arXiv:2007.13730v1 [cond-mat.supr-con] 27 Jul 2020 le othe to allel o-eoKr oainat of rotation onset Kerr observed non-zero the a from inferred was evidence Further tt tlwtmeaue rmasrnl orltdand correlated state. strongly normal a Fermi-liquid from ultraclean temperature, superconducting low unconventional at an state of emergence the for ugsino a of suggestion hoeia ok [ works theoretical t eysrn upeso ihnnmgei impuri- non-magnetic with [ by suppression indicated ties strong was same very state the ground its At unconventional an enhancement. time, [ mass normal-state experiments modest conductivity optical unity. [ and exceeding oscillations ratio quantum Wilson Meanwhile, measured in- the originally from as ferred correlations, ferromagnetic for evidence in- an with states for possibility the open plane leave far so candidate any recent clarification, [ [ heat conductivity specific thermal and example, For however. mained, chiral the to [ support of shift lent combination Knight surements the NMR after, re- from Soon therefore results and components. symmetry two reversal quires time state breaks chiral the which is symmetry by Sr allowed states in interest broad for reasons several seiec o oa a tutr [ structure gap nodal a for evidence as clpoete fSr of properties ical ecnutvt eerhsneisdsoeyi 94by 1994 [ in coworkers discovery and its Maeno since research perconductivity aaee ymty eonzdery[ early Recognized symmetry. parameter Sr 2 h eprtr n eddpnecso h NMR the of dependences field and temperature The ihteedvlpet nmn,w ealohrphys- other recall we mind, in developments these With naeigtesceso h uecnutn tt in state superconducting the of secrets the Unraveling RuO 20 d .Tu,Sr Thus, ]. shssnebe xlctydsusdi several in discussed explicitly been since has as , 4 [ 1 vdnefree aiyucnetoa superconductivit unconventional parity even for Evidence – c 3 eandeuiefrmr hnaqatrcnuy hl chi a While century. quarter a than more for elusive remained superfluid xetdfil-nue uspril epneol.A upp An only. of response sh response the quasiparticle that field-induced conclude expected We reported. previously measurements, field-dependent Here, ai [ -axis a enapirt o novninlsu- unconventional for priority a been has ] nmiuu dnicto ftesprodcigodrpa order superconducting the of identification Unambiguous p wv tt iha with state -wave p Jerzembeck wv rpe arn tt [ state pairing triplet -wave 10 13 11 2 3 2 xeiet eebt interpreted both were experiments ] RuO – 17 a lnkIsiuefrCeia hsc fSld,Dresde Solids, of Physics Chemical for Institute Planck Max RuO 1 , 15 eateto Physics of Department 3 .Chronister A. M esrmnsexcluded measurements NMR O 12 nvriyo tAdes ot ag,S nrw Y69S U 9SS, KY16 Andrews St Haugh, North Andrews, St of University 4 5 He- < 4 ctihUieste hsc line colo Physics of School Alliance, Physics Universities Scottish ]. o lmsNtoa aoaoy o lms e eio8754 Mexico New Alamos, Los Laboratory, National Alamos Los .Priual oal among notable Particularly ]. 4 .TeNRrslspublished results NMR The ]. 0 ftenra tt ucpiiiyi ucett exclud to sufficient is susceptibility state normal the of 10% 2 4 A ainlIsiuefrMtrasSine skb 305-000 Tsukuba Science, Materials for Institute National emdrlvn oteorder- the to relevant deemed 3 osiue oe system model a constitutes T .W Hicks W. C. , a ue u nyvr eety te proposed other recently, very only out ruled was c [ 8 .Ursle susre- issues Unresolved ]. 17 6 and ] 16 ngtsitmaueet r oprdt orsodn s corresponding to compared are measurements shift Knight O d vco lge par- aligned -vector 1 3 ,APS[ ARPES ], p † .I tptoward step a In ]. wv description. -wave , ∗ 19 2 µ .Pustogow A. RuO + 4 l indicate all ] 3 , .P Mackenzie P. A. , R[ SR & 5 z 5 4 .Among ]. ( a the was ] srnm,UL,LsAgls A905 USA; 90095, CA Angeles, Los UCLA, Astronomy, p Dtd uy2,2020) 28, July (Dated: a the was x 7 17 ± mea- ] , ip 18 y 1 9 ), ], † ] , ∗ .Kikugawa N. loe ttswith states allowed rsa ne td,wsmutdo igeai piezo- single-axis a on mounted was study, under crystal o lmntdaesae hrceie by characterized states are eliminated Not ln edecue h hrlsae[ in- state applied chiral an the for excludes shift field Knight plane the of reduction observed pne ywyo oprn opeiul eotdspe- re- reported results previously magnetic heat to the comparing cific of to way contributes By also the sponse. by and controlled strength, is field creation Quasiparticle feasible. tally e on o h odnaeprinto portion condensate the for bound per ue yhg-eprtr neln [ annealing high-temperature 17 by duced discussed the ap- and approach, in-plane, [ weak another Refs. take very in we for previously Here, reduced fields. are plied strength resolution signal spectral because and challenging particularly iments belled eeal o triplet-paired, for generally ngtshifts Knight eaue(5m)wievarying while mK) (25 perature ntesprodcigpae ozr susceptibility nonzero a phase, χ superconducting the In with B ntenra state, normal the In o[0] e Fig. mm), see parallel 0.2 dimension [100], longest x to the out-of- mm and the 1 to [001]-direction, corresponding plane x dimension mm shortest Single- the (3.5 with were configuration. we dimensions tuning/matching octaves, crystal top several a covering adopted frequencies low relatively re-aaee ymty npriua o Sr for particular in symmetry, order-parameter au htcnrdcsteepcainfrayo the of any for expectation the contradicts pure that value a sc c 3 O tfis ih,tems ietwyt etfrsymmetry- for test to way direct most the sight, first At si rvosNRsuiso Sr on studies NMR previous in As 2 , 4 , 2 [001] soitdwt odnaeplrzto sexpected is polarization condensate with associated .D Bauer D. E. , p B topeea 1050 at atmosphere wv re aaeesrlvn o Sr for relevant parameters order -wave 17 rbudfrtecnest magnetic condensate the for bound er k f eut a eacutdfrb the by for accounted be can results ift ( O < ral c aee ymtyo Sr of symmetry rameter Gi eysal[ small very is kG 1 2 oee,terlvn pe rtclfield critical upper relevant the However, . 17 .A Sokolov A. D. , p p K I wv rudsaeaaou to analogue state ground -wave wv cnro r tl viable. still are scenarios -wave =5/2, s ( C 17 T < T 18,Germany; 01187, n n Astronomy, and 5 1 e n .E Brown E. S. and , shifts O a.TeNRci otiigthe containing coil NMR The (a). ( 12 K d d-aiycandidates. odd-parity e B ,Japan; 3, 17 , s ) ⊥ /T 22 ∼ γ c ,USA; 5, , -.7 H/ [ MHz/T =-5.772 c :tefil retto sfixed is orientation field the ]: χ B ;and K; [ K n st efr measurements perform to is 10 nSr in y ◦ r rca rb fthe of probe crucial a are ) 3 p with , .T aiiaeacs to access facilitate To C. s F. , wv tts ec,the Hence, states. -wave , r vlae tlwtem- low at evaluated are B 23 21 eicheat pecific 2 ,w siaeteup- the estimate we ], smc sexperimen- as much as RuO aigsc exper- such making ] χ 11 n 2 2 1 RuO ,frwhich for ], h susceptibility. the RuO 4 † 6 has χ ,hr n90% in here ], sc d 24 4 /χ 4 2 ⊥ [ )i intro- is ]) RuO 6 n ,tela- the ], c . 2 < RuO d 4 10%, . k c 4 . . 2 rotator inside the mixing chamber of a bottom-loading [001] (a) O(2) (b) O(2) T = 25 mK dilution refrigerator. Sample alignment enabled in-plane [010] ◦ B = 1.50 T orientation to within 0.2 , based on RF susceptibility O(1 ) ± ⊥ O(1||) Ru O(1 ) measurements sensitive to Bc2, described in Ref. [11], and O(1||) ⊥ (arb. units) Intensity NMR discussed in the Supplemental Material [25]. 63Cu NMR [100] O(1 ) O(1 ) || relaxation rate measurements were used to determine the ⊥ Ep = 40 nJ equilibrium bath temperature T = 25 mK. As in our pre- B vious work [11], low-power RF experiments were carried O(2) 4 nJ out to make sure the results were not measurably altered 1.0

) (d) (c) heating by B = 1.38 T by RF pulse heating effects. The applied field strength =0 s

K 0.8 NMR pulses f normal B was determined to within uncertainties less than 10’s - normal 3 3 4 of µT from the NMR resonance of He in the He/ He 0.6

normal 130 nJ mixture of the dilution refrigerator. f )/( 0.4 0 500 1000 SC SC =0 Addressed first are sample heating effects by the RF s B (T) K 0.8

f 40 nJ p - b⋅E 1.14 pulses, illustrated in Fig. 1, which turned out to be a ⊥ a+ 1 0.2 0.6

f 0.94 crucial issue [11, 12]. So as to enhance sensitivity to ( 0.4 0.73 0.4 nJ this potential artifact, we examined the transients with 0.0 1 10 100 1000 -40 -20 0 20 40 the field set to 1.38 T, a value just smaller than B . c2 E (nJ) f-f (kHz) Clear evidence for warming by the RF pulsing is in- p 0 ferred from a transient response corresponding to that FIG. 1. (a) Sr2RuO4 involves three distinct oxygen sites for of the normal-state (instead of the sought-after super- field direction B k [100]. (b) The three associated 17O NMR conducting state). Shown in Fig. 1(b,c) are 17O spectra central transitions (O(1k), O(2), O(1⊥) from left to right) are corresponding to central transitions for the three sites, independent of pulse energy Ep at 1.50 T > Bc2 ≃ 1.45 T. O(1k,2,1⊥), at applied magnetic fields slightly above and (c) Also at B = 1.38T . Bc2 the normal-state spectrum is −7 below Bc2. While at 1.5 T >Bc2 the line shape remains observed for Ep ≥ 10 J. Reducing to Ep = 40 nJ leads unaffected by changing the pulse energy, a normal state to doubled spectral features, most pronounced for O(1k,⊥), spectrum is produced also at 1.38 T < B when using which we assign to coexisting normal (dashed vertical lines) c2 and superconducting (solid) contributions around the first- a pulse energy Ep = 130 nJ. Decreasing Ep to 40 nJ order transition. Further reduction of Ep reveals the pure leads to a response where a new spectral line appears for superconducting-state spectrum. (d) O(1⊥) frequencies nor- each site, indicating the coexistence of normal and super- malized to normal-state (fnormal) and zero-shift (fKs=0; see conducting phases. This data set is particularly useful, Fig. 2) positions at B < Bc2 for variable Ep. Linear fits since the macroscopic phase segregation provides a quan- (solid lines, see inset) indicate that heating is less problem- titative measure of the magnetization jump ∆M at the atic at lower field due to larger Tc(B). Knight shifts Ks were discontinuous (first-order) transition [26, 27]. Note that determined using the frequency values leveling off at Ep → 0. these data are recorded following a single-pulse excita- tion. That is, the transient NMR response corresponds to a free induction decay (FID). All shift results of the are transition frequencies at each field, generated using present work were obtained from FID measurements car- the known normal state NMR parameters [25]. Then, the ried out with RF pulse energies low enough to avoid heat- frequency differences between closed and open symbols ing, as illustrated in Fig. 1(d). are proportional to the hyperfine fields, and constitute Having established a threshold for heating effects, the product of (normal-state) Knight shifts with applied 17 we now inspect the spectra recorded at variable field field, Ks,normal γB, for O(1k), O(2) and O(1⊥). When strength. In Fig. 2 we show the NMR intensity as a decreasing the field B

0.6 (a) B

[100] T = 25 mK T = 25 mK

10 K = 0 K s 1,⊥ normal B || [100] O(1 ) K ⊥ normal 0.4 O(2) O(1 ) | normal B (T) || superconducting O(2) 8 O(1 ) O(1 ) state

1.6 ||

(%) 0.2

1.5 K

6 0 = s K

1.34 0.0

4 1.14 K1,|| normal -0.2 0.94 Bc2

2 NMRintensity units) (arb. (b) 0.74 1.0 B || [110] B || [100] 0.5 normal ε = 0 ε = 0 ε = εv 0 state 0.8 K2 K -20 0 20 40 60 80 100 1,⊥ K 17 1,||

f - B (kHz) 0.6

normal T = 25 mK K

17 /

FIG. 2. Spectra for central O NMR transitions at different s

K 0.4 field strengths, for O(1k), O(2), O(1⊥) sites, respectively left- right, plotted as intensity vs. f−17γB. The dotted curves running vertically through the spectra follow the expected 0.2 C/T from Ref. [23]: field dependence after taking into account quadrupolar and 0.09 K orbital couplings; the dashed curves also include the normal- extrap. to T=0 state hyperfine fields. See Ref. 25 for details of quadrupolar 0.0 and orbital contributions to the transition frequencies, as well 0.0 0.2 0.4 0.6 0.8 1.0 1.2 as an analysis of the sample orientation relative to B. B B / c2

FIG. 3. (a) NMR shifts K = Ks + Ko determined from the as well as the strong increase of the O(1k) quadrupo- lar component at low fields, we limited the measure- spectra in Fig. 2. While the shifts are positive and the as- signed Ko ≃ 0.0% for O(2) and O(1 ), the O(1 ) line occurs ments to B 0.24 T. In addition to the well-known ⊥ k ≥ at a positive value Ko = 0.18% at B = 0 and K1,k < 0 [6, 28]. quadrupolar effects, one has to include purely orbital (b) The field-dependent drop of NMR Knight shift determined contributions. These were evaluated in Ref. [6], yielding in the present work at T = 25 mK is compared to specific heat Ko = +0.18% for the O(1k) site and a value indistin- C/T recorded at T = 90 mK [23] as well as its T = 0 extrapo- guishable from zero for O(1⊥) and O(2). See Ref. 25 for lation [30], all normalized to the normal state value. The val- further comment. ues of Ks coincide with the zero-temperature extrapolations The shifts K , are plotted as a function of B in of C/T , providing compelling evidence that this is the con- 1k,2,1⊥ tribution of unpaired quasiparticles in the superconducting Fig. 3. Results are shown in panel (a) as total shift, state. Measurements along [110] (small open symbols) reveal K = Ks + Ko. In the normal state, K1k < 0, while a similar jump at the transition and also uniaxial strain re- K2,1⊥ > 0; each exhibits a reduction in the supercon- sults (open cyan symbols, B k [100], εaa = εv) from Ref. 11 ducting state. Bc2 is marked by the discontinuous change coincide at low B/Bc2. of each of the three sites, accompanied by a coexistence regime [cf. Fig. 1(b,c)]. Consistent with expectations (B Bc1)[31], the results indicate that diamagnetic bution), both normalized to the normal state. As shown, ≫ shielding is a small effect. Otherwise, the discontinuous the field-induced trends are similar, and particularly rele- drop ∆M (Figs. 1,2) would be similar for all three sites. vant to the open question of order-parameter symmetry. Instead, only the hyperfine field, which is much greater Simply put, at non-zero field, a Zeeman-like NMR re- for the planar sites than it is for the apical site, and sponse can originate from quasiparticles, and, in the case opposite in sign for O(1k) relative to O(2) and O(1⊥), of triplet pairing, also from the condensate. In contrast, decreases on entering the superconducting state. the specific heat is sensitive only to the quasiparticle re- The main results of this work are displayed in Fig. 3(b), sponse with no contribution from the condensate. As where the Knight shifts are compared to previous heat can be seen by inspection of Fig. 3(b), we observe no capacity results [23], Ce(B)/T (Ce the electronic contri- systematic difference between the T 0 extrapolation → 4 of the heat capacity data of Ref. [23] and the spin suscep- to see how the quest to finalize identification of the or- tibility deduced from our measurements. Taking into ac- der parameter of Sr2RuO4 develops. We believe that by count systematic uncertainties primarily associated with ruling out any pure p-wave order parameter possibility, the oxygen orbital shifts Ko, we estimate an upper limit the research we have reported here makes a significant for the condensate response of < 10% of that of the nor- contribution to that process. mal state, for fields applied both along [100] and [110] Note added: We recently learned of a proposal [41] for (see Ref. 25 for comments pertaining to Ko). Similar a mixed-parity order parameter of the form d ip [41], ± K1k,⊥ are found at B/Bc2 = 0.17 under strained condi- which would result in a reduced condensate response, rel- tions [11], where Ko is less of an issue due to larger field ative to those of the the pure p-wave states discussed resulting from the enhanced Tc and Bc2 [32]. These ob- above. For example, Ks/Kn 0.2 (chiral p-wave com- ≃ servations place such strong constraints on the magnetic ponent), Ks/Kn 0.1 (helical p-wave component). The polarizability of the condensate that we believe that they latter is at the upper≃ bound of the sensitivity of our rule out any pure p-wave order parameter for the super- present experiments. conducting state of Sr2RuO4, as we now discuss. We thank Thomas Scaffidi for sharing his manuscript The p-wave order parameters most commonly dis- with us prior to publication, and Steve Kivelson for a cussed in the context of Sr2RuO4 are the so-called chiral number of helpful discussions and for commenting on (ˆz(px ipy)) and helical (pxxˆ + pyyˆ) states. Assuming our manuscript. A.C. is grateful for support from the that the± unit vectors encoding spin directions are pinned Julian Schwinger Foundation for Physics Research. A.P. to the lattice, they are predicted in the simplest models to acknowledges support by the Alexander von Humboldt result in condensate polarizabilities of 100% (chiral) and Foundation through the Feodor Lynen Fellowship. Work 50% (helical) of the normal state value. The chiral state at Los Alamos was funded by Laboratory Directed Re- was ruled out by our previous work [11], but the helical search and Development (LDRD) program, and A.P. ac- state and certain others were not. The data presented in knowledges partial support through the LDRD. N. K. Fig. 3 allow us to go much further. Even after considering acknowledges the support from JSPS KAKNHI (Grant Fermi-liquid corrections [12] and the effects of spin-orbit No. 18K04715). The work at UCLA was supported by coupling [14], it is unclear how to reconcile an upper limit the National Science Foundation, grant number DMR- of 10% of the normal state susceptibility with any p-wave 1709304. A.C. and A.P. contributed equally. state. One could also postulate extreme situations such as a momentum independent d aligned along either [100] or [110], or an unpinned d free to rotate in response to the applied field. None can predict a spin susceptibility ∗ suppression that would be compatible with our results; These authors contributed equally † Address correspondence to: [email protected], a few remaining possibilities have been ruled out by our [email protected], [email protected]. use of both [100] and [110] fields in the current experi- [1] Andrew P. Mackenzie, Thomas Scaffidi, Clifford W. ments. We therefore assert that our measurements have Hicks, and Yoshiteru Maeno, “Even odder after twenty- ruled out any pure p-wave order parameter candidate for three years: the superconducting order parameter puzzle the superconducting state of Sr2RuO4. of Sr2RuO4,” npj Quant. Mater. 2, 40 (2017). Given this input, we close with an evaluation of the [2] Catherine Kallin, “Chiral p-wave order in Sr2RuO4,” Rep. Prog. Phys. 75, 042501 (2012). current understanding of superconductivity in Sr RuO . 2 4 [3] Andrew Peter Mackenzie and Yoshiteru Maeno, “The In isolation, our NMR findings are consistent with superconductivity of Sr2RuO4 and the physics of spin- even-parity states, such as dx2−y2 (B1g), dxy (B2g) or triplet pairing,” Rev. Mod. Phys. 75, 657–712 (2003). dxz; dyz (E g). Indeed, STM measurements are inter- [4] Y. Maeno, H. Hashimoto, K. Yoshida, S. Nishizaki, { } 1 preted as most consistent with the B1g state [33], similar T. Fujita, J. G. Bednorz, and F. Lichtenberg, “Supercon- to thermal transport experiments [9]. Accordingly, Ce/T ductivity in a layered perovskite without copper,” Nature follows the expected √B dependence [34] in the zero- 372, 532–534 (1994). [5] T. M. Rice and M. Sigrist, “Sr2RuO4: an electronic ana- temperature limit. 3 logue of He?” J. Phys.: Condens. Matter 7, L643–L648 However, several other experiments must be accounted (1995). for. These include recent ultrasound reports of a discon- [6] K. Ishida, H. Mukuda, Y. Kitaoka, K. Asayama, Z. Q. tinuity in the elastic constant c66 [35, 36], which would re- Mao, Y. Mori, and Y. Maeno, “Spin-triplet superconduc- 17 strict possible states to two-component candidates, such tivity in Sr2RuO4 identified by O knight shift,” Nature 396, 658–660 (1998). as dxz; dyz . Time-reversal symmetry breaking may prove{ crucial} [37]. Following the two-component hypoth- [7] G. M. Luke, Y. Fudamoto, K. M. Kojima, M. I. Larkin, J. Merrin, B. Nachumi, Y. J. Uemura, Y. Maeno, esis, while avoiding a reliance on interplanar pairing, has Z. Q. Mao, Y. Mori, H. Nakamura, and M. Sigrist, led to considering the possibility for coupling between “Time-reversal symmetry-breaking superconductivity in accidentally nearly degenerate single-component states Sr2RuO4,” Nature 394, 558 (1998). such as d 2 2 ; g 2 2 [38–40]. It will be intriguing [8] Jing Xia, Yoshiteru Maeno, Peter T. Beyersdorf, M. M. { x −y xy(x −y )} 5

Fejer, and Aharon Kapitulnik, “High resolution Po- G. G. Lonzarich, Y. Mori, S. Nishizaki, and Y. Maeno, lar Kerr Effect measurements of Sr2RuO4: Evidence for “Extremely strong dependence of superconductivity on broken time-reversal symmetry in the superconducting disorder in Sr2RuO4,” Phys. Rev. Lett. 80, 161–164 state,” Phys. Rev. Lett. 97, 167002 (2006). (1998). [9] E. Hassinger, P. Bourgeois-Hope, H. Taniguchi, S. Ren´e [21] a-axis stress increases Bc2 significantly by this measure, de Cotret, G. Grissonnanche, M. S. Anwar, Y. Maeno, see Ref. [32]. N. Doiron-Leyraud, and Louis Taillefer, “Vertical line [22] Yuujirou Amano, Masahiro Ishihara, Masanori Ichioka, nodes in the superconducting gap structure of Sr2RuO4,” Noriyuki Nakai, and Kazushige Machida, “Pauli para- Phys. Rev. X 7, 011032 (2017). magnetic effects on mixed-state properties in a strongly [10] Shunichiro Kittaka, Shota Nakamura, Toshiro Sakak- anisotropic superconductor: Application to Sr2RuO4,” ibara, Naoki Kikugawa, Taichi Terashima, Shinya Uji, Phys. Rev. B 91, 144513 (2015). Dmitry A. Sokolov, Andrew P. Mackenzie, Koki Irie, [23] Shuji NishiZaki, Yoshiteru Maeno, and Zhiqiang Mao, Yasumasa Tsutsumi, Katsuhiro Suzuki, and Kazushige “Changes in the superconducting state of Sr2RuO4 under Machida, “Searching for gap zeros in Sr2RuO4 via field- magnetic fields probed by specific heat,” J. Phys. Soc. angle-dependent specific-heat measurement,” J. Phys. Jpn. 69, 572–578 (2000). Soc. Jpn. 87, 093703 (2018). [24] Robin K. Harris, Edwin D. Becker, Sonia M. Cabral [11] A. Pustogow, Yongkang Luo, A. Chronister, Y.-S. Su, de Menezes, Robin Goodfellow, and Pierre Granger, D. A. Sokolov, F. Jerzembeck, A. P. Mackenzie, C. W. “NMR nomenclature. nuclear spin properties and con- Hicks, N. Kikugawa, S. Raghu, E. D. Bauer, and S. E. ventions for chemical shifts(IUPAC recommendations Brown, “Constraints on the superconducting order pa- 2001),” Pure Appl. Chem. 73, 1795–1818 (2001). rameter in Sr2RuO4 from oxygen-17 nuclear magnetic [25] See Supplemental Material for details on the discontinu- 17 resonance,” Nature 574, 72–75 (2019). ous transition at Bc2, the different contributions to O [12] Kenji Ishida, Masahiro Manago, Katsuki Kinjo, and NMR shifts and sample alignment with respect to the Yoshiteru Maeno, “Reduction of the 17O Knight Shift external magnetic field. in the superconducting state and the heat-up effect by [26] Shingo Yonezawa, Tomohiro Kajikawa, and Yoshiteru NMR pulses on Sr2RuO4,” J. Phys. Soc. Jpn. 89, 34712 Maeno, “First-order superconducting transition or (2020). Sr2RuO4,” Phys. Rev. Lett. 110, 077003 (2013). [13] A. T. Rømer, D. D. Scherer, I. M. Eremin, P. J. [27] Shingo Yonezawa, Tomohiro Kajikawa, and Yoshiteru Hirschfeld, and B. M. Andersen, “Knight shift and lead- Maeno, “Specific-heat evidence of the first-order super- ing superconducting instability from spin fluctuations in conducting transition in Sr2RuO4,” J. Phys. Soc. Jpn. sr2ruo4,” Phys. Rev. Lett. 123, 247001 (2019). 83, 083706 (2014). [14] Henrik S. Røising, Thomas Scaffidi, Felix Flicker, Gun- [28] T. Imai, A. W. Hunt, K. R. Thurber, and F. C. Chou, nar F. Lange, and Steven H. Simon, “Superconducting “17O nmr evidence for orbital dependent ferromagnetic order of Sr2RuO4 from a three-dimensional microscopic correlations in Sr2RuO4,” Phys. Rev. Lett. 81, 3006–3009 model,” Phys. Rev. Res. 1, 033108 (2019). (1998). [15] Austin W. Lindquist and Hae-Young Kee, “Dis- [29] Yongkang Luo, A. Pustogow, P. Guzman, A. P. Dio- tinct reduction of knight shift in superconducting guardi, S. M. Thomas, F. Ronning, N. Kikugawa, D. A. state of Sr2RuO4 under uniaxial strain,” (2019), Sokolov, F. Jerzembeck, A. P. Mackenzie, C. W. Hicks, E. arXiv:1912.02215 [cond-mat.supr-con]. D. Bauer, I. I. Mazin, and S. E. Brown, “Normal state 17 [16] A. P. Mackenzie, S. R. Julian, A. J. Diver, G. J. McMul- O nmr studies of Sr2RuO4 under uniaxial stress,” Phys. lan, M. P. Ray, G. G. Lonzarich, Y. Maeno, S. Nishizaki, Rev. X 9, 021044 (2019). and T. Fujita, “Quantum oscillations in the layered per- [30] We also note that recent specific heat measurements ovskite superconductor Sr2RuO4,” Phys. Rev. Lett. 76, [10, 18] differ from those of Ref. 23 by finding a larger 3786–3789 (1996). residual electronic specific heat at low temperatures. [17] A. Damascelli, D. H. Lu, K. M. Shen, N. P. Armitage, While sample quality is the most obvious source of such F. Ronning, D. L. Feng, C. Kim, Z.-X. Shen, T. Kimura, a discrepancy, it merits further experimental attention. Y. Tokura, Z. Q. Mao, and Y. Maeno, “Fermi surface, However, since those results indicate a larger quasiparti- surface states, and surface reconstruction in Sr2RuO4,” cle contribution than that of Ref. 23, they do not invali- Phys. Rev. Lett. 85, 5194–5197 (2000). date our main conclusion which is that we see no evidence [18] A. Tamai, M. Zingl, E. Rozbicki, E. Cappelli, S. Ricc`o, for a condensate contribution to the spin susceptibility. A. de la Torre, S. McKeown Walker, F. Y. Bruno, P. [31] H. Murakawa, K. Ishida, K. Kitagawa, H. Ikeda, Z. Q. D. C. King, W. Meevasana, M. Shi, M. Radovi´c, N. Mao, and Y. Maeno, “101Ru Knight Shift measure- C. Plumb, A. S. Gibbs, A. P. Mackenzie, C. Berthod, ment of superconducting Sr2RuO4 under small magnetic H. U. R. Strand, M. Kim, A. Georges, and F. Baum- fields parallel to the RuO2 plane,” J. Phys. Soc. Jpn. 76, berger, “High-resolution photoemission on Sr2RuO4 re- 024716 (2007). veals correlation-enhanced effective spin-orbit coupling [32] Alexander Steppke, Lishan Zhao, Mark E. Barber, and dominantly local self-energies,” Phys. Rev. X 9, Thomas Scaffidi, Fabian Jerzembeck, Helge Rosner, 021048 (2019). Alexandra S. Gibbs, Yoshiteru Maeno, Steven H. Simon, [19] D. Stricker, J. Mravlje, C. Berthod, R. Fittipaldi, A. Vec- Andrew P. Mackenzie, and Clifford W. Hicks, “Strong chione, A. Georges, and D. van der Marel, “Optical re- peak in Tc of Sr2RuO4 under uniaxial pressure,” Science sponse of Sr2RuO4 reveals universal Fermi-Liquid scaling 355, eaaf9398 (2017). and quasiparticles beyond Landau Theory,” Phys. Rev. [33] Rahul Sharma, Stephen D. Edkins, Zhenyu Wang, An- Lett. 113, 087004 (2014). drey Kostin, Chanchal Sow, Yoshiteru Maeno, Andrew P. [20] A. P. Mackenzie, R. K. W. Haselwimmer, A. W. Tyler, Mackenzie, J. C. S´eamus Davis, and Vidya Madha- 6

van, “Momentum-resolved superconducting energy gaps of Sr2RuO4 from quasiparticle interference imaging,” Proc. Nat. Acad. Sci. USA 117, 5222–5227 (2020). [34] Y. Matsuda, K. Izawa, and I. Vekhter, “Nodal struc- ture of unconventional superconductors probed by angle resolved thermal transport measurements,” J. Phys. Con- dens. Matter 18, R705–R752 (2006). [35] Sayak Ghosh, Arkady Shekhter, F. Jerzembeck, N. Kiku- gawa, Dmitry A. Sokolov, Manuel Brando, A. P. Macken- zie, Clifford W. Hicks, and B. J. Ramshaw, “Thermo- dynamic evidence for a two-component superconducting Order Parameter in Sr2RuO4,” (2020), arXiv:2002.06130 [cond-mat.supr-con]. [36] 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, “Jump in the c66 shear modulus at the superconducting transition of Sr2RuO4: Evidence for a two-component order parame- ter,” (2020), arXiv:2002.05916 [cond-mat.supr-con]. [37] Vadim Grinenko, Shreenanda Ghosh, Rajib Sarkar, Jean- Christophe Orain, Artem Nikitin, Matthias Elender, De- barchan Das, Zurab Guguchia, Felix Brckner, Mark E. Barber, Joonbum Park, Naoki Kikugawa, Dmitry A. Sokolov, Jake S. Bobowski, Takuto Miyoshi, Yoshiteru Maeno, Andrew P. Mackenzie, Hubertus Luetkens, Clif- ford W. Hicks, and Hans-Henning Klauss, “Split super- conducting and time-reversal symmetry-breaking tran- sitions, and magnetic order in Sr2RuO4 under uniaxial stress,” (2020), arXiv:2001.08152 [cond-mat.supr-con]. [38] Igor Zuti´candˇ Igor Mazin, “Phase-sensitive tests of the pairing state symmetry Sr2RuO4,” Phys. Rev. Lett. 95, 217004 (2005). [39] Han Gyeol Suh, Henri Menke, P. M. R. Brydon, Carsten Timm, Aline Ramires, and Daniel F. Agterberg, “Stabi- lizing even-parity chiral superconductivity in Sr2RuO4,” (2019), arXiv:1912.09525 [cond-mat.supr-con]. [40] Steven Allan Kivelson, Andrew Chang Yuan, Brad Ramshaw, and Ronny Thomale, “A proposal for recon- ciling diverse experiments on the superconducting state in Sr2RuO4,” npj Quantum Mater. 5, 43 (2020). [41] Thomas Scaffidi, (private communication). [42] B. N. Figgis, R. G. Kidd, and R. S. Nyholm, “Oxygen- 17 nuclear magnetic resonance of inorganic compounds,” Proceedings of the Royal Society of London. Series A. Mathematical and Physical Sciences 269, 469–480 (1962). [43] C. H. Pennington, D. J. Durand, C. P. Slichter, J. P. Rice, E. D. Bukowski, and D. M. Ginsberg, “Static and dynamic Cu NMR tensors of YBa2Cu3O7−δ,” Phys. Rev. B 39, 2902–2905 (1989). [44] Hidekazu Mukuda, Kenji Ishida, Yoshio Kitaoka, Ku- nisuke Asayama, Zhiqiang Mao, Yasumitsu Mori, and Yoshiteru Maeno, “Novel Character of Spin Fluctuations in Spin-Triplet Superconductor Sr 2RuO 4: 17O-NMR Study,” J. Phys. Soc. Jpn. 67, 3945–3951 (1998). 7

Supplemental Material

Evidence for even parity unconventional superconductivity in Sr2RuO4 A. Chronister1†⋆, A. Pustogow1†⋆, N. Kikugawa2, D. A. Sokolov3, F. Jerzembeck3, C. W. Hicks3, A. P. Mackenzie3,4, E. D. Bauer5, S. E. Brown1† 1Department of Physics & Astronomy, UCLA, Los Angeles, CA 90095, USA; 2National Institute for Materials Science, Tsukuba 305-0003, Japan; 3Max Planck Institute for Chemical Physics of Solids, Dresden 01187, Germany; 4Scottish Universities Physics Alliance, School of Physics and Astronomy, University of St Andrews, North Haugh, St Andrews KY16 9SS, UK; 5Los Alamos National Laboratory, Los Alamos, New Mexico 87545, USA.

The discontinuous transition at Bc2(T → 0)

While at low temperature there is good evidence that the thermal and magnetic responses both originate from field-induced quasiparticles, the thermodynamic discontinuities at the first-order transition at Bc2(T ) were previously explored in some detail [22, 26, 27]. These are thermodynamically constrained, expressed as the appropriate Clausius- Clapeyron relation, e.g.,

dB ∆S c2 = . dTc −∆M While ∆M is the total magnetization and includes diamagnetic shielding, it is dominated by the hyperfine part in Sr2RuO4 [22, 31]. The phase transition was found discontinuous for T . 0.8 K, with specific heat [27] and MagnetoCaloric Effect [26] measurements extending to temperatures as low as 90 mK. At T =200 mK, the entropy 2 jump is quoted as 10% of the normal state value, ∆S = (0.1)γN Tc, where γn= 37.5 mJ/mol-K . Also reported at 200 mK, dBc /dTc= -0.2 T/K. These combined results lead to the expectation ∆M(200mK) .6 .7χnBc (T = 200mK. 2 ≃ − 2 Since in these cases the Zeeman energy is much larger than the thermal energy scale, the fractional entropy of the transition is not expected to be strongly temperature-dependent. In that case, the magnetization discontinuity inferred from the measurements reported here compare favorably. That is, the experiments indicate a slightly smaller −3 discontinuity, ∆Ms 0.4χnBc with χsc 0.9(10) emu/mole. ≃ 2 ≃

Comment on the orbital shift for 17O

17 The O orbital shifts Ko are of considerable importance here, since we use them to reference the Knight shifts Ks(1k, 1⊥, 2) relative to “0”. Generally, Ko can extend to greater than +1500 ppm in oxygen-containing molecules, that are dominated by the paramagnetic term, and are linked to an increased 2p-bonding order [42]. In this case, the 2p σ non-bonding orbital is 3/4-filled, and the bonding 2p π orbitals of most interest here, are 7/8 filled [29]. − ∼ − ∼ in Ref. [6], the orbital shifts Ko(1 , 2) were emprically found neglibibly small by way of a so-called K χ plot, where ⊥ − temperature is an implicit parameter. In a similar fashion, Ko(1 ) 0.18%. These values were applied to the analysis k ≃ leading to Fig. 3, although we used Ko(2) = 0.02%. More specifically, if Ko(1k) is taken as vanishingly small, the Knight shift Ks(1k), originating dominantly with the p π orbitals, acquires the wrong (unphysical) sign. See Fig. 3(a), where total shift is seen to change sign as the field-induced− quasiparticle density is monotonically increased with the field strength. The paramagnetic orbital shift for O(1k) implies it follows from the specific unquenching of the angular momentum in the p π states. These states, in hybridizing with the Ru t2g orbitals form the bands α, β, γ crossing the Fermi surface.− Thus, the perturbative methods applied somewhat successfully to the cuprate superconductors [43] may be less useful here. Note also that if we adopt Ks(1⊥) > 0, a similarly unphysical field-dependent sign change is imposed on the corresponding hyperfine part, and the diamagnetic part is expected less than .02%. This is the clearest constraint on setting the stated upper bound to the condensate fraction of the shift to < 10% of Knormal. Finally, more relative uncertainty is associated with the O(2) site. Ko=0.0% was assumed in Ref. [28]. Here, we took it as +0.02%, also small on the scale of oxygen paramagnetic orbital shifts, and just 25% compared to inferred normal state hyperfine (Knight) shift. Using this value, the results for O(2) match those for O(1k, 1⊥). 8

Nuclear Hamiltonian Parameterization

17 5 For the O nucleus with spin I = 2 , the nuclear spin hamiltonian consists of two parts:

H = HQ + Hz (S1)

Hz = γI~ (1 + K) B~ (S2) · · eQ HQ = I~ V I~ (S3) 2I(2I 1)~ · · −

where Hz is the Zeeman interaction and HQ is the nuclear quadrupole interaction. K = Ks + Korb is the total shift tensor, including both orbital and hyperfine contributions. Q is the electric quadrupole moment of the nucleus and V is the electric field gradient (EFG). The quadrupolar term has the general effect of splitting the degenerate Zeeman transitions, resulting in 2I resonance frequencies. Thus, in the case of Sr2RuO4, which has three distinct oxygen sites under the application of in-plane field (B a), we expect a full 17O spectrum of 15 lines. || By measuring these NMR lines, one can probe the electronic spin susceptibility χ of the material via the strength of the hyperfine interaction (Ks). However, as Equations (1-3) imply, this effect must be differentiated from the other interactions contributing to the total Hamiltonian. If the parameters defining the nuclear quadrupole interaction are known, this can be readily done. Fortunately, the hyperfine shift, orbital shift, and EFG tensors are all independent of field in the normal state, making it possible to determine these parameters experimentally. So, by measuring 17 all fifteen O resonances at various fields in the normal state (B>Bc2), one can overdetermine the normal state shift tensor Knorm and the EFG tensor V . Then, since the quadrupolar and orbital shifts are independent of the superconducting phase transition, any discrepancy between the expected and measured resonance frequencies below Bc2 can be directly attributed to changes in Ks, hence the spin susceptibility. The nuclear spin Hamiltonian is canonically expressed in the principle axes system (PAS) of V , using the convention for the diagonal entries Vzz Vxx Vyy. By doing this, the quadrupolar part can be written compactly as: ≥ ≥

νQ 2 2 2 2 HQ = [3I I~ + η(I I )] (S4) 6 z − x − y

where, νQ is the principle axis NQR frequency, proportional to Vzz and η is the asymmetry parameter given by (Vxx Vyy)/Vzz. Since the shift tensor is also diagonal in this frame for all three oxygen sites in Sr2RuO4, the Zeeman term− can be expressed as

Hz = γ[Bx(1 + Kxx)Ix + By(1 + Kyy)Iy + Bz(1 + Kzz)Iz] (S5) where B~ is also written in the EFG basis. With the Hamiltonian written in this form, there are 7 parameters to be determined for each oxygen site: νQ, η, the three PAS components of K, and the two angles relating Bˆ to the EFG frame, θ and φ. However, as explained below, both the angles determining Bˆ can be measured independently– leaving only 5 parameter to be determined via normal state measurements.

Sample Alignment With Respect to Magnetic Field Direction

First, the out-of-plane angle can be determined independent of the NMR spectrum by utilizing the extreme anisotropy in the upper critical field for B ab and B c. Bc2 reaches a maximum of around 1.45 T with the field aligned directly in plane [26]. As mentioned|| in the main|| text, the NMR coil containing the sample is mounted on a piezoelectric step rotator with rotation axis perpendicular to the applied field. By activating the piezo until Bc2 ◦ reaches a maximum, the in-plane condition can be aligned to within 0.2 . The angle dependence of Bc2 is shown in Fig. S1. ± The in-plane angle is then checked by a posteriori visual inspection of the sample mounting using a microscope to ◦ view the sample orientation. While the in-plane condition was verified by anisotropy of Bc2 in Fig. S1,a3 angle is found between the long axis of the single crystal and the magnetic field direction. 9

1.0

c2,max 0.9 B / c2 B

Ref. [S1]

0.8

present work

-3 -2 -1 0 1 2 3

(°)

FIG. S1. The angle dependence of the upper critical field Bc2, determined from the field-dependence of the coil inductance [11], was measured with a piezo-electric rotator (blue symbols). Bc2(θ) is plotted in units of the maximum value Bc2,max = 1.42 T. The data agree well with specific heat results from Ref. [26] where Bc2,max ranges from 1.41–1.45 T for different samples and field-sweep conditions. The in-plane condition is satisfied to ±0.2◦ for our sample.

300 O(1) (a) 1.5 (b) 250 O(2) O(1) O(1)

200 || 1 ¥

150 ¤

£ 0.5

¢ ¡ (kHz) 100

B B -

0 f 50 f

0 -0.5

-50 -1 -100 1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8 B (T) B (T)

FIG. S2. Numerically calculated transition frequencies (dashed lines) compared to measured resonance frequencies (symbols) at different fields for (a) central transitions of the three oxygen sites and (b) central and satellite transitions of O(1⊥).

Normal-State Measurement of Hamiltonian Parameters

The remaining parameters are determined by fitting the output of a numerical diagonalization of the exact Hamil- 17 ton to experimentally measured O NMR transitions at three fields greater than Bc2, (B=1.6T, 4.6T, 8T). The quadrupolar parameters for Sr2RuO4 have been investigated on a different crystal in a previous study [29], and were used as a starting point for the fit. A comparison of the best fit calculation to the experimental normal-state line positions are shown in Fig. S2. The fit reproduces the measured resonance frequencies extremely well, with an average error of less than 1kHz across the three fields. The parameters extracted from the best fit are given in Table S1. These values are consistent with previously published results [44]. 10

Shift (%) NQR frequency (MHz) Asymmetry O(1) K1|| = −0.12 νQ = 0.765 η = 0.174 K1⊥ = +0.509

O(2) K2ab = +0.082 νQ = 0.6065 η = 0

TABLE S1. List of best fit Hamiltonian parameters for the different oxygen sites with θ = 0◦ and φ = 3◦. The two planar oxygen sites O(1) and O(1’) are identical without applied field and are labeled O(1). With the field aligned in the Ru-O plane (θ = 0◦), just two components of K are relevant for the O(1) site while only one is relevant for the O(2) site.

1 0° 3° 5°

(a) experim ent 1.6 T

0 O(1 )

||

0° calc. O(2)

-1

| 3°

O(1 ) 5° -2

m ax.

8.0 8.5 9.0 9.5 10.0 10.5

deviation

2 (kHz) (b) 4.6 T 1 calc f

0

- -

-1 exp f

-2 = = f

25.0 25.5 26.0 26.5 27.0 27.5 28.0

1

(c) 8.1 T

0

-1

-2

45.0 45.5 46.0 46.5 47.0 47.5 48.0

f (MHz)

FIG. S3. Difference between calculated and measured resonance frequencies, ∆f = fexp − fcalc(φ) is shown for (a) B = 1.6, (b) 4.6 and (c) 8.1 T. The experimental data (∆f = 0) are shown at the respective frequency for O(1k), O(2), O(1⊥) in blue, green and red colors, respectively. The calculated results at φ = 0◦, 3◦ and 5◦ in-plane angle with respect to B k [100] are indicated by crosses, minus and plus signs in dark grey, black and light grey color, respectively. On the right we illustrate the maximum deviations from the experimentally determined peak positions. The angle dependence becomes most pronounced at low fields, where φ = 3◦ provides the best fit at B = 1.6 T. The accumulated rms error is smallest for φ = 3◦ at all fields.

Discussion of in-plane angle uncertainty

Due to the weak dependence of the quadrupolar term on in-plane angle near 0◦ at high field, it is possible to accurately fit the normal-state spectra for a range of in-plane angles ( 0◦ 5◦). This is illustrated in Fig. S3, which shows the deviation between the predicted and measured frequencies≈ for all− 17O transitions using 0◦, 3◦, 5◦ in-plane angle fits; the predicted normal-state frequencies differ only by 1 kHz between the fits for B =1.6–8 T. While the overall deviations are smallest for φ = 3◦ (which was used for NMR± shift analysis), any systematic error introduced by uncertainty in the in-plane angle should be examined. While for fields B>Bc2 the effect of in-plane angle is small, it can have a strong impact on the expected normal- state position at lower fields. As such, this affects the ability to extract K/Knormal for B 0. To illustrate this, the ◦ ◦ → ◦ resulting K/Knormal are shown for best fits using the three in-plane angles φ =0 , 3 and 5 in Fig. S4. The O(1k) ◦ ◦ site shows a particularly strong dependence on φ: the 0 and 5 fits produce unphysical behavior, with Ks exceeding 11

◦ ◦ the normal-state value for φ =0 and changing sign for φ =5 . The O(1⊥) site has a much weaker dependence, but still shows unphysical behavior for angles deviating from 3◦. This gives further confidence in the visually determined ◦ 3 angle, but also shows that the O(1⊥) site is more robust to a small angle systematic error in evaluating the Knight shifts. Additionally, it should be noted that the apical O(2) site, although having much weaker hyperfine coupling, is completely independent of the in-plane angle due to its axial symmetry, avoiding this issue altogether.

(a) (b)

1.0

B [100] T = 25 mK

0.8

0.8

K

1,|| 0.6 0.4 normal normal

0° K K

/ / 3° / / s s K 5° 0.4 K

0.0 K

1,

0° C/T from Ref. [23]

0.2

0.00 K 3°

(extrapol. to T=0)

5°

B [100] T = 25 mK

-0.4

0.0

0 1 2 0 1 2

B (T ) B (T )

◦ ◦ ◦ ◦ FIG. S4. Difference between extracted Ks/Knormal for in-plane angles φ = 0 , 3 and 5 with respect to B k [100]. (a) φ = 0 ◦ and 5 yield strong deviations for K1k with non-physical behavior Ks < 0 and Ks > Knormal. (b) Due to generally larger Knight shift, the variations of K1⊥ are less pronounced, yielding more robust values. Still, the non-monotonous behavior upon lowering B for 5◦ is not meaningful, and also the susceptibility values smaller (0◦) than the quasiparticle contribution from ◦ specific heat, Ks/Knormal < C/Cnormal [23], are unphysical. Altogether, we conclude upon an in-plane angle φ = 3 from [100], consistently used in this work.