Concomitant appearance of conductivity and superconductivity in (111)LaAlO3/SrTiO3 interface with metal capping
R. S. Bisht, M. Mograbi, P. K. Rout, G. Tuvia, Hye-Ok Yoon, A. G. Swartz, H. Y. Hwang, Y. Dagan
CConcomitant appearance of conductivity and superconductivity in (111)LaAlO /SrTiO interface with metal capping R. S. Bisht, M. Mograbi, P. K. Rout, G. Tuvia, Hye-OkYoon,
2, 3
A. G. Swartz,
2, 3
H. Y. Hwang,
2, 3 and Y. Dagan ∗ Raymond and Beverly Sackler School of Physics and Astronomy, Tel-Aviv University, Tel Aviv, 6997801, Israel Department of Applied Physics, Geballe Laboratory for Advanced Materials,Stanford University, 476 Lomita Mall, Stanford, CA 94305, USA Stanford Institute for Materials and Energy Sciences,SLAC National Accelerator Laboratory, Menlo Park, California 94025, USA (Dated: February 16, 2021)In polar-oxide interfaces, a certain number of monolayers (ML) is needed for conductivity to ap-pear. This threshold for conductivity is explained by the accumulation of sufficient electric potentialto initiate charge transfer to the interface. Here we study the (111) SrTiO /LaAlO interface wherea critical thickness of nine epitaxial LaAlO ML is required to turn the interface from insulating toconducting and even superconducting. We show that this critical thickness decreases to 3ML whendepositing a cobalt over-layer (capping) and 6ML for platinum capping. The latter result contrastswith the (100) interface where platinum capping increases the critical thickness beyond that of thebare interface. These results suggest that the work function of the metallic capping plays an impor-tant role in both interfaces. Interestingly, for (111) SrTiO /LaAlO /Metal interfaces conductivityappears concomitantly with superconductivity in contrast with the SrTiO /LaAlO /Metal interfacewith LaAlO layer smaller than four ML (unit-cells), which are conducting but not superconducting.We suggest that this difference is related to the different sub-bands involved in conductivity for the(111) interfaces, comparing to the (100) interfaces. Our findings can be useful for superconductingdevices made of such interfaces. The interface between LaAlO and SrTiO exhibitstwo-dimensional conductivity [1], superconductivity [2],magnetism [3–8], metal-insulator transition [9], tunableRashba spin-orbit interaction [10, 11], quantum hallstates [12, 13], and one-dimensional conductivity [14, 15].While the (100) SrTiO /LaAlO interface received sig-nificant scientific attention, the (111) interface remainsless explored.In (111) SrTiO /LaAlO interface three distinct tri-angular Ti layers form one monolayer [16, 17]. In bulk,the crystal field results in a splitting of the 3d orbitalsinto e g and three degenerate t g manifolds that are fur-ther split due to the trigonal symmetry at the interfaceinto a g and e (cid:48) g orbitals. The six-fold symmetry of the ti-tanium layer is reflected in the transport properties [18].The symmetry is further reduced due to the structuraltransition of SrTiO and the interface is predicted to hostexotic superconductivity [19] and topological states [20].Four monolayers of LaAlO are needed for the forma-tion of two-dimensional electron system (2DES) at the(100) SrTiO /LaAlO interface [9]. For the (111) inter-face, the critical thickness for conductivity is nine mono-layers (ML) [16]. The (111) 2DES also exhibits super-conductivity [21], with a link between superconductivityand spin-orbit interaction [22]. Importantly, upon carrierdepletion with negative gate voltage superconductivitytransitions into a Bose-insulating state [23]. This behav-ior contrasts with the (100) interface where a weaker in-sulating state is observed for negative gate biases [24, 25]. ∗ Corresponding author: [email protected]
For spin injection and low voltage transistor applica-tions the barrier produced by the minimal four mono-layers of LaAlO required for conductivity at the bare(100) interface or by the nine monolayers at the (111) in-terface is relatively strong. Recent first principle calcula-tions [26] and experimental studies of various metal cap-ping on (100) interfaces (SrTiO /LaAlO /Metal) [27, 28]show that the critical thickness for the onset of conduc-tivity, t LAO , can be reduced relative to the bare inter-face. t LAO increases with the work function. Here westudy the problem of critical thickness for conductivityfor (111) interfaces. We have also expanded our researchto the superconducting properties of both (100) and (111)SrTiO /LaAlO /metal interfaces. We find that uponcapping the LaAlO in (111) SrTiO /LaAlO interfacewith cobalt (Co) and platinum (Pt), t LAO has been sup-pressed from 9 ML to 3 ML and 6 ML, respectively. Fur-thermore, once the (111) interface becomes conducting, italso becomes superconducting at low temperatures. Thiscontrasts with the (100) interface where superconductiv-ity is not observed in SrTiO /LaAlO /Metal interfaceswith reduced t LAO . We suggest that for (111) interface,the bands responsible for conductivity and superconduc-tivity are the same while for (100) conductivity appearsfirst in the d xy type bands, which are not superconduct-ing [17, 29].Epitaxial LaAlO films with different thickness weregrown on Ti and TiO terminated atomically smooth(111) and (100) SrTiO substrates respectively at anoxygen pressure of 1 × − Torr and temperature 780 Cusing pulsed laser deposition. The thickness was in-situmonitored by reflection high energy electron diffraction a r X i v : . [ c ond - m a t . s up r- c on ] F e b (RHEED). The samples were then transferred to a mag-nteron sputtering chamber where they were pre-annealedfor two minutes at 200 C and at a pressure of 1 × − Torr to remove surface contaminants. Metallic layers of ≈ x to pre-vent oxidation. Wire bonding was used to connect tothe sample electrically. In this configuration, the mea-sured resistance is either a parallel combination of themetal cap resistance and the 2DES at the conductingSrTiO /LaAlO interface or only the metallic cap in theabsence of 2DES. We demonstrate this by measuring thetransverse resistance R xy i.e., the Hall signal of the (111)SrTiO /LaAlO /Co/AlO x interface for 4 LaAlO ML,as shown in Fig.1. While for positive gate voltage, the2DES dominates and exhibits signal resembling the (111)SrTiO /LaAlO interface [22] when depleting the 2DESby negative gate voltage, the contribution of Co predomi-nates as manifested in an anomalous-Hall signal confirm-ing the presence of two parallel channels for conduction.This behavior is similar to (100) with cobalt capping [28]. FIG. 1. Transverse resistance of(111)SrTiO /LaAlO /Co/AlO x for 4 ML LaAlO as afunction of a perpendicular magnetic field at different gatevoltages. The inset focuses on the negative gate voltageregime. The observed anomalous-Hall signal demonstratesthe predominance of the Co layer properties in this regime.This is in contrast to the higher carrier density regime wherethe 2DES dominates. The transport studies conducted on (111)SrTiO /LaAlO /Co/AlO x interface show a reduc-tion of LaAlO critical thickness for the onset of 2DESconductivity from 9 ML to 3 ML. In Fig.2 (a) we showthe sheet resistance of (111)SrTiO /LaAlO /Co/AlO x as a function of LaAlO thickness at 40 K. ForLaAlO thickness below 3 ML, the resistance increasesby nearly five times. This indicates that 3 ML is thecritical thickness of LaAlO (t LAO ) for the onset ofconductivity with Co capping. To verify that a 2DES is formed parallel to the metallic layer, we measured theresistance versus back gate voltage. For a thin metalliclayer parallel to a 2DES, one expects gate dependentresistance due to the dominating contribution of 2DES.On the other hand, in the absence of a 2DES parallelto a metallic layer we expect the gate dependence ofthe resistance to be immeasurably small due to thesubstantial carrier density in the metal. Fig.2 (b) showsthe gate dependence of (111) SrTiO /LaAlO /Co/AlO x for 1 and 3 LaAlO ML. For the sample with a singleLaAlO ML, the normalized resistance (R/R ( − V ) )is flat as a function of gate voltage, suggesting theabsence of 2DES at the SrTiO /LaAlO interface.For 3 ML LaAlO , the data show a significant gatedependence, suggesting the formation of 2DES at(111)SrTiO /LaAlO interface. We interpret thesaturation of the resistance at negative gate voltage asa result of depletion of the 2DES and dominance of theCo capping layer. We conclude that t LAO becomes 3ML upon Co capping for (111) interface.
FIG. 2. (a) The sheet resistance ( R (cid:3) ) of (111)SrTiO /LaAlO /Co/AlO x at 40 K as a function ofLaAlO thickness, an abrupt drop in the resistance at 3ML indicates that this is the critical thickness for the for-mation of 2DES at the SrTiO /LaAlO interface. Note: Themeasured samples were not patterned, the geometrical fac-tor for the sheet resistance calculation is within an error barof ± ( − V ) ) for (111) SrTiO /LaAlO /Co/AlO x withLaAlO thickness of 1 and 3 monolayers. (c) The gatedependence of normalized resistance (R/R ( − V ) ) for (111)SrTiO /LaAlO /Pt for LaAlO thickness of 5 and 6 mono-layers. Previous studies on the metal-capped (100)SrTiO /LaAlO interface show that the criticalthickness increases with the work-function of the metal-capping layer [26]. To understand the role of the workfunction in (111) interfaces, we carried out experimentswith Pt capping. The work-function of platinum isgenerally higher than that of Co [30]. Surprisingly, wefound that unlike the (100) interface, Pt capping alsoreduces t LAO from 9 ML to 6 ML. This is demonstratedin Fig.2(c), where we show the gate dependence of(111)SrTiO /LaAlO /Pt for 5 and 6 LaAlO ML. Theabsence of gate dependence for 5ML and the strong gatedependence for 6ML suggests that t
LAO with Pt capis 6 ML. In Figure S1 (Supplemental information), wealso show the gate dependence of the extracted sheetresistance of 2DES for (111)SrTiO /LaAlO /Co/AlO x and (111)SrTiO /LaAlO /Pt for 3 ML and 6 ML ofLaAlO .While suppression of t LAO upon Co capping is ob-served for both the (111) and (100) SrTiO /LaAlO in-terfaces, a reduction of t LAO upon Pt capping is observedonly for (111) interface, whereas an increase in t
LAO isfound for the Pt capped (100) interface [28]. A plausibleexplanation for that is the dependence of the effectivework-function on the surface properties[31].The data presented above unveil that t
LAO has astrong dependence on the work function. But how doesmetal capping affect superconductivity for the (100) and(111) interfaces? To address this question, we cooleddown our samples in a dilution refrigerator. Surprisingly,we find that all the (111) samples, which show conduc-tivity on metal capping, also show superconductivity. Intable 1 and S1 (Supplemental information), we summa-rize the properties of the (100) and (111) interfaces withvarious LaAlO thickness and different metal capping.Fig.3 displays the normalized resistance (R/R (0 . K ) ) as afunction of temperature for different (100) and (111) in-terfaces at the critical thickness t LAO with Co and Ptcapping. As shown previously in Fig.1, the Co layerindeed shows an anomalous hall effect, but for such athin cobalt film, one expects the magnetic coupling tobe extremely short with a negligible effect on the 2DES.Nevertheless, to eliminate the possibility that the closeproximity ( 12˚A) of the ferromagnetic Co to the inter-face is responsible for the absence of superconductiv-ity in the (100) SrTiO /LaAlO /Co/AlO x , we measured(non-magnetic) Ag capped (100) interface with 3 ML ofLaAlO , which also shows no superconductivity. In Fig-ure S2 (Supplemental information), we show the temper-ature and gate bias dependence of the resistance for Agcapped (100) interface.To make sure that the observed superconductivity is atwo-dimensional (2D) interfacial effect and not a spuriousone resulting from the sputtering process, we studied thetemperature dependence of the perpendicular and par-allel critical field. We show in the supplementary infor-mation that they both follow the expected 2D Ginzburg-Landau temperature dependence (See Figure S3 of sup-plemental information).In Fig.4 we show the behavior of the critical temper-ature and critical fields as a function of gate voltage forCo and Pt capped (111) interface. The dome shaped gatedependence and the values of T c , the perpendicular crit-ical field H ⊥ , and the parallel critical field H (cid:107) are similarto the bare (111) SrTiO /LaAlO interface [17, 22].The important findings we report here are the sup- FIG. 3. The normalized resistance (R/R (0 . K ) )of (100) SrTiO /LaAlO (3ML)/Co/AlO x ,(100)SrTiO /LaAlO (9ML)/Pt, (111)SrTiO /LaAlO (3ML)/ Co/AlO x , and (111)SrTiO /LaAlO (6ML)/Pt. The LaAlO /SrTiO (111)interfaces capped with Co and Pt show a superconductingtransition and the corresponding critical thickness of LAO is3 and 6 Monolayer respectively. pression of t LAO for (111) interfaces both with Co andwith Pt capping. This is in contrast with the (100) in-terface where t
LAO with Pt capping has been shown toincrease relative to the bare interface[28]. Furthermore,once conductivity appears by either Pt or Co capping on(111) SrTiO /LaAlO interfaces (with LaAlO thick-ness larger than or equals to t LAO ) superconductivity isobserved at low temperatures. This is not the case for(100) SrTiO /LaAlO /metal (Fig.3).It has been shown theoretically that the work func-tion of the capping metal as well as the size and direc-tion of the internal electric dipole in the LaAlO layeraffect the degree of charge transfer from the metal tothe Ti 3d bands at the interface [26] and consequentlyinfluence t LAO . These predictions have been verifiedexperimentally [28]. A naive inspection of the inter-nal dipole field at the (111) SrTiO /LaAlO inter-face shows that it is in-fact opposite to that of (100)counterpart [16]. It is known that the different surfaceadsorbates can modify the electronic properties of theSrTiO /LaAlO interface[32]. In addition, For metals,the effective work function strongly depends on the num-ber of surface dipoles. The latter depends on the sur-face reconstructions, as well as on the atmospheric ad-sorbates. It is, therefore, possible that the deposited Ptcan have some additional surface effects on the (111) ori-entation compared to the (100) orientation. This modi-fication of the electrostatic boundary conditions may beat the origin of the suppression of t LAO with Pt cappingfor the (111) interface.The other important observation is the presence of su-perconductivity in the metal-capped (111) interface in
TABLE I. Summary of the different samples measured for various thickness of LaAlO upon different metal capping andcorresponding nature of the interface. Note: For (100) interface one monolayer corresponds to one unit-cell (see also table S1(Supplemental information) for more samples).Interface Metallayer Thickness(ML) Conducting Super-Conducting(111) Co 2 No No(111) Co 3 Yes Yes(100) Co 2 Yes No(100) Co 3 Yes No(111) Pt 5 No No(111) Pt 6 Yes Yes(100) Pt 9 Yes No(100) Ag 3 Yes NoFIG. 4. (a) The critical temperature (T c ) of (111)SrTiO /LaAlO (6ML)/Pt interface, the dark yellow circleshows the T c of the as cooled film. (b), (c) and (d)are superconducting critical temperature, perpendicular crit-ical field and parallel critical field respectively of the (111)SrTiO /LaAlO (3ML)/Co/AlO x interface as a function ofgate voltage. The values of these critical parameters are ingood agreement with that of the uncapped interface. Note:T c is defined as a temperature where the value of resistance dropsby 50% of it’s value at 0.5 K. contrast to the (100) interface as demonstrated in Table1 and in table S1 in the supplementary part.How can we understand the robustness of superconduc-tivity in (111) interfaces? Previous experimental studieson (111) SrTiO /LaAlO interfaces show that even atstrong negative gate voltages superconductivity remainsintact [22, 23]. On the theory side, the curvature ofthe Fermi contour changes quickly upon charge accu-mulation, and both the conducting and superconducting bands get an equal contribution from the three degen-erate t g orbitals[17]. By contrast, for (100) interface,there are distinct less mobile, non-superconducting bandand a mobile superconducting band due to the very dif-ferent effective masses of the light d xy and heavy d yz , d xz bands. It is possible that the inter-band repulsion[29] results in shifting the second band to higher energyleaving only the metallic state. On the (111) interface,the splitting of t g orbitals occurs due to the crystal field.The crystal field splitting may be sensitive to the detailsof the interface. For example, the lowermost band is a g in (111) SrTiO /LaAlO , but it is the e (cid:48) g in doped sur-face of (111) SrTiO crystal[33]. It is possible that bymetallic capping, the sign of the crystal field changes andthe superconducting e (cid:48) g bands become lower in energy. Inthis case, one or two mobile bands will be occupied at allgate voltages and one should see both conductivity andsuperconductivity.To summarize, capping of Co and Pt on(111)SrTiO /LaAlO interface reduces the criticalthickness for the onset of conductivity. Importantly, allconducting (111) interfaces are also superconductingat low temperatures. Our findings suggest that properchoice of metal on top of the LaAlO barrier can tunethe barrier strength for various applications such assuperconducting tunneling devices and ferromagnetictunnel junction.RSB and MM contributed equally to this work. ACKNOWLEDGMENTS
We thank Moshe Goldstein and Udit Khanna for use-ful discussions. This research was supported by the Bi-national Science Foundation under grant 2014047 and bythe Israel Science Foundation under grant : 382/17. [1] A. Ohtomo and H. Y. Hwang, Nature , 423 (2004).[2] N. Reyren, S. Thiel, A. D. Caviglia, L. F. Kourkoutis,G. Hammerl, C. Richter, C. W. Schneider, T. Kopp, A.- S. R¨uetschi, D. Jaccard, et al., Science , 1196 (2007).[3] A. Brinkman, M. Huijben, M. van Zalk, J. Huijben,U. Zeitler, J. C. Maan, W. G. van der Wiel, G. Rijnders,
D. H. A. Blank, and H. Hilgenkamp, Nature Materials ,493 (2007).[4] M. Sachs, D. Rakhmilevitch, M. Ben Shalom, S. Shefler,A. Palevski, and Y. Dagan, Physica C: Superconductivityand its applications , S746 (2010).[5] J. A. Bert, B. Kalisky, C. Bell, M. Kim, Y. Hikita, H. Y.Hwang, and K. A. Moler, Nature Physics , 767 (2011).[6] A. Ron, E. Maniv, D. Graf, J.-H. Park, and Y. Dagan,Phys. Rev. Lett. , 216801 (2014).[7] S. Seri and L. Klein, Phys. Rev. B , 180410 (2009).[8] J.-S. Lee, Y. Xie, H. Sato, C. Bell, Y. Hikita, H. Hwang,and C.-C. Kao, Nature Materials , 703 (2013).[9] S. Thiel, G. Hammerl, A. Schmehl, C. W. Schneider, andJ. Mannhart, Science , 1942 (2006).[10] M. Ben Shalom, M. Sachs, D. Rakhmilevitch,A. Palevski, and Y. Dagan, Phys. Rev. Lett. , 126802(2010).[11] A. Caviglia, M. Gabay, S. Gariglio, N. Reyren, C. Can-cellieri, and J.-M. Triscone, Phys. Rev. lett. , 126803(2010).[12] Y. Xie, C. Bell, M. Kim, H. Inoue, Y. Hikita, and H. Y.Hwang, Solid state communications , 25 (2014).[13] F. Trier, G. E. Prawiroatmodjo, Z. Zhong, D. V. Chris-tensen, M. von Soosten, A. Bhowmik, J. M. G. Lastra,Y. Chen, T. S. Jespersen, and N. Pryds, Phys. Rev. Lett. , 096804 (2016).[14] A. Ron and Y. Dagan, Phys. Rev. lett. , 136801(2014).[15] M. Briggeman, M. Tomczyk, B. Tian, H. Lee, J.-W. Lee,Y. He, A. Tylan-Tyler, M. Huang, C.-B. Eom, D. Pekker,et al., Science , 769 (2020).[16] G. Herranz, F. S´anchez, N. Dix, M. Scigaj, and J. Fontcu-berta, Sci. Rep. , 758 (2012).[17] U. Khanna, P. K. Rout, M. Mograbi, G. Tuvia, I. Leer-makers, U. Zeitler, Y. Dagan, and M. Goldstein, Phys.Rev. Lett. , 036805 (2019).[18] P. K. Rout, I. Agireen, E. Maniv, M. Goldstein, andY. Dagan, Phys. Rev. B , 241107 (2017).[19] M. Hecker and J. Schmalian, npj Quantum Materials ,1 (2018).[20] D. Doennig, W. E. Pickett, and R. Pentcheva, Phys. Rev.Lett. , 126804 (2013).[21] A. M. R. V. L. Monteiro, D. J. Groenendijk, I. Groen,J. de Bruijckere, R. Gaudenzi, H. S. J. van der Zant, andA. D. Caviglia, Phys. Rev. B , 020504 (2017).[22] P. K. Rout, E. Maniv, and Y. Dagan, Phys. Rev. Lett. , 237002 (2017).[23] M. Mograbi, E. Maniv, P. K. Rout, D. Graf, J.-H. Park,and Y. Dagan, Phys. Rev. B , 094507 (2019).[24] A. D. Caviglia, S. Gariglio, N. Reyren, D. Jac-card, T. Schneider, M. Gabay, S. Thiel, G. Hammerl,J. Mannhart, and J.-M. Triscone, Nature , 624(2008).[25] Z. Chen, A. G. Swartz, H. Yoon, H. Inoue, T. A. Merz,D. Lu, Y. Xie, H. Yuan, Y. Hikita, S. Raghu, et al.,Nature communications , 1 (2018).[26] R. Arras, V. G. Ruiz, W. E. Pickett, and R. Pentcheva,Phys. Rev. B , 125404 (2012).[27] E. Lesne, N. Reyren, D. Doennig, R. Mattana, H. Jaffr`es,V. Cros, F. Petroff, F. Choueikani, P. Ohresser,R. Pentcheva, et al., Nature Communications (2014).[28] D. C. Vaz, E. Lesne, A. Sander, H. Naganuma,E. Jacquet, J. Santamaria, A. Barth´el´emy, and M. Bibes,Advanced Materials , 1700486 (2017). [29] E. Maniv, M. B. Shalom, A. Ron, M. Mograbi,A. Palevski, M. Goldstein, and Y. Dagan, Nature Com-munications , 8239 (2015).[30] H. B. Michaelson, J. Appl. Phys. , 4729 (1977).[31] M. Mrovec, J.-M. Albina, B. Meyer, and C. Els¨asser,Phys. Rev. B , 245121 (2009).[32] Y. Xie, Y. Hikita, C. Bell, and H. Y. Hwang, NatureCommunications , 494 (2011).[33] G. M. De Luca, R. Di Capua, E. Di Gennaro, A. Sambri,F. M. Granozio, G. Ghiringhelli, D. Betto, C. Piamon-teze, N. B. Brookes, and M. Salluzzo, Phys. Rev. B ,115143 (2018). Supplemental Material for Concomitant ap-pearance of conductivity and superconductivityin (111) LaAlO /SrTiO interface on metalcappingS1: Sheet Resistance of two dimensional elec-tron system without Capping The sheet resistance of the two dimensional elec-tron system (2DES) without capping was extractedfrom the analysis of the parallel resistance modelas given in the equation 1. Where R Measured , R Metal − Overlayer ,and R DES is the experimentally mea-sured resistance of the parallel conducting channel,resistance of metal overlayer, and resistance of the2DES respectively. In our analysis we assumed R metal − capping = R Measured at strong negative gate volt-ages. Figure S1 represents the extracted sheet resis-tance of the 2DES for (111)SrTiO /LaAlO /Co/AlO x and (111)SrTiO /LaAlO /Pt samples for 3 and 6 MLLaAlO as a function of gate voltages. The divergingnature of the resistance at negative gate voltage is ex-pected for a depleted 2DES.1 R Measured = 1 R Metal − Overlayer + 1 R DES (1)
S2: Role of Ag Capping
To show that absence of superconductivity in co-capped (100) SrTiO /LaAlO interface is not relatedto ferromagnetism in the cobalt, we used silver (Ag) cap-ping. Ag has low work function and can therefore reducethe critical thickness by increasing the charge transfer.The measured Ag capped (100) interface for 3 monolay-ers (ML) of LaAlO does not show full superconductingtransition for any gate voltage. This confirms that fer-romagnetism of Co is not the cause for the absence ofsuperconductivty in (100) co-capped interfaces. In Fig-ure S2 (a) we show the sheet resistance of Ag capped(100)SrTiO /LaAlO as a function of temperature atdifferent positive gate voltages. A dip in the sheet resis-tance can be attributed due to the localized supercon-ducting regions which are not connected percolativelyeven on application of positive gate bias. These local-ized regions may arise from spurious damage from thesputtering or slight thickness nonuniformity. A simi-lar dip has been observed in the Platinum (Pt) capped(100)SrTiO /LaAlO interface for 9 ML of LaAlO .Neverthless it has been disappeared at strong positivegate bias as shown in Figure S2 (b). The data thus sug-gest there is no macroscopic superconductivity presentin (100)SrTiO /LaAlO interface below the bare criti- cal thickness upon metal capping. S3: Analysis of critical fields
The temperature dependence of the perpendicularand parallel critical fields were analyzed according tothe phenomenological Ginzburg-landau theory for the 2-
FIG. 5. S1 : The extracted sheet resis-tance of (111)SrTiO /LaAlO /Co/AlO x and(111)SrTiO /LaAlO /Pt samples for 3 and 6 ML ofLaAlO as a function of a gate voltages. The resistance wasextracted according to the equation 1.The diverging natureof the the resistance at negative gate represents the depletionof the 2DES. dimensional nature of the superconductivity. The per-pendicular and parallel critical fields in the frameworkof the Ginzburg-landau theory follow equation 2 and 3respectively. In Figure S3, we show the fit to equation1 and 2 for (111)SrTiO / LaAlO (3ML)/Co interface.The extracted perpendicular ( H ⊥ ) and parallel criticalfield ( H (cid:107) ) as a fuction of gate voltages has been shownin Figure 4 of the main text. H ⊥ = φ πξ (1 − TT c ) (2) H (cid:107) = φ √ πξ d (1 − TT c ) (3) Table S1: