Electronic origin of high-temperature superconductivity in single-layer FeSe superconductor

Abstract

The recent discovery of high-temperature superconductivity in iron-based compounds has attracted much attention. How to further increase the superconducting transition temperature (T_c ) and how to understand the superconductivity mechanism are two prominent issues facing the current study of iron-based superconductors. The latest report of high-T_c superconductivity in a single-layer FeSe is therefore both surprising and significant. Here we present investigations of the electronic structure and superconducting gap of the single-layer FeSe superconductor. Its Fermi surface is distinct from other iron-based superconductors, consisting only of electron-like pockets near the zone corner without indication of any Fermi surface around the zone centre. Nearly isotropic superconducting gap is observed in this strictly two-dimensional system. The temperature dependence of the superconducting gap gives a transition temperature T_c ~ 55 K. These results have established a clear case that such a simple electronic structure is compatible with high-T_c superconductivity in iron-based superconductors.

ARTICLE

Received 27 Apr 2012 | Accepted 31 May 2012 | Published 3 Jul 2012 DOI:10.1038/ncomms1946

Electronic origin of high-temperature superconductivity in single-layer FeSe superconductor

Defa Liu1, Wenhao Zhang2,3, Daixiang Mou1, Junfeng He1, Yun-Bo Ou3, Qing-Yan Wang2,3, Zhi Li3, Lili Wang3, Lin Zhao1, Shaolong He1, Yingying Peng1, Xu Liu1, Chaoyu Chen1, Li Yu1, Guodong Liu1, Xiaoli Dong1, Jun Zhang1, Chuangtian Chen4, Zuyan Xu4, Jiangping Hu3,5, Xi Chen2, Xucun Ma3, Qikun Xue2 & X.J. Zhou1

The recent discovery of high-temperature superconductivity in iron-based compounds has attracted much attention . How to further increase the superconducting transition temperature ( Tc) and how to understand the superconductivity mechanism are two prominent issues facing the current study of iron-based superconductors. The latest report of high- Tc superconductivity in a single-layer FeSe is therefore both surprising and signicant. Here we present investigations of the electronic structure and superconducting gap of the single-layer FeSe superconductor. Its Fermi surface is distinct from other iron-based superconductors, consisting only of electron-like pockets near the zone corner without indication of any Fermi surface around the zone centre. Nearly isotropic superconducting gap is observed in this strictly two-dimensional system. The temperature dependence of the superconducting gap gives a transition temperature Tc~ 55K.

These results have established a clear case that such a simple electronic structure is compatible with high- Tc superconductivity in iron-based superconductors.

1 National Lab for Superconductivity, Beijing National Laboratory for Condensed Matter Physics, Institute of Physics, Chinese Academy of Sciences , Beijing 100190, China.2 State Key Lab of Low-Dimensional Quantum Physics, Department of Physics, Tsinghua University , Beijing 100084 , China . 3 Beijing National Laboratory for Condensed Matter Physics, Institute of Physics, Chinese Academy of Sciences , Beijing 100190 , China. 4 Technical Institute of Physics and Chemistry, Chinese Academy of Sciences, Beijing 100190, China.5 Department of Physics, Purdue University, West Lafayette, Indiana 47907, USA. Correspondence and requests for materials should be addressed to X.J.Z. (email: [email protected] ) or to Q.X. (email: [email protected]).

1

NATURE COMMUNICATIONS | 3:931 | DOI: 10.1038/ncomms1946 | www.nature.com/naturecommunications

2012 Macmillan Publishers Limited. All rights reserved.

ARTICLE

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms1946

The latest report of high-temperature superconductivity signature in single-layer FeSe 1 is signicant because it is possible to break the superconducting transition temperature record

(maximum Tc ~ 55 K) that has been stagnant since the discovery of the iron-based superconductors in 2008 26. Such a high Tc is unexpected in the FeSe system, because the bulk FeSe exhibits a Tc of only 8 K at ambient pressure 5, although it can be enhanced to 36.7 K under high pressure 7. The Tc is still unusually high, even considering the newly discovered intercalated FeSe system A x Fe 2 y Se 2 (A=K,

Cs, Rb and Tl) with a Tc of 32K at ambient pressure 8,9 and possible Tc near 48 K under high pressure 10. Understanding the origin of high-temperature superconductivity in such a two-dimensional (2D) FeSe system is crucial to understanding the superconductivity mechanism in the Fe-based superconductors in particular, and providing key insights on how to achieve high-temperature superconductivity in general.

As the physical properties and superconductivity of a material are dictated by its underlying electronic structure, it is essential to carry out such investigations on the high Tc single-layer FeSe super-conductor that shows advantages in a number of aspects. First, among the Fe-based superconductors discovered so far 26, FeSe superconductor has the simplest crystal structure consisting only of FeSe layer that is considered to be the essential building block for superconductivity in the Fe-based compounds 5. Second, single-layer FeSe, composed of a Se Fe Se stack along the c axis with a thickness of 5.5 , is a strictly 2D system. Th is gives rise to a 2D electronic structure without complications of k z momentum along the c direction. To obtain a complete electronic structure and to pin down a gap node in the superconducting state, one has to exhaust all the three-dimensional (3D) momentum space that is usually diffi cult in typical 3D bulk superconductors 11. Th ird, the FeSe superconductor is an ideal system where the superconducting phase is pure and well characterized. Th is is in contrast to the A x Fe 2y Se 2 (A=K, Cs, Rb,

Tl and etc.) superconducting system 8,9,12 where there exists phase inhomogeneity and the real superconducting phase remains not clearly identied. Most importantly, the discovery of high- Tc super-conductivity in the single-layer FeSe 1 provides a venue to sort out the key components in achieving high-temperature superconductivity in the Fe-based compounds.

Here we report detailed investigation of electronic structure and superconducting gap in the single-layer FeSe superconductor by high-resolution angle-resolved photoemission spectroscopy (ARPES) measurements. Its Fermi-surface topology is distinct from other Fe-based superconductors. It consists only of electron-like pockets near the zone corner without indication of any Fermi surface around the zone centre. We observed nearly isotropic super-conducting gap around the Fermi surface; it rules out the existence of gap node in this 2D system. Detailed temperature dependence of the superconducting gap gives a superconducting transition temperature of Tc~55K.Th ese results indicate that such a simple distinct Fermi-surface topology is compatible with the occurrence of high-temperature superconductivity in the Fe-based superconductors. Th ey will shed lights on understanding the superconductivity mechanism and exploring new superconductors with higher Tc in the Fe-based superconductors.

ResultsFermi surface and band structure. Figure 1a shows Fermi-surface mapping of the single-layer FeSe thin lm covering multiple Brillouin zones. In comparison, we also present Fermi surface of (Tl,Rb) x Fe 2 y Se 2 superconductor ( Tc=32K) (Fig. 1b)1316 and (Ba 1.6K0.4)Fe2As 2 (Tc=35K)(Fig. 1c)1720 , as well as the Fermi surface of FeSe by band-structure calculations ( Fig. 1d ) 21. The band structure along some typical cuts ( Fig. 2d ) are shown in Fig. 2a,c . The photoemission spectra (energy distribution curves, EDCs) corresponding to Fig. 2a are shown in Fig. 2b . For the single-layer FeSe

superconductor, we only observe a clear electron-like Fermi surface around M(

,

)(denoted as

hereaer) ( Fig. 1a ) but do not observe any indication of Fermi surface around the

(0,0) point. The

Fermi surface is nearly circular; the corresponding Fermi momentum ( kF ) for the cut crossing the M point ( Fig. 2a ) is approximately 0.25 in a unit of

/a (lattice constant a=3.90) that is signicantly smaller than 0.35

/a found in (Tl,Rb) x Fe 2y Se 2 superconductor 14.

Considering that the Fermi surface around M consists of two degenerate Fermi-surface sheets ( Fig. 1d ), this yields an electron counting of about 0.10 electrons / Fe. Th is is considerably smaller than that in (Tl,Rb) x Fe 2y Se 2 superconductor (0.18 electrons per Fe when only considering the M point electron Fermi-surface sheets).

Th is value in the single-layer FeSe superconductor is closer to the optimal doping level in electron-doped Ba(Fe 1x Co x )As2 (x~0.07)

system 22 . On the other hand, the

band bottom in single-layer FeSe lies at 60 meV below the Fermi level, slightly deeper than the 50meVin(Tl,Rb) x Fe 2y Se 2 superconductor 14.The Fermi-surface size and the bandwidth together give an eective electron mass of 2.7me (me is free electron mass) in the single-layer FeSe and 6.1 m e in (Tl,Rb) x Fe 2y Se 2 superconductor. Th is indicates that in single-layer FeSe, the electrons are lighter than those in (Tl,Rb) x Fe 2y Se 2 superconductor, possibly related to a weaker electron correlation.

The observation of only electron-like Fermi-surface sheets around M makes the single-layer FeSe superconductor distinct from all the other Fe-based superconductors. Most of the Fe-based superconductors, such as (Ba,K)Fe 2As 2, are characterized by hole-like Fermi-surface sheets around the

point according

2

1

0

1

2

a

2

b

1

0

1

2

[afii9828]

[afii9828]

k y ([afii9843]/a)

k y ([afii9843]/a)

[afii9825]

[afii9826]

[afii9796]

[afii9796]

High

E

E

2

2

1

0

1

2

2

1

0

1

2

Low

kx ([afii9843]/a)

kx ([afii9843]/a)

c

d

2

M2

M1

1

0

1

2

1

0

1

2

[afii9826]

[afii9825]

[afii9796]

k y ([afii9843]/a)

k y([afii9843]/a)

[afii9796]

M3 M4

2

1

0

1

2

2

1

0

1

2

kx ([afii9843]/a)

kx ([afii9843]/a)

Figure 1 | Fermi surface of the single-layer FeSe superconductor. ( a) Fermi-surface mapping of single-layer FeSe measured at 20 K that consists only of the electron-like Fermi-surface sheet (

) around M(

,

). (b) Fermi-surface mapping of (Tl,Rb) xFe2ySe2 superconductor ( Tc=32K) that consistsof electron-like Fermi-surface sheet (

) around M and also electron-like

) around

(0,0)14. (c) Fermi-surface mapping of (Ba 1.6K0.4)Fe2As2 superconductor ( Tc=35K) that consists of hole-like

Fermi-surface sheets (

and

) around

and complex Fermi surface near M region 17. The spectral weight near the second zone centres is absent because the measured region did not cover this area. ( d) Fermi surface of

-FeSe by band-structure calculations for kz=0 (blue thick lines) that consists of hole-like Fermi-surface sheets around

and two electron-like Fermi-surface sheets around M (ref. 21). For convenience, the four equivalent M points are labelled as M1(

,

), M2(

,

), M3(

,

) and M4(

,

).

Fermi-surface sheets (

and

2

NATURE COMMUNICATIONS | 3:931 | DOI: 10.1038/ncomms1946 | www.nature.com/naturecommunications

2012 Macmillan Publishers Limited. All rights reserved.

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms1946

ARTICLE

a

[afii9828]1 [afii9828]2

[afii9826]

b d

0

M2 M1

M4

[afii9825] [afii9828]

80 meV 60 meV

EE F(eV)

EE F(eV)

0.1

0.2

0.3

0.4

EDC intensity (a.u.)

Cut1

[afii9796]

[afii9828]1

[afii9828]2

High

[afii9796]

[afii9796]

Cut10

M3

Low

0.4 0.2 0 0.4 0.2 0 0.4 0.2 0

0.2

0.4

Momentum ([afii9843]/a)

1.8 1.6

M

1.4 1.2 1

Cut2

EEF (eV) EEF (eV)

c

0

0.05

0.1

0.15

High

Cut 2 Cut 3 Cut 4 Cut 5 Cut 6 Cut 7 Cut 8 Cut 9 Cut 10

Low

1.6 1.2

1.6 1.2

1.6 1.2

1.6 1.2

1.6 1.2

1.6 1.2

1.6 1.2

1.6 1.2

1.6 1.2

Momentum ([afii9843]/a)

Figure 2 | Band structure and photoemission spectra of the single-layer FeSe superconductor. (a) Band structure along the cut 1 crossing the

point

(left panel) and along the cut 2 crossing the M3 point (right panel). The pink dashed line in the left panel shows schematically a hole-like band near the point with its top at 80 meV below the Fermi level. The purple dashed line in the right panel shows schematically an electron-like band with its bottom at 60 meV below the Fermi level. ( b) Photoemission spectra (EDCs) corresponding to image ( a) for the cut 1 (left panel) and the cut 2 (right panel). The blue line in left panel corresponds to EDC at

point whereas the two red curves in right panel correspond to EDCs at two Fermi momenta (

1 and

2 of

the electron-like band near M. (c) Detailed band-structure evolution near the M3 region from the cut 2 (left-most panel) to the cut 10 (right-most panel). ( d) The location of the momentum cuts along some high-symmetry lines and near the M3 region.

to band-structure calculations 23,24 and ARPES measurements (Fig.1c)1720 . It is also dierent from the Fermi-surface topology of the recently discovered A x Fe 1y Se 2 superconductors, where the

Fermi-surface sheets remain present around the

point(Fig.1b), although all observed Fermi-surface sheets are exclusively electron-like 1316.Th e 2D nature of the single-layer FeSe also removes any ambiguity of the possible kz complications that are common in other

Fe-based systems. We note that, compared with the band structure of (Tl,Rb) x Fe 2y Se 2 superconductor, the band structure of single-layer FeSe is not a simple rigid band shi . First, as mentioned above, the

band shape near

M is signicantly dierent leading to a lighter eective mass. Second, the much lower electron doping in FeSe superconductor is expected to shi the bands towards the Fermi level when comparing with the bands in (Tl,Rb) x Fe 2y Se 2 super-conductor. However, the top of the hole-like

band near

point in FeSe superconductor ( Fig. 2a , le panel, 80 meV below the Fermi level) is similar to that in (Tl,Rb) x Fe 2y Se 2 superconductor 14.The bottom of the electron-like

band near M in FeSe ( Fig. 2a , right panel, 60 meV below the Fermi level) is even deeper than that in (Tl,Rb) x Fe 2ySe 2 superconductor (~50meV below the Fermi level) 14, which is in opposite direction from expectation.

Superconducting gap. The identication of clear

Fermi-surface sheet near M point makes it possible to investigate the superconducting gap in this new superconductor. We start rst by examining the temperature dependence of the superconducting gap. Figure 3a shows the photoemission images along the cut near M3(itslocation shown in Fig. 3e ) at dierent temperatures. The corresponding photoemission spectra (EDCs) on the Fermi momentum are shown in Fig. 3b, where sharp peaks develop at low temperatures.

To visually inspect possible gap opening and remove the eect of Fermi distribution function near the Fermi level, these original EDCs are symmetrized ( Fig. 3c ), following the procedure commonly used in the study of high-temperature cuprate superconductors 25. For the

pocket near

M, there is a clear gap opening at low temperature (20 K), as indicated by a dip at the Fermi energy in the symmetrized EDCs ( Fig. 3c ). With increasing temperature, the dip at EF is gradually lled up and is almost invisible near 50 55 K. As it is commonly observed from ARPES on the Fe-based superconductors that the gap closes at the superconducting temperature 17,14, Fig. 3c suggests that the single-layer FeSe sample, we measured, has a Tc around (555K) (see sample temperature calibration in Methods). This value is reproducible with an independent measurement on another sample. Th e temperature dependence of the superconducting gap size ( Fig. 3d ) follows the form of the BCS theory of superconductivity. Due to technical diffi culties, it is not yet possible to directly measure superconducting transition temperature on the single-layer FeSe thin lm by transport or magnetic measurements 1.The gap size measured at 20 K i 15 meV. As a high-resolution ARPES measurement on (Tl,Rb) x Fe 2y Se 2 superconductor ( Tc=32K) gives a superconducting gap of ~9.7meV12 , if we assume the same ratio between the superconducting gap size and Tc , 15 meV gap in the single-layer FeSe would correspond to ~ 50 K that is consistent with the value from the temperature-dependent gap measurement.

Now, we come to the momentum-dependent measurements of the superconducting gap. For this purpose, we took high-resolution Fermi-surface mapping (energy resolution of 4 meV) of the

pocket

at M ( Fig. 4d ). Figure 4a shows photoemission spectra (EDCs) around the

Fermi surface measured in the superconducting state ( T= 20K); the corresponding symmetrized photoemission spectra

3

NATURE COMMUNICATIONS | 3:931 | DOI: 10.1038/ncomms1946 | www.nature.com/naturecommunications

2012 Macmillan Publishers Limited. All rights reserved.

ARTICLE

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms1946

a b c

0

65 K

60 K

55 K

50 K

45 K

40 K

35 K

30 K

25 K

20 K

EE F(eV)

0.05

0.1

0.15

65 K

60 K

55 K

50 K

45 K

40 K

35 K

30 K

25 K

20 K

High

20 K

35 K

55 K

65 K

Momentum ([afii9843]/a)

EDC intensity (a.u.)

1.6 1.2 1.6 1.2 1.6 1.2 1.6 1.2

0

e d

15

M2 M1

M4

10

5

0

Gap (meV)

[afii9796]

M3

20 40 40 40

40 40

60 0 0

80 80

80 80

EEF (meV) EEF (meV)

Temperature (K)

Low

Figure 3 | Temperature dependence of superconducting gap of the single-layer FeSe superconductor. ( a) Typical photoemission images along the momentum cut near the M3 point (blue line in ( e)) measured at different temperatures. The temperature labelled here is a usual nominal sample temperature (see sample temperature calibration in Methods). Each image is divided by the corresponding Fermi distribution function to highlight opening or closing of an energy gap. ( b) Photoemission spectra at the Fermi crossings of

Fermi surface and their corresponding symmetrized spectra ( c) measured at different temperatures. ( d) Temperature dependence of the superconducting gap. The error bars are dened as s.d. The dashed lineis a BCS gap form that gives a gap size of 15 meV at zero temperature. ( e) Location of the momentum cut near M3 point (blue line). The dashed shows schematically the electron-like Fermi-surface near M point.

are shown in Fig. 4b . We have carried out two independent measurements on two samples; the extracted superconducting gap for both samples is nearly isotropic. Th e sample # 1 shows a superconducting gap with a size of (13 2)meV whereas the other sample #2 has a gap of (15 2)meV. We note that superconductivity in single-layer FeSe is sensitive to the preparation and post-annealing conditions. Th e slight dierence in the gap size is possibly due to slightly dierent post-annealing conditions that caused a subtle variation of the superconducting phase. In both samples, no indication of zero gap is observed around the

Fermi surface. Since the measured single-layer FeSe is strictly 2D in nature, it avoids complications from 3D Fermi surface associated with bulk materials. This establishes unambiguously that there is no node in the superconducting gap (Fig. 4c).

Discussion

Although the observed Fermi-surface topology and nearly isotropic nodeless superconducting gap seem to be reminiscent to that of A x Fe 2y Se 2 superconductors 1316 , their revelation in the single-layer FeSe system with high Tc provides much profound and decisive information on understanding the physics and superconductivity mechanism of Fe-based superconductors. First, in the newly discovered A x Fe 2y Se 2 superconductors, because of the coexistence of many dierent phases, it remains controversial on which phase is truely superconducting 12.Th e single-layer FeSe system provides a clear-cut case that the Fe-vacancy-free FeSe phase can support high-temperature superconductivity. Second, the complete absence of Fermi surface around the

point removes any scattering channel between the

Fermi surface sheet and M-point Fermisurface sheet. Th is makes it more straightforward than that in A x Fe 2y Se 2 superconductors where Fermi surface sheets are still present around point albeit being electron-like. Th ird, because of the 2D nature of

the single-layer FeSe, the present observation can make a straightforward conclusion that the superconducting gap is nearly isotropic and nodeless. In 3D A x Fe 2y Se 2 superconductors, such a conclusion cannot be made unambiguously before one carries out measurements for the entire momentum space along all kz (ref. 11).

Th e present work has important implications on the pairing mechanism in Fe-based superconductors. Th e electronic structure of the Fe-based compounds involves all ve Fe 3d-orbitalsforming multiple Fermi-surface sheets: hole-like Fermi-surface sheets around

and electron-like ones around M (refs 23,24). It has been proposed that the interband scattering between the hole-like bands near

and the electron-like bands near M is responsible for electron pairing and superconductivity 24,2631.Th e absence of Fermi surface around

point in the single-layer FeSe superconductor, with the top of the hole-like band lying well below the Fermi level (80 meV below EF , Fig. 2a , le panel), completely rules out the electron scattering possibility across the Fermi-surface sheets between the

and M points. Furthermore, with the absence of Fermi surface around

, electrons can only scatter across the Fermi-surface sheets between M points that is predicted to result in d-wave superconducting gap 24,3234 . It is argued that two Fermi-surface sheets around a given M point with opposite phases can give rise to gap nodes around the Fermi surface 35 . In this case, our identication of nodeless superconducting gap does not favour such a scheme of electron scattering across M -point Fermi-surface sheets as the pairing mechanism in the Fe-based superconductors . Th is is in contrast to the low- Tc FeSe case ( Tc=8K)wherethereisaclearindication of nodes in the superconducting gap 36. Other alternative pictures, such as the interaction of local Fe-magnetic moment 3741, or orbital uctuations 42,43 , need to be invoked to understand high-temperature superconductivity in the single-layer FeSe superconductor.

NATURE COMMUNICATIONS | 3:931 | DOI: 10.1038/ncomms1946 | www.nature.com/naturecommunications

2012 Macmillan Publishers Limited. All rights reserved.

4

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms1946

ARTICLE

a

b c

[afii9797]

P1

P1

[afii9835]

P4

P4

M

5

EDC Intensity (a.u.)

80 60 40 40

P12

10

15

P8

20 meV

P8

d

[afii9796]

[afii9835]

P12

P8

P15

P15

M3

P1

20 0 0

P15

EEF (meV) EEF (meV)

40

Figure 4 | Momentum dependence of superconducting gap of the single-layer FeSe superconductor. ( a) EDCs along the

Fermi surface and ( b) their corresponding symmetrized EDCs measured at 20 K. ( c) Momentum dependence of the superconducting gap, obtained by picking the peak position in the symmetrized EDCs, along the

Fermi surface. Two superconducting samples are independently measured and their superconducting gaps are both shown with the pink solid circles representing the gap of sample # 1 and the blue solid circles representing the gap of sample # 2. The error bars are dened as s.d. ( d) The Fermi-surface mapping near M3 and the corresponding Fermi crossings for sample # 1. The violet circle represents the Fermi surface. The red solid circles represent measured Fermi momenta that are labelled as P1 to P15 (for clarity, the intermittent points are not labelled).

The labels in ( a) and ( b) correspond to the Fermi momentum labels in ( d).

Th ere are a couple of intriguing issues that merit further investigations. First, there is a disagreement between the previous tunnelling measurements 1 and our present ARPES measurements. Although two peaks at 20.1 meV and 9.0 meV are developed in the tunnelling spectrum in the superconducting state of a single-layer FeSe superconductor 1 , the present ARPES measurements do not reveal the two-gap structure in the momentum-resolved spectra. Th e maximum superconducting gap ( ~ 15 meV) in our measurement is smaller than 20.1 meV but larger than 9.0 meV. Moreover, with the identication of only one Fermi surface near M , and a nearly isotropic superconducting gap, the observation of two gaps can not be explained by multiple gaps on dierent Fermi-surface sheets, as found in other Fe-based supercouductors like (Ba,K)Fe 2Se 2 (refs 17,18). As tunnelling is a local probe, it is highly unlikely to attribute these two peaks as due to phase separation on such a small scale. Although we are carrying out experiment to ensure that there is no change in the sample during the de-capping process of the amorphous Se protection-layer in ARPES experiment, one conjecture is whether both the FeSe thin lm and the interface between FeSe and SrTiO 3 may become superconducting with dierent gap sizes. Th is comes to another prominent issue whether the observed superconductivity is due to FeSe itself or the interface between the FeSe and the SrTiO 3 substrate. Th e main electronic feature of electron-like

Fermi surface is consistent with the band-structure calculations of FeSe, and the energy gap on this particular Fermi surface closes at high temperature. Also, we do not see signature of possible 2D electron gas in the interface as no electron-like Fermi surface around the

point is observed. These observations seem to favour the attribution of superconductivity to the FeSe layer. Because the lattice constant for bulk FeSe is a=3.76(ref.5)and for SrTiO 3 is a=3.905, an apparent lattice mismatch between FeSe

and SrTiO 3 is expected to exert a strong tensile strain on the FeSe layer when FeSe is epitaxially grown on SrTiO 3(001) substrate. Such a tensile pressure on FeSe layer may have an important role, as it has been shown that superconductivity in FeSe is rather sensitive to high pressure 7,10. However, we cannot fully rule out interface super-conductivity, when considering FeSe, may become superconducting owing to the proximity eect and weak signal from the buried interface that may be beyond our detection. It remains to be explored why only the very rst layer is superconducting while the other layers above are not superconducting 1 . Clearly, the answer to these questions will provide crucial information on achieving even higher Tc in searching for new high-temperature superconductors.

Methods

Thin lm preparation and post-annealing. Th e single-layer FeSe thin lms were grown on SrTiO 3(001) substrate by the molecular beam epitaxy (MBE) method and were characterized by scanning tunnelling microscope. Th e details can be found in (ref. 1). To transfer the thin lm sample from the MBE preparation chamber to a dierent ARPES measurement chamber, the prepared thin lm is covered by an amorphous Se capping layer. Th e Se layer was desorbed by heating up the samples to 400 C for 2 h before ARPES measurements.

Sample temperature calibration. Because the FeSe thin lm on SrTiO 3 substrate is dierent from our usual cleaved sample, also owing to their dierent mounting methods on the cold nger of the cryostat, at the same temperature of thecold nger, the temperature on the FeSe thin lm is higher than that for the usual cleaved sample. Th erefore, the usual sample temperature sets a low limit to the temperature of the FeSe thin lm. To measure the temperature of the FeSe thin lm more accurately, we simulate the FeSe thin lm by gluing a piece of polycrystalline gold foil on the similar SrTiO 3 substrate. By measuring the Fermi level of the gold from our high-resolution laser photoemission, we can measure the temperature of gold on SrTiO 3 from the gold Fermi level width tted by the Fermi-distribution function. As the Fermi-level width contains a couple of contributions includingthe temperature broadening, instrumental resolution and cleanliness of gold, the

5

NATURE COMMUNICATIONS | 3:931 | DOI: 10.1038/ncomms1946 | www.nature.com/naturecommunications

2012 Macmillan Publishers Limited. All rights reserved.

ARTICLE

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms1946

estimated temperature sets an upper limit to the gold temperature that is close to the temperature of FeSe thin lm on SrTiO 3. We found the temperature measured from this gold reference is higher than the nominal usual sample temperature by ~10K. Th e real temperature of the FeSe thin lm then lies in between these two temperature limits, nominal sample temperature and gold temperature, with an error bar of 5K.

High-resolution ARPES measurements. High-resolution angle-resolved photoemission measurements were carried out on our lab system equipped witha Scienta R4000 electron energy analyzer 44. We use helium discharge lamp as the light source that can provide photon energies of h =21.218eV (helium I). The energy resolution was set at 10 meV for the Fermi-surface mapping ( Fig. 1a ) and band-structure measurements ( Fig. 2 ) and at 4 meV for the superconducting gap measurements ( Figs 3 and 4 ). Th e angular resolution is ~ 0.3 degree. The Fermi level is referenced by measuring on a clean polycrystalline gold that is electrically connected to the sample. Th e sample was measured in vacuum with a base pressure better than 5 10 11Torr.

References

1. Wang, Q. Y. et al. Interface-induced high-temperature superconductivity in single unit-cell FeSe lms on SrTiO 3. Chin. Phys. Lett. 29, 037402 (2012).

2. Kamihara, Y. et al. Iron-based layered superconductor La[O1 x F x ]FeAs (x = 0.05-0.12) with T c = 26K.J. Am. Chem. Soc. 130, 32963297 (2008).

3. Ren, Z. A. et al. Superconductivity at 55K in iron-based F-doped layered quaternary compound Sm[O 1 x F x ]FeAs.Chin. Phys. Lett. 25, 22152216 (2008).

4. Rotter, M. et al. Superconductivity at 38K in the iron arsenide (Ba1 x K x )Fe2As 2. Phys. Rev. Lett. 101, 107006 (2008).

5. Hsu, F. C. et al. Superconductivity in the PbO-type structure

-FeSe.Proc. Natl

Acad. Sci. USA 105, 1426214264 (2008).

6. Wang, X. C. et al. Th e superconductivity at 18 K in LiFeAs system . Solid State Commun. 148, 538540 (2008).

7. Medvedev, S. et al. Electronic and magnetic phase diagram of

-Fe1.01Se with superconductivity at 36.7 K under pressure . Nat. Mater. 8, 630633 (2009).

8. Guo, J. G. et al. Superconductivity in the iron selenide K x Fe 2Se 2 (0x1.0). Phys. Rev. B 82, 180520(R) (2010).

9. Fang, M. H. et al. Fe-based high temperature superconductivity with T c=31K bordering an insulating antiferromagnet in (Tl,K)Fe x Se 2 Crystals.Europhys.

Lett. 94, 27009 (2011).10. Sun, L. L. et al. Reemerging superconductivity at 48K in iron chalcogenides. Nature 483, 6769 (2012).

11. Zhang, Y. et al. Nodal superconducting gap structure in ferropnictide superconductor BaFe 2(As0.7P0.3)2. Nat. Phys. 8, 371375 (2012).

12.Mou, D. X., Zhao, L. & Zhou, X. J. Structural, magnetic and electronic properties of the iron chalcogenide A x Fe 2 y Se 2 (A=K, Cs, Rb, and Tl, etc.)

superconductors. Front. Phys. 6, 410428 (2011).13. Zhang, Y. et al. Nodeless superconducting gap in A x Fe 2Se 2 (A=K,Cs) revealed by angle-resolved photoemission spectroscopy . Nat. Mater. 10, 273277 (2011).

14. Mou, D. X. et al. Distinct Fermi surface topology and nodeless superconducting gap in a (Tl 0.58Rb0.42)Fe1.72Se 2 superconductor. Phys. Rev. Lett. 106, 107001 (2011).

15. Zhao, L. et al. Common Fermi-surface topology and nodeless superconducting gap of K 0.68Fe 1.79Se 2 and (Tl0.45K0.34)Fe1.84Se 2 superconductors revealed via angle-resolved photoemission.Phys. Rev. B. 83, 140508(R) (2011).

16. Qian, T. et al. Absence of a holelike Fermi surface for the iron-basedK 0.8Fe 1.7Se 2 superconductor revealed by angle-resolved photoemission spectroscopy.Phys. Rev. Lett. 106, 187001 (2011).

17. Zhao, L. et al. Multiple nodeless superconducting gaps in (Ba0.6K0.4)Fe2As 2 superconductor from angle-resolved photoemission spectroscopy . Chin. Phys.

Lett. 25, 44024405 (2008).18. Ding, H. et al. Observation of Fermi-surface-dependent nodeless superconducting gaps in Ba 0.6K0.4Fe 2As 2. Europhys. Lett. 83, 47001 (2008).

19. Evtushinsky, D. V. et al. Momentum dependence of the superconducting gap in Ba 1 x K x Fe 2As 2. Phys. Rev. B 79, 054517 (2009).

20. Wray, L. et al. Momentum dependence of superconducting gap, strong-coupling dispersion kink, and tightly bound Cooper pairs in the high-Tc (Sr,Ba) 1 x (K,Na) x Fe 2As 2 superconductors.Phys. Rev. B 78, 184508 (2008).

21. Tamai, A. et al. Strong electron correlations in the normal state of the iron-based FeSe 0.42Te 0.58 superconductor observed by angle-resolved photoemission spectroscopy.Phys. Rev. Lett. 104, 097002 (2010).

22. Ni, N. et al. Eects of Co substitution on thermodynamic and transport properties and anisotropic H c 2 in Ba(Fe1 x Co x )2As 2 single crystals.Phys.

Rev. B 78, 214515 (2008).23.Singh, D. J. & Du, M.- H. Density functional study of LaFeAsO1 x F x : A low carrier density superconductor near itinerant magnetism . Phys. Rev. Lett. 100, 237003 (2008).

24. Kuroki, K. et al. Unconventional pairing originating from the disconnected Fermi surfaces of superconducting LaFeAsO 1 x F x . Phys. Rev. Lett. 101, 087004 (2008).

25. Norman, M. R. et al. Phenomenology of the low-energy spectral function in high-T c superconductors.Phys. Rev. B 57, R11093 (1998).

26.Mazin, I. I.et al. Unconventional superconductivity with a sign reversal in the order parameter of LaFeAsO 1 x F x . Phy. Rev. Lett. 101, 057003 (2008).

27. Wang, F. et al. Functional renormalization-group study of the pairing symmetry and pairing mechanism of the FeAs-based high-temperature superconductor. Phys. Rev. Lett. 102, 047005 (2009).

28. Chubukov, A. V. et al. Magnetism, superconductivity, and pairing symmetry in iron-based superconductors.Phys. Rev. B 78, 134512 (2008).

29. Stanev, V. et al. Spin uctuation dynamics and multiband superconductivity in iron pnictides.Phys. Rev. B 78, 184509 (2008).

30. Wang, F. et al. Antiferromagnetic correlation and the pairing mechanism of the cuprates and iron pnictides: A view from the functional renormalization group studies.Europhys. Lett. 85, 37005 (2009).

31.Mazin, I. I. & Johannes, M. D. A key role for unusual spin dynamics in ferropnictides.Nat. Phys. 5, 141145 (2009).

32. Maier, T. A. et al. d-wave pairing from spin uctuations in the K x Fe 2 y Se 2 superconductors.Phys. Rev. B 83, 100515(R) (2011).

33. Wang, F. et al. Th e electron pairing of K x Fe 2 y Se 2. Europhys. Lett. 93, 57003 (2011).

34. Das, T. et al. Stripes, spin resonance and

d x 2 y 2pairing symmetry in FeSe-based layered superconductors.Phys. Rev. B 84, 014521 (2011).

35.Mazin, I. Symmetry analysis of possible superconducting states in K x Fe 2Se 2 superconductors.Phys. Rev. B 84, 024529 (2011).

36. Song, C. L. et al. Direct observation of nodes and twofold symmetry in FeSe superconductor. Science 332, 14101413 (2011).

37.Yildirim, T. Origin of the 150-K anomaly in LaFeAsO: Competing antiferromagnetic interactions, frustration, and a structural phase transition . Phys. Rev. Lett. 101, 057010 (2008).

38.Si, Q. M. & Abrahams, E. Strong correlations and magnetic frustration in the high T c iron pnictides.Phys. Rev. Lett. 101, 076401 (2008).

39. Wu, J. S. et al. Th eory of the magnetic moment in iron pnictides . Phys. Rev. Lett. 101, 126401 (2008).

40. Fang, C. et al. Th eory of electron nematic order in LaFeAsO . Phys. Rev. B 77, 224509 (2008).

41. Xu, C. K. et al. Ising and spin orders in the iron-based superconductors.Phys. Rev. B 78, 020501(R) (2008).

42. Saito, T. et al. Emergence of fully gapped

s-wave and nodal d-wave states mediated by orbital and spin uctuations in a ten-orbital model of KFe 2Se 2.

Phys. Rev. B 83, 140512(R) (2011).43. Lv, W. et al. Vacancy driven orbital and magnetic order in (K,Tl,Cs) y Fe 2 x Se 2. Phys. Rev. B 84, 155107 (2011).

44. Liu, G. D. et al. Development of a vacuum ultraviolet laser-based angle-resolved photoemission system with a superhigh energy resolution better than 1meV.Rev. Sci. Instrum. 79, 023105 (2008).

Acknowledgements

We thank Dunghai Lee for discussions. X.J.Z. thanks nancial support from theNSFC (10734120) and the MOST of China (973 program No: 2011CB921703 and 2011CB605903). Q.K.X. and X.C.M. thank support from the MOST of China (program No. 2009CB929400 and No. 2012CB921702).

Author contributions

D.F.L., W.H.Z., D.X.M., and J.F.H. contributed equally to this work. X.J.Z., Q.K.X. and X.C.M. proposed and designed the research. W.H.Z., Y.B.O, Q.Y.W., Z.L., L.L.W., X.C., X.C.M and Q.K.X. contributed to MBE thin lm preparation. D.F.L., D.X.M., J.F.H., L.Z., S.L.H., X.L., Y.Y.P., C.Y.C., L.Y., G.D.L., X.L.D., J.Z., Z.Y.X., C.T.C. and X.J.Z. contributed to the development and maintenance of Laser-ARPES system. D.F.L., W.H.Z., D.X.M., J.F.H., Y.B.O. and L.Z. carried out the experiment with the assistance from S.L.H., X.L. and Y.Y.P.. D.F.L., J.F.H., D.X.M., L.Z. and X.J.Z. analyzed the data. X.J.Z. wrote the paper with Q.K.X., J.F.H., D.X.M, D.F.L., L.Z., Y.Y.P. and J.P.H .

Additional information

Competing nancial interests: Th e authors declare no competing nancial interests.

Reprints and permission information is available online at http://npg.nature.com/ reprintsandpermissions/

How to cite this article: Liu, D. et al. Electronic origin of high-temperature superconductivity in single-layer FeSe superconductor. Nat. Commun. 3:931 doi: 10.1038/ncomms1946 (2012).

NATURE COMMUNICATIONS | 3:931 | DOI: 10.1038/ncomms1946 | www.nature.com/naturecommunications

2012 Macmillan Publishers Limited. All rights reserved.

6