Abstract
Pre-fabrication of metallic nanorods on biaxially textured templates has been explored in this study to introduce flux pinning centers in RE–Ba–Cu–O (REBCO, RE =rare earth) based superconductors. Pt nanorods were deposited by an electron beam assisted deposition method on LaMnO3-capped biaxially textured IBAD-(ion beam assisted deposition) substrates. Well-controlled nanorods with varying diameter (50–120 nm), length (up to 1 μm), orientation and unit cell size were grown over an area of 120–150 μm2. The nanorod-decorated samples were then deposited with Gd–Y–Ba–Cu–O ((Gd, Y)BCO) by metal organic chemical vapor deposition (MOCVD). The Pt nanorods remain in their positions during MOCVD and become embedded in the (Gd, Y)BCO matrix, although they suffer creep-induced shape deformation due to exposure to elevated temperature. Higher unit cell size, longer nanorods, and nanorods oriented at an angle to the substrate normal adversely affect the epitaxy of the (Gd, Y)BCO film due to formation of a-axis grains. The observed current-carrying capacity of the Pt nanorod sample is lower than its corresponding reference sample without any nanorods and processed under identical conditions, but it decreases at a slower rate with increasing magnetic field. Potential routes to improve the performance while retaining the desirable characteristics of controlled nanorod direction and density are discussed.
Export citation and abstract BibTeX RIS
1. Introduction
High-temperature superconductors (HTS) are being developed for various electric and magnetic applications [1]. While HTS-based REBCO superconductors carry very high currents in the absence of a magnetic field, their application potential can be greatly enhanced by addressing the problem of the significant reduction in their current-carrying capacity in the presence of a magnetic field [2, 3]. The in-field performance can be significantly improved by creating pinning defects with sizes of the order of the flux lines. Many techniques have been used in the past to pin flux lines by defects such as edge and screw dislocations [4], twin planes [5], precipitates [6], and point defects created by high-energy irradiation [7]. A substantial amount of work in recent years has been focused on the introduction of artificial pinning defects by a self-assembly process such as BaZrO3 [8], BaSnO3 [9], BaIrO3 [10], BaHfO3 [11] and Ba2YNbO6 [12] defects. Commonly used techniques for the introduction of defects by the self-assembly process include MOCVD [13] and PLD [14]. These nanodefects self-assemble along the c-axis orientation and hence act as strong pinning centers for magnetic flux lines when the field is oriented in that direction (B ∥ c). While a significant improvement in critical current at B ∥ c can be achieved by this method, the critical current (Ic) of the superconducting tape is still limited by the minimum critical current at a field angle between B ∥ ab and B ∥ c directions, due to the strong anisotropy of the Ic versus B curve. Also, since these defects are created by a self-assembly process during the growth of the superconducting film, it is difficult to control the spacing and direction of the nanorods. For this reason, a method has been explored in our study where nanorods are pre-fabricated on the buffer layer, followed by MOCVD deposition of the superconducting film. With an appropriate nanorod growth method, the size, spacing, and orientation of the nanodefects can be controlled and made independent of the growth process of the superconducting film.
In this study, nanorods have been grown on biaxially textured templates followed by REBCO film deposition. Growth of nanostructures for various applications has been done in the past by many techniques, such as laser ablation [15], solution phase method [16], chemical vapor deposition [17], template-based method [18], electrodeposition [19], carbothermal reduction [20], and metal organic chemical vapor deposition [21]. The idea of introducing such a nanostructure prior to HTS film deposition has been explored to a limited extent, e.g., by decorating the substrate with nanostructures grown by different techniques [22–28], before REBCO layer deposition. These nanostructures explored were usually nanoparticles or nanoislands. Yang et al [29] have pre-fabricated long (more than 1 μm) MgO nanorods, with very high aspect ratio, by a chemical vapor deposition process for Bi–Sr–Ca–Cu–O (BSCCO) superconductors. Recently, Khatri et al [30] have shown pre-fabrication of Ni nanorods (height up to 2 μm, diameter 10–20 nm, and density 5 × 1010 cm−2) by an electrodeposition method on LaMnO3-buffered IBAD substrates for REBCO superconductors.
In this study, we have grown Pt nanorods with high aspect ratio on LaMnO3 (LMO)-buffered IBAD substrates. Pt nanorods were grown by an electron beam assisted method, which provides an excellent means of introducing excellently controlled nanostructures in terms of size, length and orientation. While this process is not scalable, it provides the opportunity to study the basic aspects of the pre-fabricated nanorod approach, such as the effect of nanorod unit cell size, rod diameter, length and orientation, as well as subsequent REBCO growth and its interaction with the nanorods. Nanorods of length up to 1 μm and diameter 50–120 nm have been grown with varying unit cell size and angle to the substrate normal. While the size of these nanorods is significantly larger than that of the BZO precipitates formed by the in situ self-assembly process (∼5–10 nm) and Pt is likely to be far from optimal material for the purpose, the method has been used to explore and demonstrate that long nanorods can be pre-fabricated in a controlled manner on biaxially textured substrates, survive the following REBCO layer deposition by MOCVD, remain embedded in the film and result in a good superconductor.
2. Experimental details
Platinum nanorods were deposited on a LMO-capped biaxially textured IBAD template on a flexible Hastelloy substrate [31]. The Pt nanorods were grown in a specific pattern using a dual beam (SEM/focused ion beam (FIB)) FEI 235 system equipped with a Pt gas injection system by injecting metal vapor in the vicinity of the desired nanorod locations. Metal vapor is generated by bombarding the Pt source with an electron beam at a pressure less than 2 × 10−5 Torr. A hexagonal pattern of nanorods with unit cell size ranging from 250 to 750 nm were grown with length up to 1 μm, over an area of 120–150 μm2, as specified in an input pattern Matlab file that is compatible with the FIB instrument software. The typical deposition rate was 0.015 μm3 min−1. The patterned area was marked by ion milling markers into the buffer/substrate using FIB, in order to identify the exact location of the nanorod pattern on the sample after (Gd, Y)BCO growth. The samples were then deposited with (Gd, Y)BCO in a reel-to-reel MOCVD process [32], Ag sputtered and annealed in O2. Transport current measurements were carried out at liquid nitrogen temperature (77 K) using a four-probe method and Ic values were determined using an 1 μV cm−1 criterion. In order to limit the measurement only to the nanorod-patterned region, a microbridge was made around the nanorod-containing area by ion milling using FIB. Reference samples without nanorods, made under identical conditions to the nanorod-patterned samples, were used for comparison. The Ic of the reference samples was also measured across a microbridge made with the same dimensions as the test samples. Samples for cross-sectional microstructure analysis by transmission electron microscopy (TEM) were prepared using FIB, and TEM studies were performed on a JEOL 2000FX microscope.
3. Results and discussions
3.1. Nanorods growth
Nanorod patterns were grown with different length, diameter and orientation with respect to the substrate surface. Examples are shown in figure 1. Figures 1(a) and (b) are SEM images of a nanorod pattern covering an area of 12 μm × 10 μm with a unit cell size of 600 nm and nanorods grown along the surface normal, while the micrograph in figure 1(c) shows a nanorod pattern with nanorods grown at an angle of 30° relative to the surface normal.
Figure 1. Pt nanorods deposited by electron beam assisted deposition in the FIB system: (a) 12 μm × 10 μm pattern of Pt nanorods grown along the substrate normal (top view), (b) nanorods grown along the substrate normal (60° angular view) and (c) nanorods grown at an of angle (30°) to the substrate normal (60° angular view).
Download figure:
Standard image High-resolution image3.2. Effect of BaZrO3 coating on Pt nanorods
In an attempt to provide a nanorod surface more compatible with REBCO and avoid any direct interaction between Pt and REBCO during MOCVD growth, the nanorods were coated with a thin film of BaZrO3 (BZO) by sputter deposition at 670 ° C. The corresponding SEM micrographs of BZO-coated nanorods are shown in figure 2. An initial concern was that the nanorods would not survive the sputtering process due to elevated temperatures, which could cause severe deformation and detachment from the buffer surface. It is evident from the figure that creep-induced deformation has indeed affected the shape of the nanorods. However, the nanorods do survive the coating process and still remain in their original location. Because of the bending of nanorods during sputtering, BZO coating is observed only on the curved portion of nanorods which is in line-of-sight relative to the sputter deposition direction. The extent and direction of bending is not uniform among the nanorods, which could potentially be exploited for more isotropic pinning due to the curved shape of the nanorods. It is also concluded that the sputtering process is not optimal for uniform and complete coating of nanorods due to the highly directional nature of the process.
Figure 2. Microstructure of Pt nanorods after BZO sputter deposition: (a) top view and (b) angular view (52°). Arrows indicate the location where BZO is deposited on the curved nanorods.
Download figure:
Standard image High-resolution image3.3. (Gd, Y)BCO deposition by MOCVD process
Samples with pre-fabricated Pt nanorods both with and without BZO coating were deposited with (Gd, Y)BCO film by MOCVD without any dopants such as Zr. The resulting texture of the (Gd, Y)BCO grown on nanorods without BZO coating is shown in figure 3. A unit cell dimension of 600 nm, nanorod length of 280–360 nm, and a diameter 75–80 nm were employed in the nanorod pattern in the sample. The thickness of (Gd, Y)BCO film was 750 nm. The formation of a-axis oriented platelets of (Gd, Y)BCO, which form a network interspersed with the c-axis oriented (Gd, Y)BCO matrix, is evident in the figure. The pronounced growth of a-axis oriented grains is likely a combination of incompatibility of Pt with the (Gd, Y)BCO matrix, as well as the relatively large size of nanorods in relation to the unit cell size. In order to further determine how the nanorod parameters affect the epitaxy of (Gd, Y)BCO film, and thus the formation of a-axis grains, samples with different nanorod length and the same unit cell size were compared for (Gd, Y)BCO film growth. It was found out that the longer the nanorods, the more difficult it is to grow c-axis aligned (Gd, Y)BCO between the nanorods by MOCVD and more a-axis grains are formed. A similar phenomenon has been observed by varying the unit cell size at a constant nanorod length, as shown in the microstructure of 900 nm thick (Gd, Y)BCO films in figure 4. It is seen from the figure that more a-axis grains are present in sample with nanorods of a unit cell size of 400 nm than in the sample with nanorods of 600 nm unit cell size. The lower amount of a-axis grains in the sample in figure 4(b) compared with that in figure 3 in spite of the same unit cell size of 600 nm in both is attributed to a short nanorod length of 250–300 nm and thicker 900 nm (Gd, Y)BCO film in the former. The effect of nanorod orientation on the a-axis grain growth was also examined. Samples with nanorods inclined to the surface normal showed poorer texture (more a-axis grain growth) than samples with nanorods grown parallel to surface normal. Despite the formation of a-axis grains, an encouraging finding is that a c-axis oriented matrix is still achieved in the form of a smooth film even in the presence of a relatively high volume fraction of nanorods, which demonstrates the potential for growth of MOCVD-REBCO film on pre-fabricated nanorods. With optimization of the type and size of nanorods, as well as the MOCVD process, a-axis-free growth of REBCO with a high density of pre-fabricated nanorods could be achieved.
Figure 3. Microstructure of (Gd, Y)BCO deposited on Pt nanorods pattern (600 nm unit cell): (a) top view and (b) magnified angular view (52°).
Download figure:
Standard image High-resolution image
Figure 4. Nanorod pattern area showing the effect of unit cell size on a-axis grain growth during (Gd, Y)BCO deposition: (a) unit cell 400 nm and (b) unit cell 600 nm.
Download figure:
Standard image High-resolution image3.4. Transport critical current measurements
Transport critical current measurements on a Pt nanorod sample and the corresponding reference sample were done by the four-probe method. A SEM image of a typical bridge made around the patterned area for these measurements is shown in figure 5. The in-field angular dependence of critical current was measured at applied magnetic fields of 0.23 and 1.0 T at 77 K. In-field measurement results of a Pt nanorod sample described in figure 3 are compared with the corresponding reference sample in figure 6. The critical current density (Jc) values of the samples with and without nanorods at 77 K in zero applied magnetic field were measured to be 1.63 MA cm−2 and 4.93 MA cm−2 respectively. The lower critical current density values in the Pt nanorod sample are believed to be due to the presence of a high volume fraction of a-axis oriented grains. However, the critical current density of the nanorod sample is found not to decrease rapidly when the magnetic field is moved away from B ∥ a–b, which is in contrast to the characteristic of the sample without nanorods.
Figure 5. Microbridge across the Pt nanorod-patterned area (angular view 52°) shown by two parallel lines between which current passes (sample surface seen here is Ag, sputtered after (Gd, Y)BCO film deposition). Markers used for locating the nanorod deposited area can also be seen in the figure.
Download figure:
Standard image High-resolution image
Figure 6. Angular dependence of critical current density (Jc) of the Pt nanorod sample (600 nm unit cell) and the corresponding reference sample at 77 K in magnetic fields of 0.23 and 1 T (R—reference sample, Pt—Pt nanorod sample).
Download figure:
Standard image High-resolution imageThe field dependence of the critical current density of the samples with and without nanorods at B ∥ ab and B ∥ c are shown in figure 7. From a power-law dependence of Jc on magnetic field B (J = kB−α), where k is a constant, α values of 0.85 and 0.66 were calculated for the reference sample and for the Pt nanorod sample respectively in the orientation of B ∥ c. α values of 0.62 and 0.36 were calculated for the reference sample and for the Pt nanorod sample respectively in the orientation of B ∥ ab. Therefore, while the critical current density of the Pt nanorod sample is lower than that of the reference sample, its rate of decrease in Jc with applied field is lower. Although the starting lattice spacing of the nanorods used in this study is quite large compared to the flux line spacing at 1 T, interfacial effects between the Pt nanorods and the film, as well as the reduced spacing of the nanorods due to their bending, could potentially influence pinning and contribute to the lower α values.
Figure 7. Magnetic field dependence of the critical current density (Jc) at different applied magnetic fields for the Pt nanorod sample (unit cell 600 nm) and the corresponding reference sample: (a) B ∥ c, (b) B ∥ ab (R—reference sample, Pt—Pt nanorod sample).
Download figure:
Standard image High-resolution image3.5. Top-down milling
One of the potential concerns about our approach is that even if the nanorods survive the BZO coating process, they might get detached or be otherwise compromised during the MOCVD process. While a-axis oriented grains were seen on the top surface of the nanorod sample (figure 3), they are not direct evidence of nanorod survival since they may have formed even if the nanorods disintegrated in shape to form a partially covered film on the buffer surface. For this reason, a post-deposition examination of the sample was done using continuous top-down milling of a (Gd, Y)BCO-deposited nanorod-containing sample (unit cell size 500 nm, inclination angle to surface 60°) using FIB. The findings are shown in figure 8. It is clearly evident that nanorods are still at their expected locations and hence are not detached or otherwise destroyed during the MOCVD process. With further milling, the (Gd, Y)BCO film and the nanorods are eventually milled away and the buffer layer below is exposed, as shown in figure 8(b). Even after complete milling of nanorods, the a-axis grain is still seen. This is likely due to a different rate of milling along different orientations by the Ga ion beam and because the a-axis grain seen in the figure appears to be nucleated on the buffer surface which is at the end of the milling process. The findings in figure 8 are direct evidence that Pt nanorods do survive the MOCVD process and remain in their expected location, which is encouraging evidence in support of the approach and an indication that more suitable nanorods, e.g., oxide nanorods with a smaller diameter, are likely to survive the process and potentially provide strong pinning.
Figure 8. Continuous milling by ion beam of a section of Pt nanorod deposited region: (a) Pt nanorods are clearly visible and (b) with further milling, nanorods are milled away and a-axis grains of (Gd, Y)BCO are still seen.
Download figure:
Standard image High-resolution imageA cross-section TEM micrograph of the interface between the LMO buffer and (Gd, Y)BCO film is shown in figure 9, revealing a section of an embedded Pt nanorod extending from the buffer into the (Gd, Y)BCO film. It is also seen from the figure that c-axis oriented (Gd, Y)BCO grows in the immediate vicinity of the nanorod, which does not shed light into the mechanism of formation of a-axis grains in samples with nanorods.
Figure 9. Cross-sectional microstructure of a nanorod sample examined by TEM showing a Pt nanorod (indicated by arrow) at an angle of 30° from the surface normal at the interface of LMO and (Gd, Y)BCO.
Download figure:
Standard image High-resolution image
It should, however, be noted that MOCVD was done at nominal processing conditions optimal to (Gd, Y)BCO films without nanorods, i.e., no attempt was made to study the effect of MOCVD processing conditions on the resulting microstructure and performance of nanorod-containing samples.
4. Conclusion
Metallic Pt nanorods with length up to 1 μm, diameter 50–120 nm, varying orientation with respect to the surface normal and different unit cell size were deposited successfully by electron beam assisted deposition on IBAD-based Hastelloy substrate. It has been found that nanorods that are longer, that have a smaller unit cell size and that are oriented at an angle to substrate normal adversely affect the epitaxy of (Gd, Y)BCO and a-axis grains are formed. Despite their high aspect ratio, nanorods survive the high temperatures associated with MOCVD with top-down milling showing that the nanorods remain in their location during the superconductor deposition process. (Gd, Y)BCO films with pre-fabricated Pt nanorods exhibit a Jc of 1.63 MA cm−2 (unit cell 600 nm, nanorod length 300–350 nm, and diameter 75–80 nm) at 77 K in zero applied magnetic field, which is lower than the Jc of 4.93 MA cm−2 of samples without nanorods and is likely due to the abundance of a-axis grains in the former. Lower values of α are found for the Pt nanorod sample compared to the reference sample, indicating that its Jc decreases at a slower rate with increasing magnetic field in both B ∥ c and B ∥ ab field orientations.
Acknowledgment
This work was funded by National Science Foundation grant no. CMMI-1000162.