Superconducting properties of iron chalcogenide thin films

Abstract

Iron chalcogenides, binary FeSe, FeTe and ternary FeTe_xSe_1−x, FeTe_xS_1−x and FeTe:O_x, are the simplest compounds amongst the recently discovered iron-based superconductors. Thin films of iron chalcogenides present many attractive features that are covered in this review, such as: (i) easy fabrication and epitaxial growth on common single-crystal substrates; (ii) strong enhancement of superconducting transition temperature with respect to the bulk parent compounds (in FeTe_0.5Se_0.5, zero-resistance transition temperature T_c0^bulk = 13.5 K, but T_c0^film = 19 K on LaAlO₃ substrate); (iii) high critical current density (J_c ∼ 0.5 ×10⁶ A cm² at 4.2 K and 0 T for FeTe_0.5Se_0.5 film deposited on CaF₂, and similar values on flexible metallic substrates (Hastelloy tapes buffered by ion-beam assisted deposition) with a weak dependence on magnetic field; (iv) high upper critical field (∼50 T for FeTe_0.5Se_0.5, B_c2(0), with a low anisotropy, γ ∼ 2). These highlights explain why thin films of iron chalcogenides have been widely studied in recent years and are considered as promising materials for applications requiring high magnetic fields (20–50 T) and low temperatures (2–10 K).

Introduction

Iron chalcogenides are defined as binary compounds of iron and a chalcogenide element (Ch) belonging to group VIA of the periodic table. Several compounds are formed from the common valence states of Fe and Ch: FeCh, Fe₂Ch₃, Fe₃Ch₄ and FeCh₂. In some cases Fe is not stoichiometric as in Fe_1−αS (α = 0.1–0.2). Iron oxides (Fe₂O₃ and Fe₃O₄) and sulfides (mostly Fe₂S₃) are widely used in the industry and show several interesting properties, except superconductivity, so they are not discussed in this review. Thin films of binary 1:1 iron selenides and tellurides FeCh (Ch = Se, Te) were extensively studied since the early 1970s. Optical, magnetic and others physical properties were investigated [1–7], but resistivity has not been studied at temperatures below 100 K, and thus the occurrence of superconductivity was not recognized. This was probably due to the presence of iron that was historically considered detrimental for superconductivity. The first papers on superconducting properties of pure and doped 1:1 thin films appeared only after the discovery of superconductivity at 7 K in polycrystalline FeSe [8].

Superconducting iron chalcogenides with the 1:1 composition are called the ‘11 family’ of iron-based superconductors. Other families have more complex structures: ‘1111’, such as LaFeAsO_0.89F_0.11 [9] and derived compounds; ‘111’ like LiFe_1−xAs [10], and ‘122’ like BaFe₂As₂ [11]. Even if their critical temperature (T_c) is low compared with those of the 1111 or 122 families, the compounds of 11 family are attractive for at least two reasons: (i) relatively safe synthesis, owing to the absence of the poisonous and volatile arsenic and (ii) potential applications, owing to the high upper critical field B_c2(0) approaching 50 T and its low anisotropy (γ ∼ 2) [12]. The quality of 11 thin films was greatly enhanced recently, and high current densities, J_c, in magnetic field have been reported [12–16].

Excellent reviews on bulk iron chalcogenides have been published as dedicated papers [17] or as chapters in extended overviews of all iron-based superconductors [18–23]. This paper is one of the few reviews entirely dedicated to iron chalcogenide thin films after the work of Li et al [24].

This contribution is an updated review of superconducting properties of 11 thin films, with a special emphasis on recent improvements in the critical temperature, upper critical field and critical current values. This review is divided in five parts. In the first part the synthesis parameters for films fabricated by pulsed laser deposition (PLD) and other techniques are summarized and their influence on the film crystallinity and morphology is reviewed. The second part is dedicated to the superconducting critical temperature and factors that determine its difference from T_c of the parent bulk compounds. The third part concerns the critical fields and related magnetic parameters. The fourth part contains most of the updates to the previous review [24] and is dedicated to recent results on enhanced critical current density and global pinning force. The last part summarizes the state of the art and the open issues in research on iron chalcogenide thin films. The first four parts are divided in subsections dedicated to the compositions¹ reported in literature: FeTe_xSe_1−x, FeTe_xS_1−x, FeSe_x, FeTe and FeTe:O_x. Data on the PLD synthesis parameters, critical temperatures, critical fields, critical currents and global pinning forces of all the films reviewed in this contribution are summarized in tables for convenience and easy comparison.

Synthesis, morphology and crystalline structure

Most iron chalcogenide thin films reported in literature are fabricated by PLD, ablating sintered targets with an yttrium aluminum garnet (YAG) laser (fourth harmonic, λ =266 nm) [25] or a KrF excimer laser (λ =248 nm).

Several research groups explored a wide range of experimental conditions to fabricate high-quality, superconducting epitaxial thin films by PLD, such as deposition temperature; laser frequency, energy and the number of shots; and the distance between substrate and target (table 1). Besides FeTe:O_x films, which are prepared in oxygen atmosphere (P_O2 ⩽ 10⁻⁴ Torr) [26], the ablation of targets is carried out in high vacuum (10⁻⁴–10⁻⁶ Pa). PLD targets are produced by conventional sintering. For FeTe_xSe_1−x, stoichiometric amounts of Fe, Te and Se powders are sealed into a quartz tube and reacted at 400 °C for about one day; then they are pressed into pellets with a diameter of about 13 mm and a thickness of about 7 mm, resealed, and sintered at 800 °C for about 7 days. The same procedure is followed for FeTe_xS_1−x, but the first reaction is carried out at 800 °C for 12 h and the second one at 600 °C, again for 12 h.

Table 1.

Summary of deposition parameters of iron chalcogenide epitaxial films prepared by PLD on various single crystal substrates, including SrTiO₃ (STO), LaAlO₃ (LAO), LaSrAl oxide (LSAO), LaSrAlTa oxide (LSAT), LaSrGa oxide (LSGO) and yttria-stabilized zirconia (YSZ). KrF excimer laser (λ = 248 nm) was used unless noted otherwise.

Material	Substrate	Deposition T (°C)	Frequency (Hz)	Laser shots	Film thickness (nm)	Base pressure (Pa)	Laser energy	Substrate–substrate distance (cm)	Notes	Reference
FeTe0.8S0.2	MgO, STO	200–600	5–10	9000–54000	100–200	5×10−4	200–350 mJ pulse−1			[32]
FeTe0.8S0.2	MgO, STO, LAO	400	5	N/A	80–100	1.33×10−3	0.74–2 J cm−2	4–7	YAG laser	[25]
FeTe0.5Se0.5	STO, LAO	300–450	N/A	N/A	N/A	N/A	3 J cm−2	N/A		[34]
FeTe0.5Se0.5	MgO	180–500	N/A	N/A	400	N/A	N/A	N/A		[35]
FeTe0.5Se0.5	MgO, LAO, STO, LSAO, LSGO, YSZ, LSAT, Al2O3	300	10	N/A	50	1.33×10−3	N/A	N/A		[36]
FeTe0.5Se0.5	CaF2	280	10	N/A	36–175	1.33×10−3	N/A	N/A		[37]
FeTe0.5Se0.5	MgO, LAO, STO, YSZ, LSAT, LSAO, LSGO, CaF2, SrF2, BaF2	300	10	N/A	50	1.33×10−3	N/A	N/A		[38]
FeTe0.5Se0.5	STO, MgO, LAO, CaF2	250–600	3, 10, 30	9000–45000	70–400	(1–2)×10−3	350 mJ pulse−1	5		[15]
FeTe0.5Se0.5	STO	300–650	3–10	N/A	17–90	5×10−7	2 J cm−2	5		[39]
FeTe0.5Se0.5	LAO, STO, YSZ	550	3	N/A	1.2–600	5×10−7	2 J cm−2	5		[40]
FeTe0.5Se0.5	MgO, STO, LAO, YSZ	30–650	3	N/A	1.2–600	5×10−7	2 J cm−2			[41]
FeTe0.5Se0.5	MgO, STO, LAO, YSZ, CaF2, LiF	550	3	N/A	1.2–600	5×10−7	2 J cm−2	5		[16]
FeTe0.5Se0.5	Fe-buffered MgO	450	3	N/A	95	10−8	3–5 J cm−2	N/A	Fe buffer thickness = 20 nm	[14]
FeSex	MgO, STO, LAO	380	N/A	N/A	50–200	1.33×10−4	N/A	N/A		[42]
FeSex	MgO	320 and 500	N/A	N/A	140	N/A	N/A	N/A		[43]
FeSex	MgO	250–500	2	N/A	100–600	N/A	5–6 J cm−2	N/A		[44]
FeSex	STO, LAO	620	2	N/A	N/A	3×10−4 (vacuum) ∼30 (Ar)	100 mJ mm−2	N/A	XeCl laser (308 nm)	[45]
FeSe	MgO, Al2O3, STO, LAO	610	48	N/A	1500–1800	1.33×10−3	1.15 J cm−2	N/A		[50]
FeTe	LSAT, MgO, LAO	540	4	N/A	N/A	2×10−4	100 mJ mm−2	N/A	XeCl laser (308 nm)	[47]
FeTe:Ox		300–450	5	N/A	100	N/A	3 J cm−2	N/A	10−4 Torr O2	[26]

There is very limited mention of superconducting films grown by other common techniques, such as molecular beam epitaxy [27, 28], metalorganic chemical vapor deposition (MOCVD) [29] and sputtering [30]. Recently, very thick (thickness 100 μm) superconducting films of FeSe (T_c^onset = 8 K) were successfully deposited by electrochemical route [31].

This section presents the morphology and crystalline structure of epitaxial iron chalcogenide films. All these films are superconductors; however their superconducting properties, as well as the transport and magnetic properties, are described in other sections of this review. It is worth noting that the preparation of 11 films is safer than the synthesis of 1111 or 122 films, thanks to the absence of poisonous elements, but it still presents some difficulties related to the volatility of chalcogenide elements, which hinders the stoichiometry control.

FeTe_xS_1−x films

Mele et al [32] reported on 100–200 nm thick films of FeTeS grown epitaxially on MgO and SrTiO₃ (STO) substrates with a cube-on-cube orientation (figure 1). Besides an earlier paper on FeTeS films from the same authors [33], there are no other reports on FeTeS thin films prepared with an excimer laser. Yoshida et al [25] deposited epitaxial films showing the same kind of crystalline orientation using a YAG laser. Independently from the preparation technique, an interfacial reaction layer is formed between the film and STO substrate (figure 2(a)), but is absent for films on MgO (figure 2(b)).

Figure 1.

(a), (b) XRD ѳ–2ѳ patterns of FeTe_0.8S_0.2 thin films deposited at 400 °C on (a) SrTiO₃ (100) and (b) MgO (100) single-crystal substrates. The insets show the corresponding XRD rocking curves (001 reflection). (c), (d) XRD in-plane patterns (Φ scan, 101 reflection) of FeTe_0.8S_0.2 thin films deposited at 400 °C on (c) SrTiO₃ (100) and (d) MgO (100) single-crystal substrates. Reproduced with permission from [32]. ©2010 IOP Publishing.

Figure 2.

TEM cross-sectional images of FeTeS thin films deposited at 400 °C (a) on SrTiO₃ (100) single-crystal substrates, showing a possible interfacial layer; (b) on MgO (100) single-crystal substrates. The insets show the corresponding selected area electron diffraction patterns. Reproduced with permission from [32]. ©2010 IOP Publishing.

FeTe_xSe_1−x films

FeTe_xSe_1−x with x = 0.5 presents the highest T_c at the ambient pressure amongst all 11 compounds. Consequently, many groups prepared thin films of FeTe_0.5Se_0.5 before trying other compositions. High-quality epitaxial FeTe_0.5Se_0.5 films were obtained on several substrates. Figure 3 shows x-ray diffraction (XRD) patterns of one of the first high-quality FeTe_0.5Se_0.5 films deposited on STO by Si et al [34]. The ѳ–2ѳ scan shows only 00L peaks and the Φ scan suggest a four-fold symmetry with in-plane alignment between the film and substrate. The epitaxial growth of the film is confirmed by the high-resolution transmission electron microscopy (HRTEM) cross-sectional image shown in figure 4. Huang et al [35] reported on FeTe_0.5Se_0.5 films grown on MgO at various temperatures. They found that epitaxy is improved at higher deposition temperatures and a four-fold symmetry with narrow peaks is achieved at the maximum deposition temperature of 500 °C.

Figure 3.

(a) XRD ѳ–2ѳ scan of an FeSe_0.5Te_0.5 thin film on a (100) STO substrate. (b) Φ scan of the (112) peak from the FeSe_0.5Te_0.5 film and (c) from the STO substrate. Reproduced with permission from [34]. ©2009 American Institute of Physics.

Figure 4.

HRTEM image of an FeSe_0.5Te_0.5 thin film on a (100) STO substrate. Reproduced with permission from [34]. ©2009 American Institute of Physics.

Imai et al [36] systematically compared the properties of relatively thin (50 nm) FeTe_0.5Se_0.5 films grown on eight substrates and reported epitaxial growth on MgO, STO and LaAlO₃ (LAO) (figure 5). In a following report [37] they extended their research to non-oxide substrates like CaF₂, which resulted in the best superconducting properties (figure 6); SrF₂ and BaF₂ were found to be less effective than CaF₂ [38]. They state that the key to high-quality 11 thin films is the absence of interfacial reactions between the FeTeSe film and the substrate (figure 7). Mele et al [15] and Bellingeri et al [16, 39–41] confirmed that smooth, cube-on-cube epitaxial FeTe_0.5Se_0.5 films can be obtained on several substrates: MgO, LAO, STO, yttria-stabilized zirconia (YSZ), MgO, CaF₂ and LiF. Iida et al [14] reported for the first time the effective use of Fe buffer on MgO substrate. The Fe buffer grew at 45° to the MgO substrate, whereas the FeTeSe film showed the cube-on-cube epitaxial relation (figure 8).

Figure 5.

XRD in-plane patterns (Φ scans) of the 101 peak from FeSe_0.5Te_0.5 films on (a) MgO, (b) LAO, (c) STO, (d) LSAO and (e) YSZ substrates. Reproduced with permission from [36]. ©2010 Japan Society of Applied Physics.

Figure 6.

XRD Φ scans of FeSe_0.5Te_0.5 thin films grown on CaF₂(100): 115 reflection of the CaF₂ substrate (top) and 101 reflection of the FeSe_0.5Te_0.5 film (bottom). Reproduced with permission from [37]. ©2011 Japan Society of Applied Physics.

Figure 7.

HRTEM cross-sectional images of FeSe_0.5Te_0.5 thin films grown on (a) CaF₂ (100), (b) LAO (100), (c) MgO (100), (d) STO (100), (e) LSAO (001) and (f) YSZ (100) substrates. Reproduced with permission from [38]. ©2012 Japan Society of Applied Physics.

Figure 8.

(a) XRD ѳ–2ѳ scan of an FeSe_1−xTe_x bilayer in Bragg-Brentano geometry using Co-Kα radiation. (b) Rocking curve of the 001 reflection; Δω = 0.72° indicates good out-of-plane texture. (c) The 101 pole figure measurement and the corresponding Φ scans of the FeSe_1−xTe_x film. The respective Φ scans of the 110 Fe (buffer layer) and the 220 MgO (substrate) reflections are also presented. Reproduced with permission from [14]. ©2011 American Institute of Physics.

Speller et al [30] reported the first attempt to prepare FeTeSe films by molecular beam epitaxy (MBE) in a single-stage process on several substrates (MgO, STO, LaSrAlTa oxide (LSAT) and YSZ). In contrast to PLD samples, they found large compositional variations in some samples. Results of post-deposition annealing indicate that the ternary Fe(Se,Te) phase has a limited thermodynamic stability, transforming to Fe(Se,Te)₂ even at low temperatures. Epitaxy was reported only on MgO and STO substrates (figure 9), with the presence of a-axis grains rotated by 30° about the surface normal.

Figure 9.

(101) XRD Φ scan showing the in-plane alignment of a typical Fe_ySe_1−xTe_x film prepared by MBE. A scan of the substrate (111) MgO peaks is added for comparison. Reproduced with permission from [30]. ©2010 IOP Publishing.

FeSe_x films

Several reports on FeSe thin films appeared shortly after the discovery of superconductivity in the parent bulk compound. One of the first reports on epitaxial films was by Nie et al [42], who studied tensile strain induced by MgO, STO and LAO single-crystal substrates; c-axis growth with 00L peaks was obtained on all substrates.

Wang et al [43] deposited smooth tetragonal FeSe_1−x films on (001) MgO substrates by PLD. They studied the effect of low-temperature distortion on superconductivity and crystallinity. Films deposited at a low temperature of 320 °C were oriented along the c-axis. However, the (101) peaks were dominant in the films deposited at 600 °C.

Wu et al [44] presented an early report on the PLD of FeSe films on MgO. In XRD Ø scans they observed two kinds of domains with four-fold symmetry. The a-axes of major and minor domains were rotated by 45° and 0°, respectively, relatively to the MgO substrate orientation (figure 10).

Figure 10.

XRD Φ-scan of (101) and (203) peaks of an FeSe film. Two domains with four-fold symmetry are observed, which are 45° rotated to each other. The orientation of major domain is rotated by 45° from the a-axis of substrate. Reproduced with permission from [44]. ©2009 Elsevier.

Han et al [45] prepared FeSe_x (x = 0.80, 0.84, 0.88 and 0.92) thin films on STO, LSAT and LAO substrates by PLD. All films exhibited a single phase and c-axis oriented epitaxial growth. Increasing the Ar pressure reduced the film crystallinity, and the films deposited under P > 10 Pa were amorphous.

Since 2010, most research groups use PLD for growing FeTeSe films, which have higher T_c values than FeSe films. However, several FeSe films were grown by alternative techniques. One of the first examples is given by Jourdan and ten Haaf [28], who prepared FeSe by MBE and achieved the best results on YAlO₃ substrates cut in the (110) direction. Films grew epitaxially as confirmed by XRD patterns, and Ø scan demonstrates that in-plane (100) and (010) axes of the FeSe thin film were aligned parallel to the (1–10) and (001) axes of the YAlO₃ substrate (figure 11). Song et al [46] reported on MBE of stoichiometric and superconducting FeSe crystalline thin films on double-layer graphene that was formed on SiC(0001). It was shown by scanning tunneling microscopy (STM) that this new kind of substrate can yield strain-free FeSe films. Very recently, Demura et al [31] demonstrated the growth of 100 μm thick films on metallic tapes by electrochemical route. In their preliminary work, the composition ratio of Fe and Se was controlled by electric potential (figure 12) and pH value (figure 13).

Figure 11.

XRD Φ-scan of the off-specular (101) and equivalent peaks of an FeSe thin film on a YAlO₃ (110) substrate. Reproduced with permission from [28]. ©2010American Institute of Physics.

Figure 12.

XRD patterns of FeSe films electrodeposited at different electric potentials. Peaks marked by O, ∗ and ▴ indicate tetragonal FeSe, hexagonal FeSe, and hexagonal Se, respectively. Reproduced with permission from [31]. ©2012 Physical Society of Japan.

Figure 13.

XRD patterns of FeSe thin films electrodeposited at several pH values. Peaks denoted by O and ∗ indicate tetragonal FeSe and hexagonal FeSe, respectively. Reproduced with permission from [31]. ©2012 Physical Society of Japan.

FeTe films

Han et al [47] prepared epitaxial and superconducting FeTe films on LAO, LSAT, MgO and STO (figure 14). No other reports were published so far on thin films of FeTe, which is not a superconductor in the bulk form.

Figure 14.

XRD patterns of FeTe films deposited by PLD on (a) MgO, (b) STO, (c) LSAT and (d) LAO substrates. Reproduced with permission from [47]. ©2010 American Physical Society.

FeTe:O_x films

The only report on oxygen-enriched FeTe:O_x films was published in 2010 by Si et al [26]. XRD patterns show 00L peaks from the film and STO substrate, indicating a good out-of-plane alignment of the film. The Φ scan of the (112) peak from the thin film and substrate contains four peaks spaced by 90°, revealing a fourfold symmetry (figure 15).

Figure 15.

(a) XRD ѳ–2ѳ scan of an Fe_1.08Te:O_x thin film on a (100) STO substrate. (b) Φ scan of the (112) peak from the Fe_1.08Te:O_x film and (c) from the STO substrate. Reproduced with permission from [26]. ©2010 American Physical Society.

Superconducting critical temperature

The superconducting transition of iron chalcogenide thin films is strongly affected by several factors that are reviewed in this section: chemical substitutions, deviations from nominal stoichiometry, deposition temperature, laser frequency, number of laser pulses, film thickness, pressure and substrate quality. Strong differences (summarized in table 2) between the critical temperatures of the parent bulk compounds and films have been observed.

Table 2.

Summary of superconducting transition temperatures of polycrystalline (bulk) and thin films of iron chalcogenides. High-pressure data are reported for the bulk samples only. In the case of thin films, only the highest T_c for each kind of compound is reported. EDS stands for energy dispersive X-ray spectroscopy.

	Polycrystalline (bulk)						Thin film
Nominal stoichiometry	Ambient pressure		High pressure				Ambient pressure
	Tconset (K)	Tc0(K)	Tconset (K)	Tc0(K)	EDS composition	Ref.	Tconset (K)	Tc0 (K)	EDS composition	Substrate	Ref.
FeSe	12	8	37 (4 GPa)	20 (4 GPa)	N/A	[18]	12	8	FeSe0.91	STO	[50]
FeTe0.8S0.2	9	6.5	7 (1 GPa)	2 (1 GPa)	FeTe1.29S0.45	[18]	11	6.6	FeTe1.57S0.15	STO	[25]
FeTe0.5Se0.5	16	13.5	25 (2 GPa)	22 (2 GPa)	N/A	[18]	21	19	N/A	LAO	[16]
FeTe	0	0	N/A	N/A	N/A	[47]	13	9	N/A	LAO	[47]
FeTeOx	0	0	N/A	N/A	N/A	[26]	12	8	N/A	STO	[26]

Effect of chemical substitutions and deviations from stoichiometry

FeSe, the first discovered compound of the 11 family [8], has a zero-resistance transition temperature T_c⁰ = 8 K. From the early stages of research on polycrystalline materials it was made clear that substitution of Se by an isovalent element like Te enhances the superconducting transition temperature up to 16 K [38]. Later, it was found that doping at the Te site induces superconductivity in FeTe, which is not a superconductor in the bulk form [47]. The dependence of T_c on chemical substitutions in FeTe and FeSe binary compounds and the role of stoichiometry are discussed in this section.

FeTe_xS_1−x films.

The phase diagram of FeTe_xS_1−x [17] shows that the S substitution for Te suppresses magnetic ordering in FeTe and induces superconductivity with T_c⁰ = 6.5 K at low S doping levels (x = 0.8; figure 16).

Figure 16.

Phase diagram of FeTe_1−xS_x superconductor. Reproduced with permission from [17]. ©2010 Physical Society of Japan.

In one of the few reports on FeTe_0.8S_0.2 Mele et al [32] found that films deposited at 400 °C deviated from stoichiometry and had average compositions of FeTe_0.49S_0.13 and FeTe_0.83S_0.12 on STO and MgO, respectively. Consequently the T_c values were lower in these films than the stoichiometric bulk material: 3.54 K on STO and 5.37 K on MgO. Yoshida et al [25] described a similar deviation for films deposited with the YAG laser: Te and S contents were lower in the films than in the targets, and the average composition of all films was FeTe_0.58S_0.12. Using a target enriched in Te and S (composition FeTe_1.29S_0.45) they obtained a Te-rich film (composition FeTe_1.57S_0.15) with T_c⁰ = 6.6 K, which is close to the bulk T_c⁰. These results reveal that the volatility of S may severely hinder the enhancement of T_c in this system. This is the reason why the study of FeTe_xS_1−x films was almost abandoned.

FeTe_xSe_1−x films.

The phase diagram of FeTe_xSe_1−x [17] reveals that the tetragonal–orthorhombic structural transition observed in FeSe is suppressed with increasing Te concentration. The highest T_c appears in the tetragonal phase near x = 0.5. With a further increase of Te content, the T_c decreases, the antiferromagnetic ordering accompanying the tetragonal–monoclinic distortion emerges, and the bulk superconductivity disappears (figure 17).

Figure 17.

Phase diagram of FeTe_1−xSe_x superconductor. Reproduced with permission from [17]. ©2010 Physical Society of Japan.

Wu et al [44] reported a systematic study of Se doping effect in FeTe films. Their result shows that T_c⁰ increases with Te doping consistently with the phase diagram. However, it is not clear if the values reported refer to the nominal composition of the target or to the real composition of the film.

Screening of literature reveals contradictory results on the stoichiometry deviations in high-quality films prepared from a FeTe_0.5Se_0.5 target. Si et al [34] grew a 100 nm thick FeTe_0.5Se_0.5 thin film on STO with T_c^onset = 14 K. Bellingeri et al [40] reported even higher T_c^onset (21 K) for a 200 nm thick FeTe_0.5Se_0.5 film on LAO. Both authors state that there was no any deviation from the target stoichiometry. On the contrary, Mele et al [15] reported that the compositions of films grown on typical oxide substrates heavily deviate from the nominal composition: FeTe_0.08Se_0.99 on MgO, FeTe_0.14Se_0.94 on STO and FeTe_0.35Se_0.87 on LAO, all deposited at 450 °C. The deviations from stoichiometry explain the small values of T_c for these samples. Consistently, high-quality films were obtained on CaF₂ at lower deposition temperatures, so that deviations from stoichiometry were modest: FeTe_0.35Se_0.41 at 300 °C had T_c⁰ = 13.65 K (figure 18).

Figure 18.

Temperature dependence of the normalized resistance of FeTeSe thin films deposited on STO, MgO, LAO and CaF₂ single-crystal substrates at 450 °C with 45 000 laser pulses and 30 Hz repetition frequency. The inset shows details of the superconducting transition. Reproduced with permission from [15]. ©2012 IOP Publishing.

3.2. Effect of deposition temperature and number and frequency of laser pulses

The experimental parameters of PLD obviously affect the quality of the films and their superconducting properties. The majority of reports discuss the properties of the film obtained in the ‘best conditions’. Detailed discussions on the variation of film quality with experimental conditions were given only in the few cases reported here.

FeTe_xSe_1−x and FeSe films.

Wu et al [44] reported the thickness dependence of T_c for FeTe_xSe_1−x thin films (x = 0, 0.5). They found that the superconductivity of both FeSe and FeSe_0.5Te_0.5 films shows a strong thickness dependence: T_c⁰ reaches 8 K (true zero-resistance state) for FeSe_0.5Te_0.5 films thicker than 100 nm. In contrast, for FeSe thin films no true zero-resistance state exists unless the film is thicker than 300 nm (figures 19 and 20).

Figure 192.

The temperature dependence of the resistivity of FeSe_0.5Te_0.5 films. Zero resistance appears as film thickness increases to more than 100 nm. The inset shows the strong thickness dependence of onset T_c. Reproduced with permission from [44]. ©2009 Elsevier.

Figure 20.

Temperature dependence of the resistivity of FeSe films. Zero resistance appears as film thickness increases to more than 300 nm. The inset shows the strong thickness dependence of onset T_c. Reproduced with permission from [44]. ©2009 Elsevier.

Huang et al [35] analyzed the deposition temperature dependence of T_c in FeTe_0.5Se_0.5 samples deposited on MgO. Films deposited above 400 °C show semiconducting behavior, while those deposited at lower temperatures are metallic. The maximum value of T_c⁰ = 10.75 K is reached at 310 °C. Increasing the deposition temperature improves epitaxy, but degrades superconductivity and eventually produces semiconducting films (figure 21).

Figure 21.

Normalized resistance as a function of temperature for FeSe_0.5Te_0.5 films fabricated at various deposition temperatures (shown in °C in the inset). Note the different normalization values in the main figure and the inset. Reproduced with permission from [35]. ©2010 American Physical Society

Bellingeri et al [40] reported that the T_c of FeTe_0.5Se_0.5 films deposited on LAO reaches a maximum of 21 K (T_c^onset) at about 200 nm (figure 22).

Figure 22.

In-plane lattice constant a (a) and T_c (b) values versus film thickness for FeSe_0.5Te_0.5 films on LAO, STO and YSZ substrates: T_c reaches a maximum of 21 K for a thickness of ∼200 nm that corresponds to a minimum of a. (c) Resistivity vs temperature for films on LAO (thickness 9, 18, 36, 72, 150, 200, 280 and 420 nm): the room-temperature resistivity monotonically decreases with thickness. Inset shows details of the superconducting transition for some films. Reproduced with permission from [40]. ©2010American Institute of Physics.

Mele et al [15] systematically analyzed the dependence of T_c of FeTeSe/CaF₂ films on the experimental parameters, exploring a wide range of deposition temperatures (from 250 to 450 °C). They found that T_c⁰ increases with decreasing the deposition temperature, as shown in figure 23: T_c⁰ = 13.65 K for the sample deposited at 300 °C. Further improvements were obtained by lowering the deposition frequency from 30 to 3 Hz (figure 24). At 3 Hz, T_c⁰ = 16.18 K was higher than the bulk value.

Figure 23.

Resistance versus temperature for FeTeSe thin films fabricated at different temperatures by PLD on CaF₂ single-crystal substrates with 45 000 laser shots and 30 Hz repetition frequency. The inset shows details of the superconducting transition. Reproduced with permission from [15]. ©2012 IOP Publishing.

Figure 24.

Resistance versus temperature for FeTeSe thin films fabricated by PLD on CaF₂ single-crystal substrates at 300 °C at different pulse frequencies. The inset shows details of the superconducting transition. Reproduced with permission from [15]. ©2012 IOP Publishing.

Effect of lattice parameters and cell distortion

The anion height, i.e. the distance between the Fe layer and anion X (X = As, P, Se and Te), was suggested [17] as a universal parameter characterizing T_c values of all iron-based superconductors (figure 25). The maximum T_c is reached when anion height reaches 1.38 Å. Another proposed parameter [16, 48, 49] is the X–Fe–X angle α in Fe(X)₄ tetrahedra: T_c peaks when α approaches the ideal value of 109° 47′ (figure 26).

Figure 25.

(a) Anion height dependence of T_c in the typical Fe-based superconductors. Solid and open symbols indicate the data obtained at ambient pressure and under high pressure, respectively. (b) Schematic of the anion height from the Fe layer. Reproduced with permission from [17]. ©2010 Physical Society of Japan.

Figure 26.

T_c versus the tetrahedral bond angle α for FeSe_0.5Te_0.5 thin films and FeSe_xTe_1−x polycrystalline samples (red triangles and purple diamonds in the left part), in comparison with the literature data for various Fe-arsenides (blue squares in the center and right part). Inset: T_c versus ‘chalcogen height’ (that is, anion height) for FeSe_0.5Te_0.5 thin films, FeSe_xTe_1−x polycrystalline samples and intercalated FeSe (blue squares). Bracketed numbers correspond to the references in [16]. Reproduced with permission from [16]. ©2012 IOP Publishing.

According to these plots, iron chalcogenides possess too large values of anion height and too small values of α, and this explains why their T_c values are smaller than those of the other families of iron-based superconductors. If the anion height of 11 compounds is reduced and α is enlarged, T_c should increase. This means the c-axis should be compressed, and the a-axis expanded according to the Poisson effect. A simple way to achieve this effect is the application of uniaxial pressure along the c-axis. However, as reported in table 2, the result is controversial: T_c increases significantly with applied external pressure for FeSe_x and FeTe_0.5Se_0.5 bulk samples, but unexpectedly decreases for FeTe_0.8S_0.2.

Focusing on thin films, the lattice mismatch between film and substrate allows to modify the crystal parameters, simulating the application of high external pressures. Lattice mismatch is defined as M = 100(a₀ − a_s)/a_s, where a_s and a₀ are the lattice parameters of the substrate and bulk material, respectively. The crystal parameters of selected FeCh materials and typical substrates, together with the correspondent lattice mismatches, are summarized in table 3. From this table, the expected effects for 11 films grown on MgO, STO and CaF₂ are an expansion of a-axis and a compression of c-axis, with the concomitant T_c enhancement. On the other hand, there is almost no mismatch on LAO, and thus the crystal parameters and T_c should be almost unchanged.

Table 3.

Crystalline parameters of commonly used substrates and bulk iron chalcogenides, and the lattice mismatch between them.

		Lattice mismatch
Substrate	Crystalline parameter a or a/√2 (Å)	FeTe0.5Se0.5 (%) (a=3.7913 Å)	FeTe0.8S0.2 (%) (a=3.81 Å)	FeSex (%) (a=3.765 Å)
MgO	4.212	−11.08	−10.5	−11.9
SrTiO3	3.904	−2.89	−2.47	−3.56
LaAlO3	3.788	0.087	0.57	−0.61
CaF2	3.863a	−2.13	−1.62	−2.84

^aFor CaF₂, a/√2 is reported.

The systematic analysis of literature reported in the following subsections shows that the situation is much more complex and the general prediction from the Poisson effect cannot be applied.

FeTe_xS_1−x films.

Yoshida et al [25] reported that the shrinkage of c-axis with respect to the bulk value increases T_c^onset of PLD FeTe_1.57S_0.15 films prepared using a YAG laser. This result seems to confirm the general observation that that the reduction of the anion height has a beneficial effect on the superconducting properties of 11 films (figure 27).

Figure 27.

T_c^onset for FeTe_1.57S₀₁₅ thin films, fabricated by using FeTe_1.29S_0.45 target, as a function of the lattice parameter c. Reproduced with permission from [25]. ©2011 Elsevier.

FeTe_xSe_1−x films.

Bellingeri et al [16, 39–41] extensively analyzed the properties of FeTe_0.5Se_0.5 films deposited on MgO, YSZ, STO, CaF₂, LiF and LAO. They reported a biaxial compressive strain effect by LAO substrate on 11 films, with shrinkage of both the a- and c-axes. Maximum T_c^onset (21 K) was reported for the shortest a-axis of the film, while no dependence on the c-axis length was observed. Indeed, the 21 K film has α ∼ 104°, the closest to ideal one amongst the films prepared on LAO, as well as the shortest anion height (figure 28).

Figure 28.

(a) Transition temperatures of several FeSe_0.5Te_0.5 films on LAO, STO and YSZ substrates as a function of the in-plane cell parameter a. In (b) and (c), transition temperatures of films deposited on LAO are plotted as a function of the Fe–(Se,Te) bond length and of the (Se,Te)–Fe–(Se,Te) bond angle, respectively. The stars represent the corresponding bulk values. Dashed lines are guide to the eye. Reproduced with permission from [39]. ©2009 IOP Publishing.

Mele et al [15] reported that FeTe_0.5Se_0.5 films deposited on LAO had a compressed c-axis (c = 5.9293 versus 5.9784 Å in the bulk), but their T_c was lower than in the corresponding bulk material (T_c⁰ = 7.47 versus 13.5 K in the bulk). They also found that the c-axis of FeTe_0.5Se_0.5 films on MgO and STO were compressed (c = 5.9228 and 5.9116 Å, respectively), as expected, although the films exhibited lower T_c values. On the contrary, there is more clear correlation between T_c and c-axis in the case of CaF₂: c-axis is unexpectedly longer than 6 Å, and T_c is higher than in the bulk material. These results suggest the c-axis expansion and the resulting epitaxial strain along the a-axis play an important role in the T_c increase on CaF₂.

Imai et al [36] showed that a thin FeTe_0.5Se_0.5 film (50 nm) deposited on CaF₂ at 300 °C had the most elongated c-axis (5.968 Å) and the largest T_c⁰ (15.7 K). They claim that superconducting properties of the films are not affected by the a-axis and lattice mismatch, but by the ratio c/a and by the degree of the in-plane orientation. Speller et al [30] supported this idea and reported the T_c dependence on the c/a ratio for their MBE samples and several films and bulk materials reported in literature. They suggest that that T_c values are strongly affected by unit cell anisotropy, with an optimum c/a ratio of ∼1.6 corresponding to the highest T_c values of 14 K (figure 29). This is consistent with reports that T_c is very sensitive to anion height above the Fe plane in Fe-based superconductors.

Figure 29.

Relationship between c/a and onset T_c values in films deposited by low-temperature MBE (stars) compared with data from thin films and bulk FeTe_0.5Se_0.5 samples reported in literature. Reproduced with permission from [30]. ©2011 IOP Publishing.

Huang et al [35] reported that T_c of FeTe_0.5Se_0.5 films grown on MgO has a maximum for c ∼ 6 A. They suggested that rather than directly associated with the changes in c-axis, this variation of T_c may be linked to changes in the iron coordinating Se/Te tetrahedra, like the distance Δh_Se/Te. They also observed that the tetrahedral angle α does not significantly affect T_c of 11 films, while it is a crucial parameter for the other families of iron-based superconductors.

FeSe films.

In the early report by Nie et al [42], 50 nm thick FeSe films deposited on LAO were relaxed because their crystal parameters had almost the same values as in the parent bulk compound; the films had T_c^onset = 8 K, which is 4 K lower than in the bulk. Interestingly, strained films deposited on STO and MgO are not superconductors.

Later, Jung et al [50] reported the growth of epitaxial FeSe films on several substrates (STO, LAO, Al₂O₃ and MgO). Their best film (deposited on STO) exhibited almost the same critical temperature as the parent bulk material: T_c⁰ = 8 K and T_c^onset = 12 K (figure 30).

Figure 30.

(a) Resistivity versus temperatures for FeSe_1−x films grown on various substrates. (b) Enlarged view with the resistivity values normalized to those at 15 K. The inset shows the field-cooled (FC) and the zero-field-cooled (ZFC) measurements (H = 20 Oe) for an FeSe_1−x film grown on an STO substrate. Reproduced with permission from [50]. ©2010 Elsevier.

FeTe films.

Han et al [47] reported that the onset of superconductivity of 13 K has been introduced by interfacial stress, more specifically by the tensile stress, into the tetragonal non-superconducting FeTe compound. Corresponding to the maximum value of T_c, the tetrahedral angle α assumes a value of 95.4° (figure 31).

Figure 31.

(e) Lattice constants c for 32 FeTe films deposited on four different substrates. (f) Zero-resistance transition temperature T_c⁰ versus Fe–Te–Fe bond angle for four FeTe films indicated by arrows in (e). Inset schematically shows the definition of the angles 1 and 2, which are the Fe₄–Te–Fe₁ and Fe₄–Te–Fe₂ angles, respectively. Angle 2 is equivalent to the tetrahedral angle α mentioned elsewhere. Reproduced with permission from [47]. ©2010 American Physical Society.

Critical fields

Magnetic field/temperature phase diagrams of iron chalcogenide thin films reported in literature were plotted from magnetoresistance data obtained in a static field (H = 0–14 T), usually using a quantum design physical property measurement system (PPMS), or in a pulsed field (H = 0–85 T). In the first case, the values of upper critical fields H_c2(0 K) in the direction parallel and perpendicular to the c-axis are usually evaluated using the Werthamer–Helfand–Hohenberg (WHH) model [51]: H_c2(0) = −0.69T_c × (dH_c2/dT)_T=Tc. In the second case, direct extrapolation of H_c2(0) is possible. The values of coherence length ξ_c(0) are calculated from the Ginzburg–Landau relationship: ξ_c(0) = Φ₀/2π H_c2^∥(0)^1/2, where H_c2^∥ is the upper critical field in the direction parallel to the c-axis. Anisotropy parameter is defined as γ = H_c2^⊥/H_c2^∥. Table 4 summarizes H_c2(0) and other magnetic properties of 11 films reported so far.

Table 4.

Upper critical fields and related magnetic properties of 11 thin films deposited on various substrates.

Material	Substrate	Tc0 (K)	Hc2∥ (0) (T)	Hc2⊥(0) (T)	dHc2∥T (T/K)	dHc2⊥T (T/K)	ξc(0) (nm)	ξab(0) (nm)	γ	Reference
FeTe0.8S0.2	MgO	5.37	74.17	80.73	14.33	14.63	2.10	2.01	1.06	[32]
FeTe0.8S0.2	SrTiO3	3.54	62.09	N/A	14.06	N/A	2.30	N/A	N/A	[32]
FeTe0.5S0.5	MgO	12	72.5	N/A	N/A	N/A	N/A	N/A	N/A	[44]
FeTe0.5S0.5	SrTiO3	14	100	N/A	N/A	N/A	N/A	N/A	N/A	[34]
FeTe0.5S0.5	CaF2	15	150	N/A	N/A	N/A	N/A	N/A	N/A	[37]
FeTe0.5Se0.5	LAO	17.7	∼53	∼52	8.1	500	<0.12	1.5	∼1	[12]
FeSe	LAO		14	N/A	N/A	N/A	N/A	N/A	N/A	[45]
FeSe	STO		50	N/A	N/A	N/A	N/A	N/A	N/A	[50]
FeSe	YAlO3		9.2	25.7	N/A	N/A	N/A	N/A	N/A	[28]
FeTe	STO	13	123	N/A	N/A	N/A	N/A	N/A	N/A	[47]
FeTe:Ox	STO		200	N/A	23	N/A	N/A	N/A	N/A	[26]

FeTe_xS_1−x films

In the early report by Mele et al [32], an irreversibility field of H_irr = 9 T (∥c-axis) was observed at 1.94 and 3.75 K for FeTe_0.8S_0.2 films deposited on STO and MgO, respectively. Values of H_c2(0) estimated with the WHH model from 90% of the normal state resistivity, were H_c2^∥(0) = 62.1 T for the sample deposited on STO and 74.2 T for the sample deposited on MgO, and the respective ξ_c(0) values were 2.3 and 2.1 nm. In the case of MgO substrate, H_c2^⊥(0) = 80.7 T, and ξ_ab(0) = 2.01 nm. These FeTeS films were almost isotropic, with γ = 1.09.

FeTe_xSe_1−x films

Wu et al [44] have published one of the first reports on FeTeSe thin films. The upper critical field H_c2(0) of FeTe_0.5Se_0.5 on MgO was estimated at 72.5 T, which was about four times larger than that of the undoped FeSe sample.

Si et al [34] reported chalcogenide thin films with enhanced superconducting properties with respect to the corresponding bulk materials. They measured magnetoresistance in the H∥c configuration for three kinds of FeTe_0.5Se_0.5 samples: parent bulk, film deposited on STO with the same T_c as in the bulk, and film with T_c enhanced by the strain effect. H_c2 was consequently estimated from magnetoresistance curves applying the WHH model. Using T_c^onset values, H_c2(0) of all the samples was estimated at 100 T.

Tsukada et al [37] reported enhanced H_c2(0) = 120 T (evaluated using T_mid and the WHH theory) for FeTe_0.5Se_0.5 thin films deposited on CaF₂ (figure 32).

Figure 32.

(c) ρ vs T measured for FeTe_0.5Se_0.5 film in magnetic fields up to 14 T applied along the c-axis. The dashed line indicates the normal-state resistivity, and the dotted line indicates the half-rise of the normal resistivity that defines T_c^mid. (d) T_c^mid vs H_c2. Reproduced with permission from [37]. ©2011 Japan Society of Applied Physics.

Later, Tarantini et al [12] directly evaluated the upper critical fields of FeSe_0.5Te_0.5/LAO films in pulsed fields up to 85 T. They clarified that the shape of H_c2(T) curves for H∥ab in chalcogenide films is inconsistent with the conventional WHH theory: H_c2(0)^∥ was close to 53 T, i.e., about half of the values extrapolated from the WHH model; H_c2(0)^⊥ was slightly smaller so that the anisotropy parameter γ was ∼1. Furthermore, the upper critical field in the perpendicular configuration showed extraordinarily steep slopes of dH_c2/dT = 250–500 T K⁻¹ near T_c (figure 33). This behavior was attributed to the almost complete suppression of orbital pair breaking. Ginzburg–Landau coherence lengths were quite short: ξ_ab = 1.5 nm and ξ_c < 0.12 nm.

Figure 33.

(a), (b) H_c2(T) curves for two strained FeSe_1−xTe_x thin films with similar T_c measured in a 45 T dc magnet (triangles) and in a 60 T pulsed magnet (circles) (solid and open symbols correspond to H∥ab and H∥c, respectively). The black solid curves show H_c2 fits, and the red solid curves correspond to the Qvector of the Fulde–Ferrell–Larkin–Ovchinnikov (FFLO) state. The blue dashed line in (b) shows H_c2(T) calculated without taking FFLO instability into account. Fitting parameters are α = 1.2 and η = 0.9 for H∥c and αε^−1/2 = 7 for H∥ab. Here, Q is measured in the units of Q₀ = 4πk_BT_c/hv₁ε^3/4. (c) Superconducting transitions in low fields (H∥ab) for a FeSe_1−xTe_x film and a magnification close to 90% of the normal-state resistivity. (d) For the same sample, the detailed H_c2 data close to T_c show the extremely steep slope for H∥ab. The continuous lines correspond to the functions H_c2(T) ∝ (T_c − T)^1/2, indicating complete suppression of orbital pair breaking. Reproduced with permission from [12]. ©2011 American Physical Society.

Direct measurements in pulsed fields showed that the upper critical fields of 11 films can be pushed up to about 50 T by epitaxial strain and doping, overcoming typical values of parent bulk compounds that are of the order of 40 T [52–54]. The values of H_c2 (0) calculated by the WHH formula range from 14 to 100 T or more, but are generally much larger than the observed H_c2(0), suggesting the predominance of the Pauli paramagnetic effect. In any case, H_c2(0) is quite high considering the moderately low T_c in this system. For example, H_c2(0) for FeSe_1−xTe_x is almost twice as high as for Nb₃Sn, despite their same T_c = 18 K. These properties may be important for future applications of 11 compounds.

FeSe films

Han et al [45] used the half-rise of the R(T) curve at the superconducting transition to determine T_c for each field in FeSe thin films deposited on LAO. The upper critical field H_c2(0) can be estimated at 14 T according to the WHH model, as shown in the inset of figure 34. This value is close to that obtained from polycrystalline samples. Jung et al [50] prepared FeSe films with T_c^onset = 12 K on several substrates and estimated their H_c2 at 50 T using the WHH model and the values of 0.9 T_c^onset. Jourdan and ten Haaf evaluated the upper critical fields of FeSe films deposited by MBE on YAlO₃ substrates [28] in the field range 0–8 T in both orientations: they observed an almost linear dependence H_c2(T), similarly as in polycrystalline bulk samples. From the WHH theory, H_c2(0)^∥ = 9.2 T and H_c2(0) = 25.7 T, so that γ ∼ 3.

Figure 34.

(a) Dependences of T_c^onset and T_c⁰ on the Se content x for FeSe_x films deposited on LAO at 620 °C under 3 × 10⁻⁴ Pa. (b) Resistivity versus temperature in the external magnetic field for FeSe_0.88. The zero-temperature upper critical field is ∼14.0 T, as given in the inset of panel (b). Reproduced with permission from [45]. ©2009 IOP Publishing.

FeTe films

Han et al [47] introduced the onset of superconductivity at 13 K into the tetragonal non-superconducting FeTe compound by tensile stress on STO substrates. The upper critical field H_c2(0) was estimated by the WHH theory as 123 T, much higher than that of FeSe (figure 35).

Figure 35.

Influence of magnetic field on superconductivity for a 90 nm thick FeTe film on STO. Inset shows upper critical fields deduced from the midpoint transition temperatures with H_c2(0) extrapolated as 123.0 T. Reproduced with permission from [47]. ©2010 American Physical Society.

FeTe:O_x films

Si et al [26] reported that superconductivity (T_c⁰ = 8 K) can be induced in FeTe films deposited on STO by incorporation of oxygen into the Te site. By using the WHH model, H_c2(0) was estimated to be ∼200 T from the T_c onset of magnetoresistance curves (0–9 T). The linear fitting for H_c2(T) yielded dH_c2/dT = −23 T K⁻¹.

Critical currents and pinning forces

Fabrication of high-quality epitaxial films belonging to the 122 and 1111 families of iron-based superconductors was developed earlier than for films belonging to the 11 family. The initial difficulties related to the volatility of arsenic under the pulsed laser irradiation were overcome, and 1111 and 122 films were obtained with J_c values higher than in single crystals of the same composition [55, 56]. Despite their simpler structure and easier synthesis route in comparison with films belonging to other families, the quality improvement of epitaxial 11 films took a little longer time, and reports on high-J_c iron chalcogenides are still relatively few.

Table 5 summarizes J_c and global pinning force (F_p) values for the 11 films reported so far. Most of reports concern J_c of FeTe_0.5Se_0.5. No data are available on critical currents of FeSe or FeTe:O_x thin films.

Table 5.

Critical currents and pinning characteristics of FeTe_0.8S_0.2 and FeTe_0.5Se_0.5 thin films deposited on various substrates. Values of T_c⁰, self-field and in-field J_c (H∥c direction) and F_p^max were found in the main text of the references or evaluated from the J_c–H plots. Please check references: [26] did not use LAO or FeTe_0.5Se_0.5. Reference is [13] and second substrate is IBAD-buffered hastelloy, not STO.

Material	Substrate	Tc0 (K)	Jc (MA cm−2) H∥c			Type of pinning	Fpmax (GN m−3)	Reference
			Self-field (H=0)	Intermediate field (H = 5 T)	High field (H = 14 T)
FeTe0.8S0.2	MgO	5.37	0.04 @ 2 K	0.006 @ 2 K	N/A	N/A	N/A	[32]
FeTe0.8S0.2	SrTiO3	3.54	0.0001 @ 2 K	0.00002 @ 2 K	N/A	N/A	N/A	[32]
FeTe0.5Se0.5	CaF2	15.3	0.05 @ 12 K	0.03 @ 10 K	0.042 @4.2 K	N/A	N/A	[37]
FeTe0.5Se0.5	LaAlO3	15	0.4 @ 4.2 K	0.3 @4.2 K	0.2 @4.2 K	N/A	20 @ 4.2 K, 20 T	[13]
FeTe0.5Se0.5	SrTiO3 IBAD-buffered hastelloy	11	0.15 @ 4.2 K	0.1 @ 4.2 K	0.08 @4.2 K	N/A	8 @ 4.2 K, 20 T	[13]
FeTe0.5Se0.5	Fe-buffered MgO	17	0.5 @ 4.2 K	0.06 @ 4.2 K	N/A	Intrinsic (⊥c-axis)	N/A	[14]
FeTe0.5Se0.5	LaAlO3	19	0.12 @ 4 K	0.1 @ 4 K	N/A	Intrinsic (⊥c-axis)	N/A	[57]
FeTe0.5Se0.5	CaF2	16.18	0.41 @ 4.2 K	0.28 @ 4.2 K	N/A	c-axis correlated (∥c-axis)	22 @ 4.2 K, 9 T	[16]
FeTe0.5Se0.5	LaAlO3	19	0.12 @ 4 K	0.1 @ 4 K	N/A	Intrinsic (⊥c-axis)	N/A	[17]
FeTe0.5Se0.5	SrTiO3	15.1	0.4 @ 4 K	0.2 @ 4 K	N/A	c-axis correlated (∥c-axis)	N/A	[17]

Critical currents

FeTe_xS_1−x films.

In an early paper by Mele et al [32] on optimized FeTeS films deposited on MgO and STO the self-field J_c was reported as 4 × 10⁴ A cm⁻² for the film deposited on MgO. It was even lower for films deposited on STO under the same conditions, due to the formation of reactive interface (see section 2.1). The low J_c can be attributed to the low values of T_c (in both cases the T_c was lower than in the bulk material).

FeTe_xSe_1−x films.

A relatively high J_c value was first achieved by Tsukada et al [37] for FeTe_0.5Se_0.5 thin films deposited on CaF₂. They reported a very slow suppression of J_c above 10 T: at T = 4.5 K, J_c (10 T) = 5.9 × 10⁴ A cm⁻² and J_c (14 T) = 4.2 × 10⁴ A cm⁻² (figure 36). Si et al [13] reported the magnetic field dependence of J_c for FeTe_0.5Se_0.5 thin films deposited on LAO and flexible metallic substrates buffered with ion-beam assisted deposition (IBAD) (figure 37). On LAO, J_c (0 T, 4 K) was 5 × 10⁵ A cm⁻² and the field dependence was not pronounced. On IBAD-buffered substrates, J_c was 2 × 10⁵ A cm⁻² at 4 K and 0 T with almost similar in-field behavior. Iida et al [14] studied the effect of Fe buffer between MgO single crystals and FeTeSe thin film. They reported J_c (0 T, 4.2 K) =0.5 × 10⁶ A cm⁻². They found that for H∥c J_c values were quickly decreasing with increasing H, indicating less flux pinning in this direction, whereas the reduction of J_c with H was very small for H⊥c below 16 K (figure 38). Eisterer et al [57] reported on critical current densities of four FeTeSe thin films of different thickness prepared on LAO under similar deposition conditions; J_c at 4.5 K (B∥c) was ∼10⁵ A cm⁻². The same field dependence in all samples was observed at 10 and 15 K. More recently, Bellingeri et al [16] reported J_c plots for FeTe_0.5Se_0.5 films deposited on STO and LAO. The critical current density of two films behaved very differently: the film on LaAlO₃ had higher J_c for the field parallel than perpendicular to the ab planes, while the opposite held for the SrTiO₃ sample. The J_c values of the SrTiO₃ sample, at liquid He temperatures, stayed above 10⁵ A cm⁻² in the entire field range (figure 39). Similar results were reported by Mele et al [15] for FeTe_0.5Se_0.5 films grown on CaF₂: the J_c in zero field was 0.43 × 10⁶ A cm⁻² at T = 2 K and remained above 0.26 × 10⁶ A cm⁻² at 9 T (figure 40). Notably, the decrease of J_c did not accelerate much in high fields at 4.2 K (J_c = 0.41 × 10⁶ A cm⁻² at 0 T and 0.23 × 10⁶ A cm⁻² at 9 T). Recently, Higashikawa et al [58] reported a study of J_c in FeTe_0.5Se_0.5 thin films deposited on CaF₂ by means of scanning Hall probe microscopy. From the 2D-plane distribution of the current (figure 41), they estimated J_c at 4.2 K as ∼1 MA cm⁻², three times higher than the value directly measured by the four-probe method (figure 42). This fact suggests that FeTeSe thin films have a great potential for practical applications at low temperatures. Independently from the substrate, the common feature of all reports on FeTeSe thin films is a weak dependence of J_c on magnetic field at liquid helium temperatures. This property is beneficial for high-field magnet applications.

Figure 36.

(a) Current-density dependence of electric field in an FeTe_0.5Se_0.5 film. The maximum current available in PPMS is 5 mA corresponding to J ∼ 7 × 10⁴ A cm⁻². (b) Part of the same data plotted in the linear scale. (c) Applied-field dependence of J_c at T = 2.0, 4.5, 7.5, 10.0 and 12.0 K. Typical data for FeSe_0:5Te_0:5 single crystal, BaFe_1.8Co_0.2As₂ thin film and MgB₂ thin film are also plotted for comparison. Reproduced with permission from [37]. ©2011 Japan Society of Applied Physics.

Figure 37.

Critical current densities of FeSe_0.5Te_0.5 films on (a) LAO substrate and (b) flexible metallic tape (Hastelloy) buffered with IBAD at various temperatures with magnetic field parallel (solid symbols) and perpendicular (open symbols) to the ab plane (tape surface). Reproduced with permission from [13]. ©2011 American Institute of Physics.

Figure 38.

(a) J_c–H characteristics measured in H∥c and (b) H⊥c at several temperatures. Chain lines indicate a J_c criterion of 15 A cm⁻² for μ₀H_irr. Reproduced with permission from [14]. ©2011 American Institute of Physics.

Figure 39.

J_c versus magnetic field for samples deposited on LaAlO₃ (left panel) and on SrTiO₃ (right panel) for different temperatures (4 and 6 K for LaAlO₃ and 4 and 8 K for SrTiO₃—this means nearly the same T = T_c) and with H perpendicular (solid symbols) and parallel (open symbols) to the ab planes. Reproduced with permission from [16]. ©2012 IOP Publishing.

Figure 40.

Magnetic field dependence of critical current density at T = 2, 4.2 and 10 K, in the field direction B∥c, for FeTeSe thin films fabricated by PLD at 300 °C and 3 Hz on single-crystal CaF₂ substrates. Reproduced with permission from [15]. ©2012 IOP Publishing.

Figure 41.

Critical current distribution measured with a Hall probe on a 1 cm × 0.2 cm × 200 nm FeTe_0.5Se_0.5 thin film fabricated by PLD on CaF₂. Reproduced with permission from [58].

Figure 42.

Critical current versus temperature (B = 0 T) for 200 nm thick FeTe_0.5Se_0.5 thin film fabricated by PLD on CaF₂. J_c evaluated by transport measurements (PPMS) is compared with J_c calculated from Hall probe maps with two different electric field criteria (E_c). Reproduced with permission from [58].

Angular dependences of J_c were reported for several of the FeTeSe films described above. In the FeTeSe/Fe/MgO films of Iida et al [14] the angular-dependent critical current densities (J_c(H)) at 8 K in several magnetic fields (figure 43) always showed a peak at ѳ = 90° and 270° (H⊥c) owing to the intrinsic pinning. There is no evidence of c-axis correlated pinning (peak for H∥c). Bellingeri et al [16] and Mele et al [15] reported a completely different behavior for FeTe_0.5Se_0.5 thin films deposited on STO, LAO and CaF₂. On LAO, J_c monotonically increased with increasing angle, with no indication of c-axis-correlated defects. J_c was independent of the field orientation up to 50° and the peak for H∥ab was quite sharp: this is consistent with correlated pinning centers parallel to the film. In the case of the film deposited on SrTiO₃, pronounced anisotropy and a large peak at ѳ = 0° (c-axis peak) was observed at all temperatures and magnetic fields, indicating the presence of a strong pinning mechanism (figure 44). Similar behavior was observed on CaF₂ [15], although the c-axis correlated peak was very broad and appeared at B ≽ 3 T (figure 45). The effect on STO and CaF₂ resembles the one observed in 122 films [56] and in YBa₂Cu₃O_7−δ films with embedded nanocolumns [59]. Furthermore, the presence of nanorods parallel to c-axis of the film was confirmed by Bellingeri et al [16] using surface STM, but not by Mele et al [15] who employed cross-sectional TEM. In the paper by Eisterer et al [57], the angular dependence of J_c shows only the intrinsic peak (B⊥c) except for one sample that at 15 K shows the B∥c peak.

Figure 43.

Angular dependence of critical current density J_c(ѳ) of FeSe_1−xTe_x film measured in several magnetic fields at 8 K. Reproduced with permission from [14]. ©2011 American Institute of Physics.

Figure 44.

Angular dependence of the normalized critical current density at various temperatures and fields for FeTe_0.5Se_0.5 films deposited on LaAlO₃ and SrTiO₃. Reproduced with permission from [16]. ©2012 IOP Publishing.

Figure 45.

Angular dependence of critical current density J_c measured at H = 1, 3, 5 T and T = 2, 4.2 K on FeTeSe thin films fabricated by PLD at 300 °C and 3 Hz on CaF₂ single-crystal substrates. Ø is the angle between the c-axis of the sample and applied magnetic field B. Reproduced with permission from [15]. ©2012 IOP Publishing.

In summary, the reports on angular dependence of J_c are contradictory and further investigations are required to clarify whether c-axis correlated pinning may be induced by the substrate.

Global pinning force

5.2.1. FeTe_xSe_1−x films.

Calculation of global pinning force F_P = J_c × B in the range 0–30 T reported in figure 46 shows that F_p^max is about 20 GN m⁻³ at 4.2 K for films deposited on LAO [13]. Films deposited on CaF₂ show a similar trend up to 9 T, the maximum magnetic field available for measurements by Mele et al [15]. Furthermore, Si et al [13] reported (figure 46) an interesting comparison at 4.2 K between FeTeSe/STO films, YBaCu oxide (YBCO)-coated conductors and metallic superconducting wires. FeTeSe performed better than metallic superconductors in high fields, whereas YBCO was superior in the entire ranges of field and temperature. However, it must be kept in mind that YBCO has problems with grain boundaries and anisotropy that seems to be absent in FeTeSe films (see previous sections and other chapters of this special issue). The analysis of pinning force by the Kramer`s law (continuous lines in figure 46) reveals that the FeTeSe system can be explained invoking pinning by point-like defects. Since the coherence length of FeTeSe is short (1–2 nm), improvements in J_c and F_p are expected by controlled introduction of engineered defects at the nanoscale, as in ceramic (YBCO) and 122 iron-based superconductors.

Figure 46.

(a) J_c and (b) F_p at about 4.2 K of FeSe_0.5Te_0.5 films compared with the literature data of 2G YBCO wire, TCP Nb₄₇Ti and Nb₃Sn. For YBCO and FeSe_0.5Te_0.5 the field direction is parallel to the c-axis. Solid lines in (b) are Kramer's scaling approximations. Reproduced with permission from [13]. ©2011 American Institute of Physics.

Conclusions

In July 2008, the discovery of superconductivity in polycrystalline FeSe generated much interest in the scientific community, with numerous papers published in the same year on bulk materials and single crystals. At the beginning, the main purpose was to clarify the mechanism of superconductivity in presence of a ferromagnetic element. Furthermore, the early evidence that the upper critical field of FeSe and related iron-chalcogenide compounds is quite large for their low T_c favored them for practical applications. Consequently, since 2009 many research groups were attracted to the fabrication and characterization of iron chalcogenide thin films. To date, several high-quality films with excellent transport properties in magnetic field have been produced and tested.

A variety of attractive features of iron-chalcogenide thin films was described in this review and can be summarized as follows.

• (1)Epitaxial growth on common substrates. In comparison with compounds belonging to other families of iron-based superconductors, thin films of iron chalcogenides are relatively easy to grow by PLD. Cube-on-cube orientation and epitaxial growth are obtained on common single-crystal substrates (MgO, STO, LAO, CaF₂ and so on) under mild vacuum conditions (∼10⁻⁴ Pa) and at moderate deposition temperatures (300–500 °C).
• (2)Enhancement of T_c with respect to parent bulk compounds. Except for thin films of FeSe, which exhibits almost the same T_c⁰ ∼ 8 K in the bulk and films, iron chalcogenide thin films have higher maximum T_c values than their parent bulk compounds. FeTe_xS_1−x shows T_c⁰ = 6.6 versus 6.5 K in the bulk; FeTe_xSe_1−x films have T_c⁰ = 18 K on LAO and 16.15 K on CaF₂ (T_c⁰ ∼ 14 K in the bulk); FeSe films have T_c⁰ ∼ 6 K versus 4 K in the bulk; and, remarkably, FeTe films exhibit T_c⁰ = 9 K on LAO substrates, and FeTe:O_x films have T_c⁰ = 8 K on STO, while the parent bulk compound are not superconductors.
• (3) High and almost isotropic upper critical field. Direct evaluation in pulsed fields reveals that the upper critical field at 0 K is about 50 T in the case of FeTe_0.5Se₍₀₎₅, which is quite high for compounds with relatively small T_c. Consequently, iron chalcogenide films may find applications requiring high fields (20–50 T) and low temperatures (2–10 K).
• (4)High critical currents with low magnetic field dependence. In the recent papers, impressive values of 0.5–0.6 MA cm⁻² at 4.2 K, 0 T have been reported for the critical current density of FeTe_0.5Se_0.5. Furthermore, the dependence of J_c on magnetic field at liquid helium temperatures is weak, and this represents another advantage for in-field applications. Coherence length is on the order of few nanometers, and thus the in-field critical current can be enhanced by introduction of nanosized artificial pinning centers, as in the case of YBCO films.
• (5)Feasibility of thin films deposited on buffered metallic substrates. The feasibility of Fe buffers on single crystals for deposition of iron chalcogenide thin films has been demonstrated. Besides that, epitaxial films have been prepared on technical metallic substrates buffered with IBAD, without significantly reducing the critical current and upper critical field with respect to the films prepared on bare single crystals. A further progress in this direction will open the door for practical applications, similarly to what is happening now with the second generation of YBCO-coated conductors.

On the other hand, several issues remain open and are expected to be solved in the near future:

• (1)Control of stoichiometry. Chalcogenide elements (especially S and Se) are volatile and easily evaporate during the preparation of PLD targets. Evaporation may occur also in the PLD chamber during the fabrication of the films. The result is that the content of chalcogenides in the films is below the stoichiometric value, and therefore T_c may be lower than expected. Targets with stoichiometric excess of chalcogenides, and/or in-plume PLD might help obtaining films with the desired stoichiometry.
• (2)Factors that enhance T_c in films compared to bulk materials. It is believed that the enhancement of T_c in iron chalcogenide films with respect to the parent bulk compounds is related to strain induced by the substrate. However, the experimental behavior of the films deviates from simple Poisson effect, and it is not possible to predict the change in T_c from the lattice mismatch. To solve this puzzling issue, a systematic analysis of the variation of lattice mismatch, composition and thickness with experimental conditions is required for all types of iron chalcogenide films.