The second generation high temperature superconducting (HTS) coated conductors are promising high performance practical materials that can be used at liquid nitrogen temperature (77 K). As shown in Fig. 1, the second generation superconducting wire consists of a multilayer thin film structure (substrate/buffer layer/superconducting layer/protective layer, etc.). Remarkable progress has been made in the fabrication of the second generation HTS coated conductors with excellent superconducting properties in the past several years [1–10]. Initially, researches were mainly focused on determining the characterizing material properties. The development of such practical materials has fueled global competition.

Fig. 1 The second wire structure consisting of a multilayer thin film structure (substrate/buffer layer/superconducting layer/protective layer, etc.)

Figure options

As shown in Fig. 2, various institutions or companies from all over the world have endeavored to promote the development of coated conductors. Coated conductors are being commercialized. However, a scale-up of the thickness is limited to about 2 μm owing to fundamental issues related to the thin film growth process using rolling assisted biaxially textured substrates (RABiTS) and ion beam assisted deposition (IBAD) [11]. The growth of a single layer REBCO film with high critical current I_c over 1 000 A/cm-width has been widely recognized to be very challenging owing to the thickness effect. To accelerate the use of REBCO HTS wires, their price should be reduced considerably by cost-effective, scalable fabrication of longlength, high I_c HTS wires.

Fig. 2 History of R&D of high-temperature superconducting wires (both wires have approximately the same cross-sectional area of 1 mm2) [12]

Figure options

The main obstacle toward achieving a higher current is that the critical current density J_c of the superconducting films drops precipitously as the film thickness is increased [13–15], as shown in Fig. 3. Superconducting films are required to carry 1 000 A/cm-width, which will require a coating thickness greater than 1 μm [16]. Hence, improvements are required in order to increase the commercial viability of HTS coated conductors. Although this limitation was first observed decades ago, the thickness effect was still one of the biggest challenges in the future.

Fig. 3 The critical current density of YBCO decreases sharply with the increase of film thickness [11, 16, 31]

Figure options

Regarding the investigation of the mechanism of thickness dependence of the critical current density in YBCO films, much work has been reported in recent years [17–25]. In 1999, Foltyn et al. [26] considered that the possible reasons for the thickness dependence of Jc were the impurity diffusion from the substrate to contaminate the YBCO layer nearest to the interface, microstructural evolution in the form of degradation of crystallinity, increasing surface roughness and micro porosity.

In 2002, Jia et al. [27] first demonstrated a multilayer architecture to enhance the supercurrent carrying capability of thick YBCO films. Degradation of crystalline quality, change in microstructures, and decrease in the number of vortex pinning centers have been proposed as possible reasons for the problem but a consensus has not been reached yet.

Recently, the geometrical self-field effect has been theoretically analyzed and determined to be an essential factor in the ubiquitous thickness dependence of J_c [28, 29]. It is often forgotten that the self-field effect unavoidably reduces J_c as the thickness of YBCO films rises, and thereby the total current increases. It was proposed that the critical current measurements should be carried out in an external field perpendicular to the film surface such that the external field exceeded the maximum value of the perpendicular component of self-field [30].

Nevertheless, the mechanism responsible for this decay is still not entirely understood. In this paper, we review some factors related to the thickness effect of the second generation superconducting films in detail and describe the recent progress in improving the performance of superconducting thin films.

In order to investigate the thickness effect, our group also performed fundamental research related to metal organic deposition (MOD) coating process parameters, particularly on the influence of different speeds of Czochralski process on the final film thickness. Figure 4 shows the relationship between the withdraw speed and thickness. It can be observed that, with the increase of withdraw speed, the final film thickness initially exhibits an increasing trend. However, as the withdraw speed increases beyond 70 mm/ min, with the film thickness increasing further, the final film thickness undergoes a declining trend. Similarly, the relationship between the density and viscosity is shown in green in Fig. 4. Moreover, we can also observe the relationship between the density and final film thickness of different buffer layers in Fig. 4, as shown with a red line. A theoretical model between the thickness t and withdrawal speed of substrate U was reported [32],

(1)

Fig. 4 Withdraw speed dependence of thickness and the relationship between density thickness (red line) and viscosity (green line)

Figure options

where α is the viscosity of the coating solution, ρ the density of the solution, and g is the gravitational acceleration. In this case, the exponent is 0.5, which is slightly larger than our case, probably owing to the effect of drying. The above relationship plays an extremely important role in achieving high current.

Furthermore, different heating rates and thicknesses of each layer have a great impact on the critical current density. From Fig. 5, it is evident that with the increase in thickness of each layer, the critical current density of the film shows a downward trend. When the heating rate is less than 5 ℃/min, the critical current density of the film shows a downward trend. When the heating rate is in the range of 5–25 ℃/min, the effect of heating rate on the critical current density is relatively small. Further increasing the heating rate will cause a decrease in the critical current density.

Fig. 5 The relationship between thickness per each coating and critical current density (black line) and the dependence of heating rate (red line)

Figure options

In order to obtain good transport characteristics, one of the main challenges is to maintain a well-aligned crystalline structure along the length of the superconducting film and reduce the number of film defects such as voids, cavities, and second phase as much as possible. In other words, the film must be highly homogeneous. As reported by many researchers, cuprate superconductors usually fall exponentially with the increase in film thickness, which results from the deterioration of microstructure and epitaxial texture.

As evident from Fig. 6, there are three important stages in microstructure evolution. The first stage is characterized by a rapid coarsening of both the grain and the grain boundary (the first 1.7 μm), which corresponds to the degradation of the in-plane texture. Therefore, it is possible that this rapid coarsening of grains and grain boundaries in the first few hundreds of nanometers is related to the initial drop in J_c.

Fig. 6 SEM images of the surface morphology variation with different film thicknesses (0.2–9 μm) (the scale is the same for all the images) [33]

Figure options

In the second stage (1.7–5.2 μm), the coarsening of both the grain and the grain boundary slows down, which corresponds to a plateau in the deterioration curve of the inplane texture. At a film thickness of approximately 5.2 μm, the third stage begins with a gradual microstructure transition from dendritic grains, which are well connected to equiaxial grains that may have weaker grain connectivity.

Furthermore, in this stage, the effective cross-sectional area for supercurrent flow declines significantly owing to the development and coarsening of grain boundary, and possibly an increased porosity. Consequently, the evolution of microstructure is one of the factors that lead to the decrease of the critical current density. The evolution of superconducting film microstructures cannot be ignored while considering the thickness effect.

Solovyov et al. [34] studied the grain size effect of superconducting films. As shown in Fig. 7, both in physical vapor deposition (PVD) and MOD, with the increase in grain size, the critical current density declines. However, the increase in film thickness is accompanied by an increase in the grain size. In other words, it is important to control the grain size in order to achieve high critical current density.

Fig. 7 Grain size dependence of critical current density for superconducting films [34]

Figure options

As evident in Fig. 8, the root mean square (RMS) value of surface roughness of the 200 nm-thick film obtained by pulsed laser deposition (PLD) is 23.6 nm. As the film thickness increases to 1 030 nm, the RMS value of surface roughness is 64.6 nm. For further increase of the film thickness to 1 450 nm, there is a slight increase in the RMS value of surface roughness from 64.6 nm to 68.7 nm. The increase in surface roughness will result in the growth of a-axis grains. In other words, surface roughness is also an important factor resulting in the degradation of critical current density.

Fig. 8 Surface roughness of GdBCO films with various thicknesses [35]

Figure options

The O2+/O3-mode for YBCO films is subject to c-axis orientation and it should be strong enough for a highquality epitaxial YBCO film with a perfect orthorhombic lattice. The O4 mode for YBCO films is subject to a-axis oriented grains.

As we can see from Fig. 9, the peak intensity of the O2+/O3+ mode enhances as the thickness increases. In the case of thinner samples such as YBCO (80 nm) and YBCO (320 nm), the O2+/O3+ mode peaks are nearly absent, whereas they are distinctly perceived in thicker films, such as YBCO (2 000 nm). In other words, with the increase in film thickness, the percentage of a-axis oriented grains increases, resulting in the decrease of critical current density.

Fig. 9 Raman spectra and characteristic phonon modes for the studied samples [36]

Figure options

Residual stresses will inevitably be induced in a thin film owing to the structural variation and thermal misfit between the thin film and substrate. The structure and properties of films are, therefore, dependent on the residual stress to some degree. Many studies have shown that the residual stress may cause the film to delaminate from the substrate. The emergence of tensile stress will cause the degradation of films, resulting in microcracks, holes, and a-axis grains. It is vital to study the multilayer structure in order to reduce the degradation of the critical current density owing to increasing film thickness.

The Williamson Hall method is an old and effective method to analyze the film internal strain ε using X-ray diffraction (XRD) [37]. The relationship of the internal strain ε and the integral breadth β of each (0 0 L) peak of the GdBCO film is expressed as

(2)

where θ is the Bragg angle position of each (0 0 L) peak and λ is the value of X-ray wavelength (λ = 0.154 18 nm).

Figure 10a shows β²cos²θ variation as a function of sin²θ for the GdBCO film with different thicknesses. Using the linear fit slopes obtained in Fig. 10a, the residual stresses calculated using Eq. (1) are 0.101, 0.076, 0.086, and 0.091 for the four GdBCO films, respectively. The corresponding film thicknesses are 200 nm, 1 030 nm, 1 450 nm, and 2 100 nm, respectively.

Fig. 10 a Williamson-Hall plot for GdBCO films with different thicknesses and b residual stress in YBCO films with different thicknesses [38–41]

Figure options

It is concluded that the thinnest film has the highest residual stress whereas the 1 030 nm-thick film has the lowest residual stress. With further increase in the film thickness, the film residual stresses increase.

In Fig. 10b, monolayer YBCO films thinner than t₀ (approximately 1 μm) were under compressive stress [39–41], which would be relieved by the increase in film thickness. There is a tendency for the compressive stress to reduce with the increase in the thickness. For films in the range of 0.5 μm—t₀ (approximately 1.0 μm), YBCO thin films have the lowest residual stress, resulting in a nearly free-standing film. As the film thickness is increased beyond t₀, tensile stress appears in YBCO films. With the accumulation of tensile stress, cracks, wrinkles, and a-axis emerge, thus affecting the quality of the film. The above analysis is in accordance with Fig. 10a. Consequently, we should manipulate the residual stress in a state of compressive stress in order to reduce the thickness effect.

The 3D strong pinning is produced when vortex lines are chopped into short discontinuous segments by strong pinning defects, which enforce depinning by the bending of vortex segments into critically sized vortex loops, whose diameter is equal to the separation d between pins. The central part of Fig. 11 shows the direction of magnetic field (B), critical current (J) and lorentz force (L), respectively.

Fig. 11 a Illustration of the 3D strong pinning and b 2D collective pinning

Figure options

Thus, J_c is determined by the pin spacing d and is independent of the film thickness t when d t. This case corresponds to the present MOD films. The correlation length parallel with the field direction is denoted by L_c. As shown in Fig. 11b, continuous vortex lines in the 2D collective pinning case occur in a dense array of weaker pins, which leads to both an inverse square root dependence on the thickness and a lower J_c, as observed in many typically made PLD films. Further, the behavior of J_c (t) can be estimated by the 2D collective pinning theory for a rigid vortex line in a thin film where the bending distortions of the vortex are negligible [42].

It is believed that the decline of pinning force is also an important issue in the investigation of the thickness effect. In order to reduce the thickness effect, artificial pinning centers should be introduced in thick superconducting films.

The O1s spectra consist of two peaks. The main peak at binding energy E_B = 528–528.5 eV is related to the O characteristic of the CuO₂ plane and CuO chains in superconducting REBCO [43]. The small peak at E_B = 530–530.5 eV may be associated with non-superconducting phases [44].

As shown in Fig. 12, the intensity of the two peaks decreases with an increase in the film thickness from 200 nm to 2 100 nm, which indicates that there is less oxygen content for the upper layer of the thicker film compared to the thinner films. Moreover, the curve integral area of the four samples decreases as the film thickness increases from 200 nm to 2 100 nm. This demonstrates that there is less oxygen content for the upper layers of the thicker film. The two trends are not obvious for the 200 nm-thick film and the 1 030 nm-thick film.

Fig. 12 The O1s spectra measured for GdBCO films with different thicknesses [35]

Figure options

However, when the film thickness increases to 1 450 nm, the two trends become obvious. The above analysis implies that the oxygen content is insufficient for the upper layers of the thicker film, especially for a film thicker than 1 030 nm. The oxygen content is different for layers of different depths in a thick film. The oxygen content is almost unchanged for the bottom layer from 0 to approximately 1 030 nm, which is demonstrated in Fig. 12. However, the oxygen content reduces for the upper layers from 1 030 nm to 2 100 nm. The oxygen content reduction is accompanied with the emergence of a-axis grains in the upper layers of the thick film. It indicates that the oxygen deficiency in the upper layers of films beyond a thickness of 1 030 nm mainly results from the emergence of a-axis grains.

The conventional path for increasing the thickness of MOD-based YBCO films has been the deposition and decomposition of multiple thin layers, which are subsequently reacted as a single layer. The individual layers can be prepared with or without rare earth nanoparticle pinning to tune the infield anisotropy [45].

This process has been used at American Superconductor (AMSC) for growing a 2 μm (nominal thickness) film, produced by three individual coating and decomposition steps, with I_c exceeding 500 A/cm-width (77 K, self-field). However, Fig. 13 shows the presence of a copper-rich region between the individual layers.

Fig. 13 TEM micrograph (left) of a triple-coat (0.8 μm/0.6 μm/ 0.6 μm) precursor film after decomposition (EDAX map on the right shows the formation of a Cu-rich layer at the interface between the individual layers. The top two layers do not contain Dy) [46]

Figure options

Through-thickness I_c characterization of fully processed films revealed that J_c of the film above the interface was lower than that below the interface. Through-thickness X-ray texture analysis of the film determined that the decrease in J_c was associated with a reduced YBCO texture in the region above the interface, suggesting that the epitaxial growth of YBCO was partially disrupted as it crossed the copper-rich interface, which was shown in Fig. 13.

Although it has been demonstrated that high throughthickness J_c can be obtained in thick YBCO films (approximately 2 μm) built up from very thin ( < 0.2 μm) layers [47, 48], this method is not consistent with a highrate, low-cost manufacturing process. Copper-rich layers and segregation of elements are also factors resulting in the decrease of critical current density.

Extensive reports on improving superconducting performance via various methods have been published recently. For instance, MacManus-Driscoll et al. [49] performed research linked to rare earth ion size and determined that superior critical current density could be obtained in mixed RE-123 films. Moreover, a multilayer structure such as YBCO/CeO₂ is effective in weakening the thickness effect. This multilayer approach makes it possible to fabricate very high critical current coated conductors without being limited by the rapid decrease in J_c for thicker YBCO single-layer films. Finally, co-evaporation process and inclined substrate deposition (ISD) process also obtained good results. 5 μm-thick SmBCO films with the highest I_c value of 1 540 A/cm-width (77 K, self-field) and 8 μm-thick SmBCO films with the highest I_c value of 1 100 A/ cm can be achieved via co-evaporation process and ISD process, respectively.

In order to achieve high critical current density, many studies have been performed on rare earth ion size [49]. Table 1 shows the results of the samples prepared in this study. The film with the highest self-field J_c (between 3.7 MA/cm² and 4.7 MA/cm²) was composed of YSmBCO.

Table 1 Mixed RE-123 films of the same variance (σ2 = (6–8) × 10^-6 nm²) but different average ionic radii

Table options

As evident from Fig. 14, the improvement of J_c in decorated samples over a range of thickness is the evidence that Ag is effective at creating pinning centers not only in thin films but also in thick films.

Fig. 14 Thickness dependence of the critical current density for films grown on substrates decorated with 15 pulses on Ag, at different applied magnetic fields from 0 to 2 T [50]

Figure options

As evident from Fig. 15, remarkably high critical current values of 3 963 A/12 mm and 2 797 A/12 mm have been achieved at 3 T and 5 T, respectively, in the 2.2 μm-thick 20% Zr (mole fraction)-added film. These values are more than twice the critical current values obtained for the 0.9 μm-thick 15% Zr (mole fraction)-added film. Hence, heavy doping can improve the superconducting performance of thin films, especially when used in a strong field.

Fig. 15 Magnetic field dependence of critical currents of a 2.2 μm GdYBCO tape with 20% Zr (mole fraction) addition and the best 0.9 μm thick (Gd, Y)BCO tapes with 7.5%, 15% and 25% Zr addition (mole fraction) at 30 K in the orientation of the field along the c-axis [51]

Figure options

In order to decrease the deterioration of epitaxial texture and microstructure, multilayer structures are introduced in thick superconducting films. For instance, the structure of YBCO/GdBCO consisting of three units of YBCO/GdBCO was obtained via PLD.

Figure 16a shows the structure model of YBCO/GdBCO films. The thickness of each layer is approximately 0.5 μm and YBCO/GdBCO is treated as a unit with a thickness of 1.0 μm. Contrary to a monolayer film, the cross-section of a multilayer film is dense and devoid of a large number of defects, which indicates that residual stress is released periodically by intermittent multilayer deposition method.

Fig. 16 a The structure model of YBCO/GdBCO films [37] and b the structure model of YBCO/CeO2 [27]

Figure options

Figure 16b shows the structure of YBCO/CeO₂. This multilayer approach makes it possible to fabricate very high critical current coated conductors without being limited by the rapid decrease in J_c for thicker YBCO singlelayer films. By an interlayer of CeO₂ between the adjacent YBCO layers, the structural defects induced during the increase of thickness of the YBCO layer reduce. This interlayer also provides a starting template for the subsequent growth of a high quality YBCO layer.

As shown in Fig. 17, deposition parameters can be uniformly controlled in a large chamber. Consequently, kilometer-long HTS wires can be fabricated by merely increasing the drum size with significant modification of its basic structure. Use of inexpensive metal elements as evaporation sources can be an additional advantage, reducing the total processing cost of the wires.

Fig. 17 Schematic outline of evaporation using drum in dual chamber (EDDC) process with the gradient technique (a EDDC apparatus consisting of an evaporation chamber, shield plate, and reaction chamber, b thickness distribution of each element along the substrate tape length, c and d slit geometry at the shield plate and the arrangement of substrate tape on the drum) [52]

Figure options

Using this method, we can obtain 12 cm-long, 5 μm-thick SmBCO films with the highest I_c value of 1 540 A/ cm-width (77 K, self-field), which corresponds to J_c of 3 MA/cm². The aforementioned value of I_c is the highest ever reported for any length of HTS film on either single crystal or metallic substrates. Consequently, the thickness effect declines dramatically and high J_c is maintained even as the film thickness approaches 5 μm.

As we can see from Fig. 17, deposition parameters influencing the composition, phase, and texture of the films were optimized via a unique combinatorial method that is broadly applicable to co-evaporation of other promising complex materials containing several cations. Unlike previous reports, no exponential decay of I_c with increasing thickness is observed in this study for samples fabricated under optimized conditions. The lack of any degradation in the structural properties with increasing thickness plays a critical role in maintaining high J_c, and consequently high I_c in these thick films. The EDDC technique and the combinatorial method prove invaluable in the fabrication of other complex, high-performance, and functional materials [52].

We can see clearly from Fig. 18 that the critical current density of DyBCO films with a thickness 6 μm is close to that of thinner DBCO films with a thickness of 2 μm, whose J_c ranges from 1.6 MA/cm² to 2.6 MA/cm². This indicates that the critical current density does not decline significantly with an increase in thickness. In other words, the thickness effect is weakened.

Fig. 18 Critical current per centimeter tape width versus DyBaCuO film thickness of ISD coated conductors (the red asterisk indicates the sample investigated by SEM and TEM) [53]

Figure options

ISD yields a non-zero component of the growth direction parallel to the DyBCO (a, b)-plane since DyBCO grows on a faceted MgO surface and avoids the a-axis growth. Highly biaxially textured DyBCO films at all thicknesses can be obtained; thus, no significant decrease in the critical current density occurs with the increase in film thickness. Therefore, the ISD technology offers a unique possibility to overcome thickness limitations in coated conductor technology.

The thickness dependence of J_c in the second generation superconducting films is reviewed in the present paper. It is believed that the phenomenon of thickness effect is caused by a series of factors such as the emergence of a-axis oriented grains, oxygen deficiency, evolution of microstructure, segregation of elements, residual stress, surface roughness, and decline of pinning force, etc. Recent progress in enhancing the superconducting performance of thin films especially in terms of critical current density is summarized. Further, with the increase in film thickness, more artificial defects are required as effective pinning centers for a higher current density. It is concluded that the artificial multilayer architecture is very helpful in maintaining the epitaxial texture of REBCO films, which allows an integrated understanding of the thickness effect.