. Hicks, J. Schmalian, M. Nicklas, and A. P. Mackenzie, High-sensitivity heat-capacity measurements on Sr₂RuO₄ under uniaxial pressure, Proc. Natl. Acad. Sci. USA 118, e2020492118 (2021).DOI
- [41]Y.-S. Li, M. Garst, J. Schmalian, S. Ghosh, N. Kikugawa, D. A. Sokolov, C. W. Hicks, F. Jerzembeck, M. S. Ikeda, Z. Hu, B. J. Ramshaw, A. W. Rost, M. Nicklas, and A. P. Mackenzie, Elastocaloric determination of the phase diagram of Sr₂RuO₄, Nature 607, 276 (2022).DOI
- [42]S. Ghosh, A. Shekhter, F. Jerzembeck, N. Kikugawa, D. A. Sokolov, M. Brando, A. P. Mackenzie, C. W. Hicks, and B. J. Ramshaw, Thermodynamic evidence for a two-component superconducting order parameter in Sr₂RuO₄, Nat. Phys. 17, 199 (2021).DOI
- [43]S. Benhabib, C. Lupien, I. Paul, L. Berges, M. Dion, M. Nardone, A. Zitouni, Z. Q. Mao, Y. Maeno, A. Georges, L. Taillefer, and C. Proust, Ultrasound evidence for a two-component superconducting order parameter in Sr₂RuO₄, Nat. Phys. 17, 194 (2021).DOI
- [44]S. A. Kivelson, A. C. Yuan, B. Ramshaw, and R. Thomale, A proposal for reconciling diverse experiments on the superconducting state in Sr₂RuO₄, npj Quantum Mater. 5, 1 (2020).DOI
- [45]G. M. Luke, Y. Fudamoto, K. M. Kojima, M. I. Larkin, J. Merrin, B. Nachumi, Y. J. Uemura, Y. Maeno, Z. Q. Mao, Y. Mori, H. Nakamura, and M. Sigrist, Time-reversal symmetry-breaking superconductivity in Sr₂RuO₄, Nature 394, 558 (1998).DOI
- [46]V. Grinenko, S. Ghosh, R. Sarkar, J.-C. Orain, A. Nikitin, M. Elender, D. Das, Z. Guguchia, F. Brückner, M. E. Barber, J. Park, N. Kikugawa, D. A. Sokolov, J. S. Bobkowski, T. Miyoshi, Y. Maeno, A. P. Mackenzie, H. Luetkens, C. W. Hicks, and H.-H. Klauss, Split superconducting and time-reversal symmetry-breaking transitions in Sr₂RuO₄ under stress, Nat. Phys. 17, 748 (2021).DOI
- [47]V. Grinenko, D. Das, R. Gupta, B. Zinkl, N. Kikugawa, Y. Maeno, C. W. Hicks, H.-H. Klauss, M. Sigrist, and R. Khasanov, Unsplit superconducting and time reversal symmetry breaking transitions in Sr₂RuO₄ under hydrostatic pressure and disorder, Nat. Commun. 12, 3920 (2021).
- [48]J. Xia, Y. Maeno, P. T. Beyersdorf, M. M. Fejer, and A. Kapitulnik, High Resolution Polar Kerr Effect Measurements of Sr₂RuO₄: Evidence for Broken Time-Reversal Symmetry in the Superconducting State, Phys. Rev. Lett. 97, 167002 (2006).pubmed.ncbi.nlm.nih.gov/17155427
- [49]S. K. Tolpygo, J.-Y. Lin, M. Gurvitch, S. Y. Hou, and J. M. Phillips, Effect of oxygen defects on transport properties and T_c of YBa₂Cu₃O₆+x: Displacement energy for plane and chain oxygen and implications for irradiation-induced resistivity and T_c suppression, Phys. Rev. B 53, 12462 (1996).
- [50]S. K. Tolpygo, J.-Y. Lin, M. Gurvitch, S. Y. Hou, and J. M. Phillips, Universal T_c suppression by in-plane defects in high-temperature superconductors: Implications for pairing symmetry, Phys. Rev. B 53, 12454 (1996).
- [51]Dataset for Ruf et al. “Controllable suppression of the unconventional superconductivity in Sr₂RuO₄ via high-energy electron irradiation” (2024); DOI to be inserted here.
- [52]L. Klein, Y. Kats, N. Wiser, M. Konczykowski, J. W. Reiner, T. H. Geballe, M. R. Beasley, and A. Kapitulnik, Negative deviations from Matthiesen’s rule for SrRuO₃ and CaRuO₃, Europhys. Lett. 55, 532 (2001).
- [53]N. Haham, M. Konczykowski, B. Kuiper, G. Koster, and L. Klein, Testing dependence of anomalous Hall effect on resistivity in SrRuO₃ by its increase with electron irradiation, Phys. Rev. B 88, 214431 (2013).DOI
- [54]A. Legris, F. Rullier-Albenque, E. Radeva, and P. Lejay, Effects of electron irradiation on YBa₂Cu₃O₇−δ superconductor, J. Phys. (Paris) I 3, 1605 (1993).
- [55]T. Li, H. Qing, and L. Xhengming, Analytic Fitting to the Mott Cross Section of Electrons, Radiat. Phys. Chem. 45, 235 (1995).
- [56]M. J. Boschini, C. Consolandi, M. Gervasi, S. Giani, D. Grandi, V. Ivanchenko, P. Nieminen, S. Pensotti, P. G. Rancoita, and M. Tacconi, An expression for the Mott cross section for electrons and positrons on nuclei with Z up to 118, Radiat. Phys. Chem. 90, 39 (2013).
- [57]M. Tinkham, Effect of Fluxoid Quantization on Transitions of Superconducting Films, Phys. Rev. 129, 2413 (1963).DOI
- [58]F. E. Harper and M. Tinkham, The Mixed State in Superconducting Thin Films, Phys. Rev. 172, 441 (1968).DOI
- [59]M. Uchida, M. Ide, M. Kawamura, K. S. Takahashi, Y. Kozuka, Y. Tokura, and M. Kawasaki, Anomalous enhancement of upper critical field in Sr2RuO4 thin films, Phys. Rev. B 99, 161111 (2019).
- [60]J. W. P. Hsu and A. Kapitulnik, Superconducting transition, fluctuation, and vortex motion in a two-dimensional single-crystal Nb film, Phys. Rev. B 45, 4819 (1992).pubmed.ncbi.nlm.nih.gov/10002120
- [61]A. A. Abrikosov and L. P. Gor’kov, Contribution to the theory of superconducting alloys with paramagnetic impurities, Sov. Phys. JETP 12, 1243 (1961).
- [62]P. Hohenberg, Anisotropic superconductors with non-magnetic impurities, Sov. Phys. JETP 18, 834 (1964).
- [63]A. A. Abrikosov, Influence of the gap anisotropy on superconducting properties, Physica C 214, 107 (1993).DOI
- [64]R. J. Radtke, K. Levin, H.-B. Schüttler, and M. R. Norman, Predictions for impurity-induced Tc suppression in the high-temperature superconductors, Phys. Rev. B 48, 653 (1993).DOI
- [65]L. S. Borkowski and P. J. Hirschfeld, Distinguishing d-wave superconductors from highly anisotropic s-wave superconductors, Phys. Rev. B 49, 15404 (1994).DOI
- [66]L. A. Openov, Critical temperature of an anisotropic superconductor containing both nonmagnetic and magnetic impurities, Phys. Rev. B 58, 9468 (1998).DOI
- [67]P. B. Allen, New method for solving Boltzmann’s equation for electrons in metals, Phys. Rev. B 17, 3725 (1978).DOI
- [68]A. A. Golubov and I. I. Mazin, Effect of magnetic and nonmagnetic impurities on highly anisotropic superconductivity, Phys. Rev. B 55, 15146 (1997).DOI
- [69]M. Björck and G. Andersson, GenX: an extensible X-ray reflectivity refinement program utilizing differential evolution, J. Appl. Crystallogr. 40, 1174 (2007).DOI
- [70]J. Paglione, C. Lupien, W. A. MacFarlane, J. M. Perz, L. Taillefer, Z. Q. Mao, and Y. Maeno, Elastic tensor of Sr2RuO4, Phys. Rev. B 65, 220506 (2002).DOI
- [71]A. Seneshyn, D. M. Trots, J. M. Engel, L. Vasylychko, H. Ehrenberg, T. Hansen, M. Berkowski, and H. Fuess, Anomalous thermal expansion in rare-earth gallium perovskites: a comprehensive powder diffraction study, J. Phys. Condens. Matter 21, 145405 (2009).
- [72]T. Vogt and D. J. Buttrey, Low-temperature structural behavior of Sr2RuO4, Phys. Rev. B 52, R9843 (1995).pubmed.ncbi.nlm.nih.gov/9980128
- [73]J. Wu, H. P. Nair, A. T. Bollinger, X. He, I. Robinson, N. J. Schreiber, K. M. Shen, D. G. Schlom, and I. Božović, Electronic nematicity in Sr2RuO4, Proc. Natl. Acad. Sci. USA 117, 10654 (2020).DOI
- [74]L. Capogna, A. P. Mackenzie, R. S. Perry, S. A. Grigera, L. M. Galvin, P. Raychaudhuri, A. J. Schofield, C. S. Alexander, G. Cao, S. R. Julian, and Y. Maeno, Sensitivity to Disorder of the Metallic State in the Ruthenates, Phys. Rev. Lett. 88, 076602 (2002).pubmed.ncbi.nlm.nih.gov/11863926
- [75]D. Forsythe, S. R. Julian, C. Bergemann, E. Pugh, M. J. Steiner, P. L. Alireza, G. J. McMullan, F. Nakamura, R. K. W. Haselwimmer, I. R. Walker, S. S. Saxena, G. G. Lonzarich, A. P. Mackenzie, Z. Q. Mao, and Y. Maeno, Evolution of Fermi-Liquid Interactions in Sr2RuO4 under Pressure, Phys. Rev. Lett. 89, 166402 (2002).
- [76]T. Akima, S. NishiZaki, and Y. Maeno, Intrinsic Superconducting Parameters of Sr2RuO4, J. Phys. Soc. Jpn. 68, 694 (1999).DOI
- [77]Y. Maeno, K. Yoshida, H. Hashimoto, S. Nishizaki, S.-i. Ikeda, M. Nohara, T. Fujita, A. P. Mackenzie, N. E. Hussey, J. G. Bednorz, and F. Lichtenberg, Two-Dimensional Fermi Liquid Behavior of the Superconductor Sr2RuO4, J. Phys. Soc. Jpn. 66, 1405 (1997).
- [78]N. E. Hussey, A. P. Mackenzie, J. R. Cooper, Y. Maeno, S. Nishizaki, and T. Fujita, Normal-state magnetoresistance of Sr2RuO4, Phys. Rev. B 57, 5505 (1998).DOI
- [79]M. E. Barber, A. S. Gibbs, Y. Maeno, A. P. Mackenzie, and C. W. Hicks, Resistivity in the Vicinity of a van Hove Singularity: Sr2RuO4 under Uniaxial Pressure, Phys. Rev. Lett. 120, 076602 (2018).










