LETTERSPUBLISHED ONLINE: 24 NOVEMBER 2014 | DOI: 10.1038/NMAT4153

Superconductivity above 100K in single-layerFeSe films on doped SrTiO3

Jian-Feng Ge1, Zhi-Long Liu1, Canhua Liu1,2*, Chun-Lei Gao1,2, Dong Qian1,2, Qi-Kun Xue3*, Ying Liu1,2,4and Jin-Feng Jia1,2*Recent experiments on FeSe films grown on SrTiO3 (STO)suggest that interface e�ects can be used as a means toreach superconducting critical temperatures (Tc) of up to80K (ref. 1). This is nearly ten times the Tc of bulk FeSeand higher than the record value of 56K for known bulkFe-based superconductors2. Together with recent studies ofsuperconductivity at oxide heterostructure interfaces3–6, theseresults rekindle the long-standing idea that electron pairingat interfaces between two di�erent materials can be tailoredtoachievehigh-temperaturesuperconductivity7–12. Subsequentangle-resolved photoemission spectroscopy measurementsof the FeSe/STO system revealed an electronic structuredistinct from bulk FeSe (refs 13,14), with an energy gapvanishing at around 65K. However, ex situ electrical transportmeasurements1,15 have so far detected zero resistance—thekey experimental signature of superconductivity—only below30K. Here, we report the observation of superconductivitywith Tc above 100K in the FeSe/STO system by means ofin situ four-point probe electrical transport measurements.This finding confirms FeSe/STO as an ideal material forstudying high-Tc superconductivity.

The search for superconductors with a Tc above the liquidnitrogen temperature (77K) led to the discovery of high-Tccuprates with a Tc as high as 130K more than two decades ago16.Even though the value of Tc is only 26K in the first Fe-basedsuperconductor, LaFeAsO (ref. 17), subsequent work on a seriesof Fe-based superconductors showed that the highest Tc underambient pressure was as high as 56K—found in SmFeAsO (ref. 2).So far, superconductors with a Tc> 77 K have been limited to thecuprates. Recently, interface effects have been employed to enhancethe superconductivity in FeSe. Single-layer films of FeSe grown onSrTiO3(001) substrates, referred to below as FeSe/STO, were foundto exhibit a superconducting energy gap, ∆, as large as 20.1meV,detected by in situ scanning tunnelling microscopy/spectroscopy(STM/STS) measurements at 4.2 K (ref. 1). A Tc value as high as86K would be expected if the ratio of 2∆/kBTc=5.5 found in bulkFeSe (Tc=9.4 K (ref. 18)) were applicable for the FeSe/STO system.This work, together with the earlier work on superconductingoxide interfaces3–6, demonstrates that the interface between twodifferent materials provides not only a rich system for studying two-dimensional (2D) superconductivity, but also a potential pathway tohigh-Tc superconductivity7–12.

Indeed, recent angle-resolved photoemission spectroscopyexperiments on the FeSe/STO system revealed a different electronicstructure from those of bulk FeSe, and a possible occurrence of

superconductivity around 65K (refs 13,14). An ex situ transportmeasurement performed on FeSe/STO protected by multiple layersof FeTe and an amorphous Si overlay revealed a zero-resistance Tcof 23.5 K and an onset Tc> 40K (ref. 15). Evidently the additionof protection layers suppresses superconductivity in single-layerFeSe. In this work, we report electrical transport measurements onsingle-layer films of FeSe grown on Nb-doped SrTiO3 substratesusing an in situ four-point probe (4PP) technique. We found thatsuperconductivity could be obtained even at a temperature as highas 109K.

Single-layer films of FeSe were grown on Nb-doped SrTiO3(001)surfaces by the same method as reported previously1, employing amolecular beam epitaxy (MBE) system equippedwith STM/STS and4PP capabilities. The growth process was monitored by reflectionhigh-energy electron diffraction (RHEED; Fig. 1a), which allowsthe precise control of film growth needed to achieve one unit-cell thickness. The crystal nature of the films was confirmed bySTM imaging at both large and atomic scales, as shown in Fig. 1band c, respectively.

The 4PP technique was first used in the study of superconductingsingle-layer Pb films grown on Si(111) surfaces by Hasegawa’sgroup12,19. There are two basic measurement configurations thathave been proposed since the 1950s (refs 20,21 and SupplementaryInformation): C1423 and C1234 (Fig. 2a). In C1423 (C1234), ad.c. current, I14 (I12), is applied through Tips 1 and 4 (1 and 2)while the voltage drop, V23 (V34), is measured between Tips 2 and3 (3 and 4). The applicability of the 4PP technique to extremelythin films is well documented12. Further limits exist for detectingfilms grown on an insulating substrate, as the feedback requiredto control the contact between the film and the tip is extremelydifficult. Nevertheless, the 4PP technique is a powerful tool forinvestigating superconductivity in films that cannot be taken out ofan ultrahigh vacuum system, or for interfaces that are not accessibleby surface probes.

Before 4PP measurements, the contacts between each tip andsamples were examined, as shown in Supplementary Fig. 1, andall contacts were Ohmic. To test the capability of detectingsuperconductivity using our 4PP, we performed measurements onan optimally doped Bi2Sr2CaCu2O8+δ single crystal (Tc∼91K); themeasured I–V curves are shown in Supplementary Fig. 3. Fromour measurements, Tc was found to be ∼90K, which indicates themeasured sample temperature deviates little from the real transitiontemperature (Supplementary Information). It should be noted thatsimilar critical current (Ic) values were obtained for two differentconfigurations (Supplementary Fig. 3).

1Key Laboratory of Artificial Structures and Quantum Control (Ministry of Education), Department of Physics and Astronomy, Shanghai Jiao TongUniversity, 800 Dongchuan Road, Shanghai 200240, China. 2Collaborative Innovation Center of Advanced Microstructures, Nanjing 210093, China.3Department of Physics, Tsinghua University, Beijing 100084, China. 4Department of Physics and Materials Research Institute, Pennsylvania StateUniversity, University Park, Pennsylvania 16802, USA. *e-mail: canhualiu@sjtu.edu.cn; qkxue@mail.tsinghua.edu.cn; jfjia@sjtu.edu.cn

NATUREMATERIALS | ADVANCE ONLINE PUBLICATION | www.nature.com/naturematerials 1© 2014 Macmillan Publishers Limited. All rights reserved.

LETTERS NATUREMATERIALS DOI: 10.1038/NMAT4153

500 nm12 nm

a

c

b

Figure 1 | Growth and characterization of a high-quality single-layer film ofFeSe. a, RHEED pattern taken just after growth of the film. b,c, Large-scaleand atomic-scale STM images of the film, taken at 7 K with Vs=3.00 V andIt= 100 pA and at 3 K with Vs=0.05 V and It=250 pA, respectively.

Two typical I–V curves collected on FeSe/STO at 3K inC1423 and C1234 are shown in Fig. 2b and c, respectively. Thedata demonstrate explicitly that the film is superconducting, withthe critical current defined by the current value for which thesuperconducting top layer can no longer short the conductingsubstrate with a finite resistance. Even though interpretation of thefinite voltage in the superconducting I–V curves is complicated, theessentially zero voltage seen at low currents cannot be a result ofartificial effects of the contacts, as all four tips of the 4PPhaveOhmiccontactswith the sample individually (Supplementary Information).The Ic values so extracted are similar for both measurementconfigurations, whichwas also found on the Bi2Sr2CaCu2O8+δ singlecrystal. The critical current density can be very roughly estimatedin the following way. The contact size is around 10 µm. The depthof the current should be the film thickness plus a certain interfaceregion. For convenience we use a depth of 1 nm. Therefore acritical current of 4.1mA at 3K, for example, will lead to a criticalcurrent density of around 1.3× 107 Acm−2 at the contact point,where the critical current density should be the highest. This valueis one order of magnitude higher than that of FeSe single-layerfilms grown on an insulating STO substrate with a protectingFeTe capping layer15, and of the same order as that of layeredYBa2Cu3O7−x films22.

Linear I–V curves were obtained occasionally even at very lowtemperatures (see Supplementary Figs 1 and 2), possibly owingto inhomogeneity of the sample or to hard contacts between atip (or tips) and the sample, which results in local damage to thesingle FeSe layer. We have confirmed the latter case by intentionallypressing the 4PP farther into the surface until the I–V curveschange from a superconducting to anOhmic behaviour, as shown inSupplementary Fig. 2. In the present experiment, we moved the 4PPto various locations to collect I–V curves at a fixed temperature. Thefilm was designated superconducting when the superconductingI–V curve was obtained in at least one location. It is found that ifone superconducting I–V curve could be obtained in one location,it could be obtained in some other locations, as expected.

The results of the electrical transport measurements on theFeSe/STO system are summarized in Fig. 3. As shown in Fig. 3a,at temperatures above 109K, the I–V curves are linear inthe full measured current range, in strong contrast to those ofthe superconducting curves taken below 109K. Therefore, Tc of thesingle-layer FeSe must be ∼109K. To make a connection withthe more traditional resistance versus temperature plot, we also plotthe resistance values extracted from the linear fit to I–V curves

VV

C1234C1423a

b c

∼20°100 μm or 10 μm

Side view of 4PP

SrTiO3 substrate

FeSe film

Ic = 4.1 mA

V23

(m

V)

V34

(m

V)

I14 (mA)−8 −4

Ic = 4.0 mA

1 2 3 4

1 2 3 4

−0.10

−0.05

0.00

0.05

0.10

0 4 8I12 (mA)

−8 −4 0 4 8−0.10

−0.05

0.00

0.05

0.10

Figure 2 | 4PP transport measurement set-up. a, Schematic of the 4PPtransport measurement set-up. The numbers are used to denote thecontacting tips. All four tips can touch the sample surface gently at aninclined angle of∼20◦. The inset shows schematically the measurementconfiguration C1234. b,c, Typical superconducting I–V curves taken with atip separation of 10 µm at 3.0 K with measurement configurations C1423and C1234, respectively.

taken at fixed temperatures (Fig. 3b). The R–T plot shown in Fig. 3bwas completed by collecting I–V curves at different locations fordifferent temperatures to avoid damage to the FeSe film from anypossible tip–sample shift while the temperature was varied. Weperformed the 4PP measurement on another sample with the tipsfixed at one location while temperature was varied, and obtainedR–T data showing a Tc value of around 99K (SupplementaryFig. 4). It should be emphasized that the value of the conductionin FeSe/STO above 109K was dominated by that of the STOsubstrate rather than the FeSe film, as the conducting substratehas a much larger cross-sectional area and, therefore, a smallresistance. Essentially, below Tc the sample was shorted by thesuperconducting film of FeSe, whereas above Tc the conductionwas dominated by the conducting substrate. In fact, this is thereason for the apparent sharp transition around 109K in Fig. 3b.In short, the 4PP does reveal a superconducting phase transitionin the single-layer FeSe film; the normal-state resistance cannotbe determined using this technique if the film is grown on aconducting substrate.

A cubic-to-tetragonal structural phase transition has beenobserved previously in a vacuum annealed surface of STO at105K (ref. 23), raising the question of the effect of this transitionof the substrate on our measurements. Given that our extensive4PP control measurements on bare STO (insets of Fig. 3a,b)yielded no superconducting I–V curves, the possibility that thesuperconducting I–V curves observed on single-layer FeSe grownon STO actually came from the substrate can be ruled out. Thenormal-state resistance extracted from the I–V curves shown inFig. 2b,c are roughly five times greater than that in Fig. 3a becausethe two sets of data were collected using two 4PPs with differentprobe separations (100 µm and 10 µm, respectively).

It should be emphasized that, below 109K, we could alwaysobtain superconducting I–V curves with a zero-resistance state,while Ic increases gradually from 0.2mA at 109K to 4.0mA at 3K,as shown in Fig. 3c. The temperature dependence of Ic can be fitted

2 NATUREMATERIALS | ADVANCE ONLINE PUBLICATION | www.nature.com/naturematerials© 2014 Macmillan Publishers Limited. All rights reserved.

NATUREMATERIALS DOI: 10.1038/NMAT4153 LETTERS

R (m

Ω)

R (m

Ω)

Cooling by LHe Cooling by LN

C1423C1234

108104100T (K)

−0.8 −0.4 0.0 0.4 0.8

138.0 K 108.8 KBare Nb:SrTiO3

Cubic–tetragonalphase transition

131.6 K 94.0 K113.9 K 92.9 K109.8 K

−0.8−0.4

0.00.40.8

−0.8 −0.4

−10 −5 0 5 −8 −4 0 4 810B (T)

V 34

(μV

)

I14 (mA)

V 23 (

mV

)

V 34

(μV

)

−2

−1

0

1

2C1234

a

c d e

b

I12 (mA)

I c (m

A)

I c (m

A)

T (K)

T (K)

I12 (mA)0.0 0.4 0.8

Bare Nb:SrTiO3

4

3

2

1

0

−1140120100806040200

1.4

1.2

1.0

0.8

12080400

5

Ic(T) = Ic(0)(1 − T/Tc)3/2

Tc = 111 K

4

3

2

1

0

4

3

2 −0.1

0.0

0.1

Figure 3 | Temperature and external magnetic-field dependence of the electrical transport properties of the single-layer film of FeSe. a, Typical I–Vcurves taken at various temperatures around 109 K. The inset shows a typical I–V curve for a bare Nb-doped STO surface taken at 99.2 K. b, Temperaturedependence of the resistance obtained from a linear fit to the I–V curves. Above 79 K the sample was cooled by liquid N2. The inset shows the temperaturedependence of resistance taken on a bare STO surface. c, Temperature dependence of Ic extracted from the measured superconducting I–V curves. Thesuperimposed curve is the fitted result following the Ginzburg–Landau model. d, Out-of-plane magnetic-field dependence of Ic at 3 K. Error bars in c,d aredefined by the di�erence between positive and negative critical current values obtained from the corresponding I–V curves. e, I–V curve taken in a magneticfield of 11 T at 3 K. Data in a–c and d,e are taken with a 4PP with tip separations of 100 µm and 10 µm, respectively.

100989694

Data from aData from b

100 K99 K98 K97 K96 K95 K

T (K) T (K)

T (K)

0H (T)

10 T

C1423 C1423

8 T 6 T 4 T 2 T 0 T

100500T (K)

120

80

40

0100806040200

R (m

Ω)

R (m

Ω)

R (m

Ω)

2.0

1.5

1.0

0.5

0.0

2.01.51.00.50.0

1086420μ

0H

(T)

μ

0H

(T)

μ0H = 10 Tμ

1101051009590

2.0a b c

1.5

1.0

0.5

0.0

10

8

6

4

2

0

H/Hc = 1 − (T/Tc)2

Figure 4 | Influence of an external magnetic field on the zero resistance detected in 4PP transport measurement on another FeSe/STO sample.a, Temperature dependence of the resistance in fixed magnetic fields ranging from 0 to 10 T. The inset shows the R–T curve for the whole range ofexperimental temperatures in a fixed magnetic field of 10 T. b, Magnetic-field dependence of the resistance at fixed temperatures ranging from 95 K to100 K. All the I–V curves from which the sample resistance values were deduced were collected in the configuration C1423 using a 4PP with a 100-µm tipseparation. c, Relation between the critical values of magnetic field and temperature. Filled triangles and circles are experimental results deduced froma and b, respectively. Error bars are defined by the di�erence of temperatures where the measured resistance reaches zero. The line is a fit to an empiricalequation (see text). The inset shows the Hc–T curve plotted using this equation.

to the Ginzburg–Landaumodel, Ic(T )= Ic(0)(1−T/Tc)3/2 (ref. 24),

which gives Ic(0)=3.8±0.2mA and Tc= (111±4)K. This agreeswell with the Tc value obtained directly from the I–V curves at fixedtemperatures. Furthermore, we investigated the superconductingI–V curves under an external magnetic field. It was found that Icdecreased with increasing magnetic field, and remains non-zero upto 11 T (Fig. 3d). As shown in Fig. 3e, a superconducting I–V curvecan still be observed at 11 T at 3K, demonstrating the robustness ofsuperconductivity in the FeSe/STO film.

So far, in situ magnetic measurements on the FeSe/STO systemare unavailable. A diamagnetic signal around 85K was observedrecently in an FeSe/STO system in an ex situ study25. But theFeSe/STO system was capped with a thin-film protective layer ofFeTe, followed by amorphous Si, which is expected to affect the filmproperties drastically. We instead measured, in situ, the influence ofexternalmagnetic fields on the transport properties of our FeSe/STOsystem around Tc. In this study, the FeSe/STO sample was firstcooled with liquid helium and then heated gradually to about 110K

NATUREMATERIALS | ADVANCE ONLINE PUBLICATION | www.nature.com/naturematerials 3© 2014 Macmillan Publishers Limited. All rights reserved.

LETTERS NATUREMATERIALS DOI: 10.1038/NMAT4153

for 4PP transport measurements in fixed magnetic fields up to10 T. A sharp superconducting transition was always observed, asshown in Fig. 4a, but Tc decreases with increasing magnetic field asexpected. 4PPmeasurements were carried out at fixed temperatureswhile the magnetic field was varied, as shown in Fig. 4b. At 100K,zero resistance never occurs, even at zero field, consistent withthe R–T results shown in Fig. 4a. The critical field, Hc, increasesremarkably with decreasing temperature. At 95K, Hc was found tobe greater than 10 T.

The results presented above show that the observed zero-resistance state detected by in situ 4PP measurement, whichwas destroyed by a sufficiently strong magnetic field, is asuperconducting state. In Fig. 4c, we show values of the criticalfield obtained at different temperatures (from Fig. 4a,b), and a fit toan empirical equation23, H/Hc=1− (T/Tc)2. This fit gives a zero-temperature µ0Hc= (116±12)T and zero-field Tc= (99.3±0.2)K.It is interesting to note that previous ex situ transport measurementsin a pulsedmagnetic field yielded a critical field above 52 T (ref. 15).The value of Tc obtained in the fitting result is slightly smaller thanthat found in Fig. 3—possibly owing to sample quality.

In summary, our in situ 4PP electrical transport measurementsshowed that single-layer films of FeSe grown on a conductiveSTO substrate are superconducting at temperatures as high as109K. Although the mechanism of high-Tc superconductivity insingle-layer FeSe/STO has yet to be understood, interface-enhancedelectron–phonon couplingmay have played a role in this remarkablephenomenon1,7–9,26,27, giving a strong boost to the search for high-Tcsuperconductivity in artificial interface systems.

MethodsThis experiment was carried out in an STM–MBE system (made by Unisoku)consisting of two connected ultrahigh vacuum (UHV) chambers with a basepressure of 3×10−10 torr. A self-designed 4PP can be mounted onto the STM tipstage for in situ electrical transport measurements, while the original STMworking function is retained. Single-layer FeSe films were grown on theTiO2-terminated and Nb-doped SrTiO3(001) (STO) substrate, following themethod reported previously1. A single-crystal STO chip was ultrasonicallycleaned with ethanol and acetone first. After degassing over night at 600 ◦C inUHV, the STO wafer was kept at 950 ◦C under a Se atmosphere for 30min andthen gradually cooled down to 550 ◦C for FeSe growth. RHEED was used tomonitor the growth process, during which Se and Fe were co-deposited onto theSTO surface with a flux ratio of ∼20:1. Post-annealing at the growth temperaturewas carried out for 30min after the Se and Fe fluxes were stopped. The film wasimmediately transferred from the MBE growth chamber to the STM chamber forSTM and 4PP transport measurements. A chemically etched tungsten tip wasused for the STM experiments.

The 4PP is composed of four tips, which are in turn linearly aligned with aroughly equal separation. Two 4PPs were used in the experiments. One is madeof silicon cantilevers coated with Au with a tip separation of ∼10µm(Supplementary Fig. 5a), whereas the other is made of Cu wire with a tipseparation of ∼100µm (Supplementary Fig. 5b). These four tips are fixed on aceramic plate mounted on a 4PP holder that we designed to accommodate fourelectrodes. We redesigned the tip stage of the STM system so that the 4PP tipholder can also be inserted on the STM tip stage in situ without affecting theSTM function when the STM tip holder is put back on the stage. The approach ofthe four tips towards the sample surface was precisely controlled by using theSTM positioning system. The four tips are inclined to the sample surface at anangle of ∼20◦ so that they touch the sample surface gently even though they arenot of exactly the same length. In the transport measurement, I–V curves werecollected at a fast rate (about 20 s per curve) while the temperature of the systemwas raised slowly (around 3Kh−1). In the 4PP electrical measurement, theelectric current was provided by a Keithley 2400 SourceMeter while the voltagewas measured using a Keithley 2182A Nanovoltmeter. The electronics are capableof detecting sample resistances of the order of 10−6�. To eliminate thermalvoltages in the circuit, the d.c. measurement current was reversed in the I–Vcurve measurements.

The thermometer used in the experiment is a Cernox resistor type providedby Lake Shore Cryotronics (Model: CX-1030-SD-HT-0.3M). During thisexperiment, the thermometer reads 295.7 K at room temperature and 77.6 Kwhen immersed in the liquid nitrogen. Given that the thermometer is thermallyanchored at the sample stage, which is thermally linked to a liquid 4He or N2

bath, and the film sits on top of the substrate (Supplementary Fig. 6), the sample

temperature cannot be lower than the cold stage temperature measured by thethermometer (Supplementary Information).

Received 11 June 2014; accepted 30 October 2014;published online 24 November 2014

References1. Wang, Q. Y. et al. Interface-induced high-temperature superconductivity in

single unit-cell FeSe films on SrTiO3. Chin. Phys. Lett. 29, 037402 (2012).2. Wu, G. et al. Superconductivity at 56K in samarium-doped SrFeAsF. J. Phys.

Condens. Matter 21, 142203 (2009).3. Reyren, N. et al. Superconducting interfaces between insulating oxides. Science

317, 1196–1199 (2007).4. Gozar, A. et al.High-temperature interface superconductivity between metallic

and insulating copper oxides. Nature 455, 782–785 (2008).5. Kozuka, Y. et al. Two-dimensional normal-state quantum oscillations in a

superconducting heterostructure. Nature 462, 487–490 (2009).6. Richter, C. et al. Interface superconductor with gap behaviour like a

high-temperature superconductor. Nature 502, 528–531 (2013).7. Ginzburg, V. L. On surface superconductivity. Phys. Lett. 13, 101–102 (1964).8. Cohen, M. H. & Douglass, D. H. Superconductive pairing across electron

barriers. Phys. Rev. Lett. 19, 118–121 (1967).9. Strongin, M. et al. Enhanced superconductivity in layered metallic films. Phys.

Rev. Lett. 21, 1320–1323 (1968).10. Zhang, T. et al. Superconductivity in one-atomic-layer metal films grown on

Si(111). Nature Phys. 6, 104–108 (2010).11. Uchihashi, T., Mishra, P., Aono, M. & Nakayama, T. Macroscopic

superconducting current through a silicon surface reconstruction with indiumadatoms: Si(111)-(

√7×√3)-In. Phys. Rev. Lett. 107, 207001 (2011).

12. Yamada, M., Hirahara, T. & Hasegawa, S. Magnetoresistance measurements ofa superconducting surface state of In-induced and Pb-induced structures onSi(111). Phys. Rev. Lett. 110, 237001 (2013).

13. He, S. L. et al. Phase diagram and electronic indication of high-temperaturesuperconductivity at 65K in single-layer FeSe films. Nature Mater. 12,605–610 (2013).

14. Tan, S. Y. et al. Interface-induced superconductivity and strain-dependent spindensity waves in FeSe/SrTiO3 thin films. Nature Mater. 12, 634–640 (2013).

15. Zhang, W. et al. Direct observation of high-temperature superconductivity inone-unit-cell FeSe films. Chin. Phys. Lett. 31, 017401 (2014).

16. Schilling, A., Cantoni, M., Guo, J. D. & Ott, H. R. Superconductivity above130K in the Hg–Ba–Ca–Cu–O system. Nature 363, 56–58 (1993).

17. Kamihara, Y., Watanabe, T., Hirano, M. & Hosono, H. Iron-based layeredsuperconductor La[O1−xFx ]FeAs (x=0.05−0.12) with Tc=26K. J. Am. Chem.Soc. 130, 3296–3297 (2008).

18. Song, Y. J. et al. Superconducting properties of a stoichiometric FeSe compoundand two anomalous features in the normal state. J. Korean Phys. Soc. 59,312–316 (2011).

19. Hasegawa, S. et al. Electrical conduction through surface superstructuresmeasured by microscopic four-point probe. Surf. Rev. Lett. 10, 963–980 (2003).

20. Smits, F. M. Measurement of sheet resistivities with the four-point probe. BellSyst. Tech. J. 37, 711–718 (1958).

21. Valdes, L. B. Resistivity measurements on germanium for transistors. Proc. Inst.Radio Eng. 42, 420–427 (1954).

22. Zhu, Y. Y., Tsai, C. F. & Wang, H. Y. Atomic interface sequence, misfit strainrelaxation and intrinsic flux-pinning defects in different YBa2Cu3O7−δ

heterogeneous systems. Supercond. Sci. Technol. 26, 025009 (2013).23. Nestler, T. et al. Increased cubic–tetragonal phase transition temperature and

resistivity hysteresis of surface vacuum annealed SrTiO3. Appl. Phys. A 105,103–109 (2011).

24. Tinkham, M. Introduction to Superconductivity 2nd edn, Ch. 4 (DoverPublications, 2004).

25. Sun, Y. et al.High temperature superconducting FeSe films on SrTiO3

substrates. Sci. Rep. 4, 6040 (2014).26. Lee, J. J. et al. Evidence for pairing enhancement in single unit cell FeSe on

SrTiO3 due to cross-interfacial electron–phonon coupling. Preprint athttp://arxiv.org/abs/1312.2633 (2013).

27. Coh, S., Cohen, M. L. & Louie, S. G. Structural template increaseselectron–phonon interaction in an FeSe monolayer. Preprint athttp://arxiv.org/abs/1407.5657 (2014).

AcknowledgementsWe acknowledge helpful discussions with X. Ma, L. Wang, W. Zhang, D. Feng, F. Zhang,N. Samarth and T. Leggett. Financial support from the National Basic Research Programof China (Grant Nos 2012CB927400, 2011CB921902, 2013CB921902 and2011CB922200), NSFC (Grant Nos 11227404, 11374206, 91021002, 11274228, 10904090,11174199, 11134008, 11274229 and 1147198) and the Shanghai Committee of Science

4 NATUREMATERIALS | ADVANCE ONLINE PUBLICATION | www.nature.com/naturematerials© 2014 Macmillan Publishers Limited. All rights reserved.

NATUREMATERIALS DOI: 10.1038/NMAT4153 LETTERSand Technology, China (Grant Nos 12JC1405300, 13QH1401500 and 10JC1407100) isgratefully acknowledged.

Author contributionsJ-F.G. and Z-L.L. conducted the experiments. C.L., Q-K.X. and J-F.J. designed theexperiments and provided financial and other supports for the experiments. C.L, Y.L.,Q-K.X., J-F.G., C-L.G., D.Q. and J-F.J. analysed the data. C.L, Y.L., Q-K.X., and J-F.J.wrote the paper.

Additional informationSupplementary information is available in the online version of the paper. Reprints andpermissions information is available online at www.nature.com/reprints.Correspondence and requests for materials should be addressed to C.L., Q-K.X. or J-F.J.

Competing financial interestsThe authors declare no competing financial interests.

NATUREMATERIALS | ADVANCE ONLINE PUBLICATION | www.nature.com/naturematerials 5© 2014 Macmillan Publishers Limited. All rights reserved.