- Email: [email protected]
Ni-15% Cr composite substrate for coated conductor application
Acta Materialia 51 (2003) 4919–4927 www.actamat-journals.com
Development of high strength and strongly cube textured Ni-4.5% W/Ni-15% Cr composite substrate for coated conductor application V. Subramanya Sarma ∗, J. Eickemeyer, A. Singh, L. Schultz, B. Holzapfel Leibniz Institute for Solid State and Materials Research (IFW), Helmholtz strasse 20, 01069, Dresden, Germany Received 30 April 2003; received in revised form 10 June 2003; accepted 16 June 2003
Abstract The development of thin, mechanically stronger and highly cube textured substrates is of great technological importance for increasing the engineering current density of the coated conductors. Nickel is a suitable substrate for this in view of its ability to form strong cube texture after heavy rolling and annealing and its excellent oxidation resistance. However, nickel is very soft (yield strength ~40 MPa) and this limits the processing to thin tapes. The ferromagnetism of Ni is also undesirable for ac application of coated conductors in magnetic fields. In the present paper we report on the development of Ni-4.5 at.% W/Ni-15 at.% Cr composite substrates of 80 and 40 µm thickness with strong cube texture, high yield strength (~200 MPa) and reduced magnetisation losses. The strong cube texture was obtained through an optimised two-step recrystallisation annealing following heavy cold working. It was found that the presence of noncube texture forming alloy (Ni-15% Cr) in the inner core of the composite had no adverse affect on the growth of cube textured grains on the surface (Ni-4.5% W) even at a low substrate thickness of 40 µm. A significant improvement in the texture/misorientation distribution was observed in the CeO2 buffer layer deposited on the composite substrate. 2003 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. Keywords: Coated conductors; Recrystallisation; Cube texture; Ni-alloys; Composite
1. Introduction The Rolling Assisted Biaxially Textured Substrates (RABiTS) method of Ni and Ni-based alloys (NiCu, NiFe, NiV, NiCr, NiW) is a very promising approach for the profitable production of long lengths of high temperature superconducting ∗ Corresponding author. Tel.: +49-351-4659203, fax: +49351-4659320. E-mail address: [email protected] (V. Subramanya Sarma).
(HTS) YBa2Cu3O7⫺d (YBCO) tapes capable of carrying high currents in magnetic fields at 77 K [1–28]. In this method, the desired strong biaxial texture in the superconducting YBCO film is achieved by epitaxial growth of a buffer and YBCO film on a highly textured substrate. Initially it has been shown that Ni is ideally suited as a substrate material due to its ability to form strong a cube texture after heavy cold rolling and recrystallisation. In addition its oxidation resistance and the small lattice mismatch allows epitaxial growth of buffer (for e.g. CeO2 + yttria stabilized zirconia
1359-6454/$30.00 2003 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. doi:10.1016/S1359-6454(03)00334-3
4920
V. Subramanya Sarma et al. / Acta Materialia 51 (2003) 4919–4927
(YSZ)) and YBCO films. In YBCO films on Ni substrates, current densities exceeding 106 Acm⫺2 have been achieved [1–3]. To achieve a high engineering current density (i.e., the current in the HTS film divided by the cross section of the whole tape) very thin tapes are desirable. But the low tensile strength of Ni limits the possibility of producing very thin HTS tapes by a continuous reel-toreel deposition process. For these reasons the strengthening of Ni substrate material is important. The need for a strong cube texture limits the possibilities of strengthening the substrate. Further, the ferromagnetism of Ni leads to hysteresis losses in coated conductors in alternating current (AC) applications [20], but it is not of great importance for direct current (DC) applications. There are two known metallurgical routes of strengthening the recrystallised substrate material i.e., solid solution and precipitation strengthening. It was shown that the substrate strength could be increased by factor 3 with substantial additions of Cr, V or W (ⱖ5 at.%, compositions are given in at.% throughout the paper) without losing the strong cube texture [8–11]. The additions (⬎10%) of Cr and V also suppress the Curie temperature to below 77 K, thus reducing the magnetization losses in the coated conductors [8,20]. However, the ease of Cr/V oxide formation on the surface at the buffer and YBCO deposition temperatures (these are prepared in oxidizing atmospheres) makes the epitaxial growth difficult. Higher strength levels can also be achieved by precipitation of Al2O3 particles through controlled internal oxidation of a Ni-1% Al alloy [29,33]. However, the formation of Al2O3 particles on the surface degrades its quality rendering it unsuitable for further epitaxial deposition. These surface oxidation problem can be overcome by the design of a composite substrate [30–33]. Using such an approach we recently reported the development of Ni-3% W/Ni-10% Cr-1.5% Al composite substrate of 80 µm thickness with high strength (room temperature yield strength ~175 MPa) and strong cube texture [32]. In this investigation, a cube texture forming alloy (Ni-10% Cr1.5% Al) was used in the core of the composite [32]. In the present paper, we report on the development of strongly cube textured and high strength Ni-4.5% W (outer sleeve)/Ni-15% Cr (a non-cube
texture forming alloy as the core) composite substrate of 80 and 40 µm thickness. Also we present the magnetisation behaviour of the composite tape at 77 K and texture results in the CeO2 buffer layer grown epitaxially by pulsed laser deposition (PLD) technique. 1.1. Design of the composite tapes The critical issues in the design of the composite tapes are discussed in a recent paper [32]. For brevity some important points are recalled here. It is known that the recrystallisation cube texture in Ni and Ni alloys is significantly improved with a fine starting grain size. This can be achieved through an intermediate recrystallisation annealing before the application of heavy (⬎95%) cold rolling [21]. Therefore, the problem with using pure Ni in the outer layer of the composite is with regard to achieving the required strong cube texture due to the wide differences in the recrystallisation temperatures between Ni (typically around 300–400 °C) and Ni-Cr, Ni-V alloys (typically 700–800 °C depending on the amount of alloy content). It should be mentioned that a coarse starting grain size has a strong detrimental effect on the final cube texture. Therefore during intermediate annealing of a composite with pure Ni outer layer and highly alloyed core would lead either to fine grain size in Ni and no recrystallisation in the core or coarse grained outer layer with recrystallised core. These microstructures and all in between are not useful to develop the very strong cube textures needed for the RABiTS application [33]. It is thus necessary to have the outer surface and the inner core materials with similar recrystallisation temperatures. Also of importance during the rolling of composite is the strength difference between the outer sleeve and the inner core. Too strong differences will result in inhomogeneous deformation leading to defects (surface cracks) in the tapes. It was reported that tungsten (W) is a good choice for solid solution strengthening without the problems of oxidation. Ni-W-alloys form a strong cube texture up to 5% W [11,21,28]. This is proved by a critical current density of 1.2 MAcm−2 achieved on an YBCO film deposited on such a substrate [11–12,28]. For the core of the composite Ni-Cr
V. Subramanya Sarma et al. / Acta Materialia 51 (2003) 4919–4927
and Ni-V alloys are suitable as the Curie temperature is suppressed to below 77 K with relatively small additions of Cr and V (~12%) [8,20]. Greater alloying additions (⬎13% of Cr and ⬎10% of V) also weaken the recrystallisation cube texture in these alloys though strength increases with increasing alloying content [34]. Since the substrate thickness is small (⬍ 80 µm) and the number of grains in the thickness direction are few (2–4), it is likely that the presence of a non-cube texture forming alloy in the core of the composite could have a detrimental effect on the growth of the cube oriented grains on the surface during recrystallisation. This issue is the focus of the current investigation.
2. Experimental Ni-4.5% W and Ni-15% Cr alloys were prepared by melting the elements having a purity of 99.98% in an induction furnace and casting them in a cylindrical mould of 32 mm diameter. The Ni- 4.5% W alloy was hot rolled at 1100 °C to a square 22 × 22 mm2 rod. The Ni-15% Cr alloy was hot forged at 1100 °C to a 10 mm diameter and inserted into the Ni-4.5%W rod [32]. This pre-form was hot and cold rolled to 80 and 40 µm thickness with the final cold deformation being ⬎95%. The hot rolling serves two purposes 1) to achieve good bonding between the outer and inner layers and, 2) to recrystallise the outer layer and the inner core. Samples cut from the long tape were recrystallised at 900, 950 or 1000 °C for 30 min to 2 h in high vacuum. The microtextures and misorientation distributions were investigated by Electron Back Scattered Diffraction (EBSD) technique in the scanning electron microscope (JEOL JSM 6400) with Channel (HKL Technology, Denmark) software. The Xray pole figures, Phi (f) and Omega (w) scans were performed in a Philips texture goniometer with CuKα radiation. The yield and tensile strengths of the tapes were measured with an INSTRON 8500 testing machine with 25 mm extensometer on samples of 10 cm length. The loading direction was along the [100] crystallographic direction of the biaxially textured substrates. Magnetisation measurements were made at 77 K in a Quantum Design PPMS ACMS magnetometer. Pulsed laser deposition
4921
(PLD) of CeO2 buffer layer (40 nm thickness) was carried out using a standard oxide PLD set up and with deposition parameters described in detail in [35].
3. Results and discussion 3.1. Microstructure and texture The microstructure after the hot rolling stage shows a completely recrystallised and fine grained structure in the outer sleeve and inner core of the composite (Fig. 1). Another important point is the strength difference between the outer sleeve and the core. From the Vickers microhardness measurements (with 98 mN of load) it can be seen that the strength difference is insignificant (Fig. 1). These two conditions seem to be important in achieving strong cube texture following further cold working and recrystallisation [32–33]. The grain boundary misorientation distributions (GBMDs) obtained from the microtexture measurements for the recrystallised 80 and 40 µm tapes are shown in Figs. 2(a) and (b) respectively. It is clear that substantial fractions of twin boundaries are present following recrystallisation annealing at 950 and 1000 °C for 30 min. Annealing for longer time (1000 °C/2 h) was carried out with a view to reduce/eliminate the twin boundaries during grain growth and this did result
Fig. 1. Microstructure of the Ni-4.5% W/Ni-15% Cr composite after the intermediate hot rolling.
4922
V. Subramanya Sarma et al. / Acta Materialia 51 (2003) 4919–4927
in the reduction of twin boundary fraction (Fig. 2(a)). But a secondary recrystallisation or abnormal grain growth occurred during which 45° misoriented grains started to grow at the expense of cube grains (Fig. 2(a)). This can also be clearly observed from the EBSD mapping (Fig. 2(c)). In view of this result, annealing for longer times or higher temperatures (⬎1000 °C) may not be beneficial for cube grain growth. While secondary recrystallisation or abnormal grain growth has been observed in Ni and Ni-Cr alloys [8,36], no such result has been reported so far in the Ni-W alloys. So to develop strong cube textures a two-step annealing procedure was devised. The basic idea is to nucleate few cube grains at relatively low temperature and selectively grow these grains with a steady increase in temperature. In the two step process the samples were heated to 700 °C held for 30 min. followed by a steady increase in temperature (100 K/h) to 1000 °C and held for 30 min. Fig. 3(a) and (b) shows the EBSD maps of the 80 and 40 µm composite tapes following the two-step annealing treatment. It is clear that the two-step annealing procedure was successful in obtaining very strong cube textures with a very negligible fraction of twin and other misoriented grains (Fig. 3(a), (b) and (c)). This can also be seen from the background corrected logarithmic scale {111} X-ray pole figure for the 80 µm thickness substrate (Fig. 4). From the X-ray pole figure and the EBSD data, the volume fraction of the cube texture component (defined as with in 15° from the ideal cube orientation {001}⬍100⬎) was calculated to be ~99.5%. The X-ray f and w scans (in and about the rolling direction) of the 80 µm composite are shown in Fig. 5(a), (b) and (c) respectively. From these scans the FWHM for the in plane texture and out of plane texture (in and about the rolling direction) were estimated (by Gaussian fit) to be 6.7, 4.7 and 4.8° respectively. Fig. 6(a) shows the sectional view micrograph of the composite (80 µm thickness) and Fig. 6(b) shows the sectional EBSD map after Fig. 2. Misorientation distribution of the Ni-4.5% W/Ni-15% Cr composite after recrystallisation in (a) 80 µm, (b) 40 µm thickness and (c) EBSD mapping of the composite following recrystallisation at 1000 °C for 2 h (grain boundaries ⬎15° are indicated by a black line).
V. Subramanya Sarma et al. / Acta Materialia 51 (2003) 4919–4927
4923
Fig. 4. X-ray {111} pole figure of the Ni-4.5% W/Ni-15% Cr composite substrate of 80 µm following two step annealing.
recrystallisation. From the EBSD map it can be seen that the core of the composite has many noncube oriented grains. This indicates that the texture in the inner core does not have a detrimental effect on the growth of cube grains on the surface of the composite during recrystallisation. This result is significant as it gives scope for further optimization of mechanical and magnetic properties (see below). 3.2. Tensile properties
Fig. 3. EBSD maps after two step recrystallisation in (a) 80 µm and (b) 40 µm thicknes tapes and (c) misorientation distributions in 80 and 40 µm thickness tapes (grain boundaries ⬎15° are indicated by a black line).
Since the superconducting coating does not withstand a strain above 0.5% in compression and 0.2% in tension without degradation, the stress at low strains (e.g. 0.2% offset yield strength) is critical for the application as a substrate material [37]. The room temperature stress-strain response of the 80 and 40 µm composite tapes following the twostep annealing are given in Fig. 7. Stress–strain data of pure Ni tape of 80 µm thickness is also included for comparison. It can be seen that the composite exhibits increased yield strength (~200 MPa) i.e., by a factor of 5 when compared to pure Ni (Fig. 7). The thinner (40 µm) composite exhibited slightly higher yield stress (Fig. 7) when compared with the 80 µm tape. It should be noted that the yield stress of thin films is strongly influenced by dimensional constraint on dislocation motion (due to increased contribution of image forces on dislocations), which results in a pronounced size (inverse thickness) effect [38]. In the present case
4924
V. Subramanya Sarma et al. / Acta Materialia 51 (2003) 4919–4927
Fig. 6. Cross sectional (a) microstructure and, (b) EBSD map of the Ni-4.5% W/Ni-15% Cr composte of 80 µm thickness after recrystallisation (grain boundaries ⬎15° are indicated by a black line).
Fig. 5. (a) X-ray f scan and (b) w scan of the Ni-4.5% W/Ni15% Cr composite substrate (80 µm thickness) after two step recrystallisation annealing.
Fig. 7. Room temperature tensile properties of the Ni-4.5% W/Ni-15% Cr composite at 80 and 40 µm thickness after two step recrystallisation.
V. Subramanya Sarma et al. / Acta Materialia 51 (2003) 4919–4927
4925
the increased strength of the composite comes from solid solution strengthening of the matrix and also possibly the constraint on dislocation motion by the presence of interfaces due to the difference in texture between the outer layer and inner core. 3.3. Magnetisation behaviour The temperature dependence of the mass magnetisation M(T) of the composite measured in an applied magnetic field of 5 kOe is shown in Fig. 8(a) along with the data for Ni, Ni-5%W (composition close to the outer sleeve) and Ni-12% Cr. It can be seen that Ni-12% Cr is non-magnetic at 77 K and the saturation magnetisation of the composite is suppressed when compared with that of Ni-5% W substrate. The M-H hysteresis loops at 77 K (operating temperature of the coated conductor) for the Ni, Ni-5%W, and the composite (after recrystallisation) are shown Fig. 8(b). High resolution M-H hysteresis loop measurements are shown in Fig. 8(c). From Fig. 8(b) and (c) it can be seen that the hysteresis losses are reduced in the composite substrate when compared with the pure Ni and Ni-5% W substrate. 3.4. Texture in the buffer layer The standard architecture of the RABiTS method uses a thin ~50 nm layer of CeO2 and ~500 nm of YSZ before the deposition of the HTS YBCO film. The purpose of the buffer layers is to prevent the diffusion of substrate material (Ni) into the HTS YBCO film during deposition as this destroys the superconductivity in the HTS YBCO film. The buffer layers should also be biaxially textured for obtaining an ab-axis orientation in the HTS YBCO film. Using PLD, a 40 nm thick CeO2 film was deposited on the composite substrate and the texture was measured by EBSD. The {111} pole figure calculated from the EBSD data is shown in Fig. 9(a). It shows a strong biaxial (45° rotated cube) texture. The misorientation distribution in the buffer layer shows a maximum close to 4° with virtually negligible fraction of high angle (⬎12°) boundaries (Fig. 9(b)). Interestingly the maximum of the grain boundary distribution function is significantly lower than that for the sub-
Fig. 8. (a) Mass magnetisation of the recrystallised Ni-4.5% W/Ni-15% Cr composite, Ni, Ni-5% W and Ni-12% Cr as a function of temperature and, (b) magnetisation loops at 77 K for Ni, Ni-5% W and the composite.
4926
V. Subramanya Sarma et al. / Acta Materialia 51 (2003) 4919–4927
Fig. 9. (a) {111} pole figure and, (b) misorientation distribution in the 40 nm CeO2 buffer layer deposited by pulsed laser deposition on the Ni-4.5% W/Ni-15% Cr composite substrate.
strate (~7°) (Fig. 9(b)), This can also be seen from the FWHMs in X-ray f and w scans on the substrate (Fig. 5) and buffer layer (Fig. 10). The improvement in the in-plane texture is ~2° and the out of plane texture ~1°. The above results indicate a self-texturing mechanism during CeO2 film growth, which was up to now only shown during growth of YBCO film [39]. Further investigations are necessary to understand the exact growth mechanism. The texture quality in the buffer layer is excellent for further deposition of YSZ and YBCO films. From the EBSD data of the substrate and the buffer layer, current limiting path simulations [39] were performed to estimate the likely enhancement of critical current density (Jc) due to the texture improvement in the buffer layer. The
Fig. 10. (a) X-ray f scan and (b) w scan of the pulsed laser deposited 40 nm CeO2 buffer layer on the Ni-4.5% W/Ni-15% Cr composite substrate of 80 µm thickness.
results showed significant improvement in the buffer layer i.e., the simulated current density was 68% of single crystal Jc when compared with the 36% of single crystal Jc in the substrate. The experimental results of YBCO deposition and the current transport measurements will be reported separately.
4. Conclusions Strongly cube textured and high strength (yield strength ~200 MPa) composite (Ni-4.5% W/Ni15% Cr) substrates of 80 and 40 µm thickness with reduced magnetisation losses were developed for coated conductor application. The strong cube tex-
V. Subramanya Sarma et al. / Acta Materialia 51 (2003) 4919–4927
ture was obtained by an optimized two-step recrystallisation annealing treatment. The presence of non-cube forming alloy (Ni-15% Cr) in the core of the composite had no adverse affect on the growth of cube textured grains on the surface even at low substrate thickness of 40 µm. A significant improvement in the texture/misorientation distribution was observed in the CeO2 buffer layer deposited on the composite substrate. Acknowledgements The authors thank Mr. Frey, Mr. Kuszinski, Mr. Opitz, Mr. Seifert, Mr. Wolf, Mr. Neumann, Mr. Trinks, Mr. Klauss, and Mrs. Grundlich for their help in the experimental work. This work is supported by the Federal Ministry of Education, Science and Technology, Germany under contract No. 13N7267A. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13]
Norton DP, et al. Science 1996;274:755. Goyal A, et al. Appl Phys Lett 1996;69:1795. Norton DP, et al. Mater Sci Eng B 1998;56:86. Ceresara S, Boffa V, Petrisor T, Fabbri F, Scardi P. Int J Modern Physics B 1999;13:1035. Eickemeyer J, Selbmann D, Opitz R, maher E, Prusseit W. Physica C 2000;341–348:2425. Reger N, deBoer B, Eickemeyer J, Opitz R, Holzapfel B, Schultz L. Inst Phys Conf Ser 167:2000;331. Tuissi A, Masini A, Villa E, Evetts JE. Int J Modern Physics B, 2000;14:3139. de Boer B, et al. Acta Mater 2001;49:1421. de Boer B, et al. IEEE Trans App Supercond 2001;11:3477. de Boer B, et al. Physica C 2001;351:38. Eickemeyer J, et al. Supercond Sci Technol 2001;14:152. Knauf J, Semerad R, Prusseit W, de Boer B, Eickemeyer J. IEEE Trans on Applied Superconductivity, 2001;11:2885. Gladstone TA, Moore JC, Wilkinson AJ, Governor CRM, IEEE Trans Appl Supercond 2001;11:2923.
4927
[14] Nekkanti RM, et al. IEEE Trans Appl Supercond 2001;11:3321. [15] Thieme CLH, et al. IEEE Trans Appl Supercond 2001;11:3329. [16] Rodionov DP, Gervaseva IV, Khlebnikova YV, Sokolov BK. Physic of Metals and Metallography 2001;92:232. [17] Glowacki BA, et al. J Mater Sci 2002;37:157. [18] Tomov RI, Kursumovic A, Majoros M, Kang DJ, Glowacki BA, Evetts JE. Supercond Sci Technol 2002;15:598. [19] Satyamurthy, et al. J Mater Res 2002;17:1543. [20] Thompson JR, Goyal A, Christen DK, Kroeger DM. Physica C 2002;370:169 [21] Eickemeyer J, et al. Physica C 2002;372–376:814. [22] Sarma VS, de Boer B, Reger N, Eickemeyer J, Opitz R, Holzapfel B. Mater Sci Forum 2002;408–412:1561. [23] Varesi E, et al. Physica C 2002;372–376:763. [24] Tuissi A, Villa E, Zamboni M, Evetts JE, Tomov RI. Physica C 2002;372–376:759. [25] Goyal A, et al. Physica C;2002:251. [26] Nast R, Obst B, Goldacker W. Physica C 2002;372– 376:733. [27] Tomov R, et al. Physica C 2003;383:323. [28] Varesi E, et al. Supercond Sci Technol 2003;16:498. [29] de Boer B, Sarma VS, Reger N Eickemeyer J, Holzapfel B. Physica C 2002;372–376:798. [30] Goyal A. US Patent No. 6,180,570; 2001. [31] Watanabe T, et al. HTS conductors, processing and applications, The 2002 International Workshop on Superconductivity (The 5th joint ISTEC/MRS Workshop), Honolulu, Hawai, USA; June 24–27, 2001, pp. 21–24. [32] Subramanya Sarma V, de Boer B, Eickemeyer J, Holzapfel B. Scripta Mater 2003;48:1167. [33] Subramanya Sarma V, de Boer B, Eickemeyer J, Holzapfel B. Acta mater 2003;51:3769. [34] de Boer B, Reger N, Holzapfel B. Proc. First Joint Int Conf On Recrystallisation and Grain Growth, Gottstein G, Moldov DA, editors. Berlin Heildelberg: Springer-Verlag; 2001, p. 1355. [35] Fernandez L, Holzapfel B, Schindler F, de Boer B, Schultz L. Physica C 2002;372–376:656. [36] de Boer B, Reger N, Opitz R, Eickemeyer J, Holzapfel B, Schultz L. Proc ICOTOM 12, Szupnar JA, editor. Ottawa: NRC Research Press; 1999, p. 944. [37] Goyal A, et al. Mater Res 1997;12:2924. [38] Arzt E. Acta mater 1998;46:5611. [39] Holzapfel B, et al. IEEE Trans on Applied Superconductivity 2001;11:3872.
Development of high strength and strongly cube textured Ni-4.5% W/Ni-15% Cr composite substrate for coated conductor application V. Subramanya Sarma ∗, J. Eickemeyer, A. Singh, L. Schultz, B. Holzapfel Leibniz Institute for Solid State and Materials Research (IFW), Helmholtz strasse 20, 01069, Dresden, Germany Received 30 April 2003; received in revised form 10 June 2003; accepted 16 June 2003
Abstract The development of thin, mechanically stronger and highly cube textured substrates is of great technological importance for increasing the engineering current density of the coated conductors. Nickel is a suitable substrate for this in view of its ability to form strong cube texture after heavy rolling and annealing and its excellent oxidation resistance. However, nickel is very soft (yield strength ~40 MPa) and this limits the processing to thin tapes. The ferromagnetism of Ni is also undesirable for ac application of coated conductors in magnetic fields. In the present paper we report on the development of Ni-4.5 at.% W/Ni-15 at.% Cr composite substrates of 80 and 40 µm thickness with strong cube texture, high yield strength (~200 MPa) and reduced magnetisation losses. The strong cube texture was obtained through an optimised two-step recrystallisation annealing following heavy cold working. It was found that the presence of noncube texture forming alloy (Ni-15% Cr) in the inner core of the composite had no adverse affect on the growth of cube textured grains on the surface (Ni-4.5% W) even at a low substrate thickness of 40 µm. A significant improvement in the texture/misorientation distribution was observed in the CeO2 buffer layer deposited on the composite substrate. 2003 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. Keywords: Coated conductors; Recrystallisation; Cube texture; Ni-alloys; Composite
1. Introduction The Rolling Assisted Biaxially Textured Substrates (RABiTS) method of Ni and Ni-based alloys (NiCu, NiFe, NiV, NiCr, NiW) is a very promising approach for the profitable production of long lengths of high temperature superconducting ∗ Corresponding author. Tel.: +49-351-4659203, fax: +49351-4659320. E-mail address: [email protected] (V. Subramanya Sarma).
(HTS) YBa2Cu3O7⫺d (YBCO) tapes capable of carrying high currents in magnetic fields at 77 K [1–28]. In this method, the desired strong biaxial texture in the superconducting YBCO film is achieved by epitaxial growth of a buffer and YBCO film on a highly textured substrate. Initially it has been shown that Ni is ideally suited as a substrate material due to its ability to form strong a cube texture after heavy cold rolling and recrystallisation. In addition its oxidation resistance and the small lattice mismatch allows epitaxial growth of buffer (for e.g. CeO2 + yttria stabilized zirconia
1359-6454/$30.00 2003 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. doi:10.1016/S1359-6454(03)00334-3
4920
V. Subramanya Sarma et al. / Acta Materialia 51 (2003) 4919–4927
(YSZ)) and YBCO films. In YBCO films on Ni substrates, current densities exceeding 106 Acm⫺2 have been achieved [1–3]. To achieve a high engineering current density (i.e., the current in the HTS film divided by the cross section of the whole tape) very thin tapes are desirable. But the low tensile strength of Ni limits the possibility of producing very thin HTS tapes by a continuous reel-toreel deposition process. For these reasons the strengthening of Ni substrate material is important. The need for a strong cube texture limits the possibilities of strengthening the substrate. Further, the ferromagnetism of Ni leads to hysteresis losses in coated conductors in alternating current (AC) applications [20], but it is not of great importance for direct current (DC) applications. There are two known metallurgical routes of strengthening the recrystallised substrate material i.e., solid solution and precipitation strengthening. It was shown that the substrate strength could be increased by factor 3 with substantial additions of Cr, V or W (ⱖ5 at.%, compositions are given in at.% throughout the paper) without losing the strong cube texture [8–11]. The additions (⬎10%) of Cr and V also suppress the Curie temperature to below 77 K, thus reducing the magnetization losses in the coated conductors [8,20]. However, the ease of Cr/V oxide formation on the surface at the buffer and YBCO deposition temperatures (these are prepared in oxidizing atmospheres) makes the epitaxial growth difficult. Higher strength levels can also be achieved by precipitation of Al2O3 particles through controlled internal oxidation of a Ni-1% Al alloy [29,33]. However, the formation of Al2O3 particles on the surface degrades its quality rendering it unsuitable for further epitaxial deposition. These surface oxidation problem can be overcome by the design of a composite substrate [30–33]. Using such an approach we recently reported the development of Ni-3% W/Ni-10% Cr-1.5% Al composite substrate of 80 µm thickness with high strength (room temperature yield strength ~175 MPa) and strong cube texture [32]. In this investigation, a cube texture forming alloy (Ni-10% Cr1.5% Al) was used in the core of the composite [32]. In the present paper, we report on the development of strongly cube textured and high strength Ni-4.5% W (outer sleeve)/Ni-15% Cr (a non-cube
texture forming alloy as the core) composite substrate of 80 and 40 µm thickness. Also we present the magnetisation behaviour of the composite tape at 77 K and texture results in the CeO2 buffer layer grown epitaxially by pulsed laser deposition (PLD) technique. 1.1. Design of the composite tapes The critical issues in the design of the composite tapes are discussed in a recent paper [32]. For brevity some important points are recalled here. It is known that the recrystallisation cube texture in Ni and Ni alloys is significantly improved with a fine starting grain size. This can be achieved through an intermediate recrystallisation annealing before the application of heavy (⬎95%) cold rolling [21]. Therefore, the problem with using pure Ni in the outer layer of the composite is with regard to achieving the required strong cube texture due to the wide differences in the recrystallisation temperatures between Ni (typically around 300–400 °C) and Ni-Cr, Ni-V alloys (typically 700–800 °C depending on the amount of alloy content). It should be mentioned that a coarse starting grain size has a strong detrimental effect on the final cube texture. Therefore during intermediate annealing of a composite with pure Ni outer layer and highly alloyed core would lead either to fine grain size in Ni and no recrystallisation in the core or coarse grained outer layer with recrystallised core. These microstructures and all in between are not useful to develop the very strong cube textures needed for the RABiTS application [33]. It is thus necessary to have the outer surface and the inner core materials with similar recrystallisation temperatures. Also of importance during the rolling of composite is the strength difference between the outer sleeve and the inner core. Too strong differences will result in inhomogeneous deformation leading to defects (surface cracks) in the tapes. It was reported that tungsten (W) is a good choice for solid solution strengthening without the problems of oxidation. Ni-W-alloys form a strong cube texture up to 5% W [11,21,28]. This is proved by a critical current density of 1.2 MAcm−2 achieved on an YBCO film deposited on such a substrate [11–12,28]. For the core of the composite Ni-Cr
V. Subramanya Sarma et al. / Acta Materialia 51 (2003) 4919–4927
and Ni-V alloys are suitable as the Curie temperature is suppressed to below 77 K with relatively small additions of Cr and V (~12%) [8,20]. Greater alloying additions (⬎13% of Cr and ⬎10% of V) also weaken the recrystallisation cube texture in these alloys though strength increases with increasing alloying content [34]. Since the substrate thickness is small (⬍ 80 µm) and the number of grains in the thickness direction are few (2–4), it is likely that the presence of a non-cube texture forming alloy in the core of the composite could have a detrimental effect on the growth of the cube oriented grains on the surface during recrystallisation. This issue is the focus of the current investigation.
2. Experimental Ni-4.5% W and Ni-15% Cr alloys were prepared by melting the elements having a purity of 99.98% in an induction furnace and casting them in a cylindrical mould of 32 mm diameter. The Ni- 4.5% W alloy was hot rolled at 1100 °C to a square 22 × 22 mm2 rod. The Ni-15% Cr alloy was hot forged at 1100 °C to a 10 mm diameter and inserted into the Ni-4.5%W rod [32]. This pre-form was hot and cold rolled to 80 and 40 µm thickness with the final cold deformation being ⬎95%. The hot rolling serves two purposes 1) to achieve good bonding between the outer and inner layers and, 2) to recrystallise the outer layer and the inner core. Samples cut from the long tape were recrystallised at 900, 950 or 1000 °C for 30 min to 2 h in high vacuum. The microtextures and misorientation distributions were investigated by Electron Back Scattered Diffraction (EBSD) technique in the scanning electron microscope (JEOL JSM 6400) with Channel (HKL Technology, Denmark) software. The Xray pole figures, Phi (f) and Omega (w) scans were performed in a Philips texture goniometer with CuKα radiation. The yield and tensile strengths of the tapes were measured with an INSTRON 8500 testing machine with 25 mm extensometer on samples of 10 cm length. The loading direction was along the [100] crystallographic direction of the biaxially textured substrates. Magnetisation measurements were made at 77 K in a Quantum Design PPMS ACMS magnetometer. Pulsed laser deposition
4921
(PLD) of CeO2 buffer layer (40 nm thickness) was carried out using a standard oxide PLD set up and with deposition parameters described in detail in [35].
3. Results and discussion 3.1. Microstructure and texture The microstructure after the hot rolling stage shows a completely recrystallised and fine grained structure in the outer sleeve and inner core of the composite (Fig. 1). Another important point is the strength difference between the outer sleeve and the core. From the Vickers microhardness measurements (with 98 mN of load) it can be seen that the strength difference is insignificant (Fig. 1). These two conditions seem to be important in achieving strong cube texture following further cold working and recrystallisation [32–33]. The grain boundary misorientation distributions (GBMDs) obtained from the microtexture measurements for the recrystallised 80 and 40 µm tapes are shown in Figs. 2(a) and (b) respectively. It is clear that substantial fractions of twin boundaries are present following recrystallisation annealing at 950 and 1000 °C for 30 min. Annealing for longer time (1000 °C/2 h) was carried out with a view to reduce/eliminate the twin boundaries during grain growth and this did result
Fig. 1. Microstructure of the Ni-4.5% W/Ni-15% Cr composite after the intermediate hot rolling.
4922
V. Subramanya Sarma et al. / Acta Materialia 51 (2003) 4919–4927
in the reduction of twin boundary fraction (Fig. 2(a)). But a secondary recrystallisation or abnormal grain growth occurred during which 45° misoriented grains started to grow at the expense of cube grains (Fig. 2(a)). This can also be clearly observed from the EBSD mapping (Fig. 2(c)). In view of this result, annealing for longer times or higher temperatures (⬎1000 °C) may not be beneficial for cube grain growth. While secondary recrystallisation or abnormal grain growth has been observed in Ni and Ni-Cr alloys [8,36], no such result has been reported so far in the Ni-W alloys. So to develop strong cube textures a two-step annealing procedure was devised. The basic idea is to nucleate few cube grains at relatively low temperature and selectively grow these grains with a steady increase in temperature. In the two step process the samples were heated to 700 °C held for 30 min. followed by a steady increase in temperature (100 K/h) to 1000 °C and held for 30 min. Fig. 3(a) and (b) shows the EBSD maps of the 80 and 40 µm composite tapes following the two-step annealing treatment. It is clear that the two-step annealing procedure was successful in obtaining very strong cube textures with a very negligible fraction of twin and other misoriented grains (Fig. 3(a), (b) and (c)). This can also be seen from the background corrected logarithmic scale {111} X-ray pole figure for the 80 µm thickness substrate (Fig. 4). From the X-ray pole figure and the EBSD data, the volume fraction of the cube texture component (defined as with in 15° from the ideal cube orientation {001}⬍100⬎) was calculated to be ~99.5%. The X-ray f and w scans (in and about the rolling direction) of the 80 µm composite are shown in Fig. 5(a), (b) and (c) respectively. From these scans the FWHM for the in plane texture and out of plane texture (in and about the rolling direction) were estimated (by Gaussian fit) to be 6.7, 4.7 and 4.8° respectively. Fig. 6(a) shows the sectional view micrograph of the composite (80 µm thickness) and Fig. 6(b) shows the sectional EBSD map after Fig. 2. Misorientation distribution of the Ni-4.5% W/Ni-15% Cr composite after recrystallisation in (a) 80 µm, (b) 40 µm thickness and (c) EBSD mapping of the composite following recrystallisation at 1000 °C for 2 h (grain boundaries ⬎15° are indicated by a black line).
V. Subramanya Sarma et al. / Acta Materialia 51 (2003) 4919–4927
4923
Fig. 4. X-ray {111} pole figure of the Ni-4.5% W/Ni-15% Cr composite substrate of 80 µm following two step annealing.
recrystallisation. From the EBSD map it can be seen that the core of the composite has many noncube oriented grains. This indicates that the texture in the inner core does not have a detrimental effect on the growth of cube grains on the surface of the composite during recrystallisation. This result is significant as it gives scope for further optimization of mechanical and magnetic properties (see below). 3.2. Tensile properties
Fig. 3. EBSD maps after two step recrystallisation in (a) 80 µm and (b) 40 µm thicknes tapes and (c) misorientation distributions in 80 and 40 µm thickness tapes (grain boundaries ⬎15° are indicated by a black line).
Since the superconducting coating does not withstand a strain above 0.5% in compression and 0.2% in tension without degradation, the stress at low strains (e.g. 0.2% offset yield strength) is critical for the application as a substrate material [37]. The room temperature stress-strain response of the 80 and 40 µm composite tapes following the twostep annealing are given in Fig. 7. Stress–strain data of pure Ni tape of 80 µm thickness is also included for comparison. It can be seen that the composite exhibits increased yield strength (~200 MPa) i.e., by a factor of 5 when compared to pure Ni (Fig. 7). The thinner (40 µm) composite exhibited slightly higher yield stress (Fig. 7) when compared with the 80 µm tape. It should be noted that the yield stress of thin films is strongly influenced by dimensional constraint on dislocation motion (due to increased contribution of image forces on dislocations), which results in a pronounced size (inverse thickness) effect [38]. In the present case
4924
V. Subramanya Sarma et al. / Acta Materialia 51 (2003) 4919–4927
Fig. 6. Cross sectional (a) microstructure and, (b) EBSD map of the Ni-4.5% W/Ni-15% Cr composte of 80 µm thickness after recrystallisation (grain boundaries ⬎15° are indicated by a black line).
Fig. 5. (a) X-ray f scan and (b) w scan of the Ni-4.5% W/Ni15% Cr composite substrate (80 µm thickness) after two step recrystallisation annealing.
Fig. 7. Room temperature tensile properties of the Ni-4.5% W/Ni-15% Cr composite at 80 and 40 µm thickness after two step recrystallisation.
V. Subramanya Sarma et al. / Acta Materialia 51 (2003) 4919–4927
4925
the increased strength of the composite comes from solid solution strengthening of the matrix and also possibly the constraint on dislocation motion by the presence of interfaces due to the difference in texture between the outer layer and inner core. 3.3. Magnetisation behaviour The temperature dependence of the mass magnetisation M(T) of the composite measured in an applied magnetic field of 5 kOe is shown in Fig. 8(a) along with the data for Ni, Ni-5%W (composition close to the outer sleeve) and Ni-12% Cr. It can be seen that Ni-12% Cr is non-magnetic at 77 K and the saturation magnetisation of the composite is suppressed when compared with that of Ni-5% W substrate. The M-H hysteresis loops at 77 K (operating temperature of the coated conductor) for the Ni, Ni-5%W, and the composite (after recrystallisation) are shown Fig. 8(b). High resolution M-H hysteresis loop measurements are shown in Fig. 8(c). From Fig. 8(b) and (c) it can be seen that the hysteresis losses are reduced in the composite substrate when compared with the pure Ni and Ni-5% W substrate. 3.4. Texture in the buffer layer The standard architecture of the RABiTS method uses a thin ~50 nm layer of CeO2 and ~500 nm of YSZ before the deposition of the HTS YBCO film. The purpose of the buffer layers is to prevent the diffusion of substrate material (Ni) into the HTS YBCO film during deposition as this destroys the superconductivity in the HTS YBCO film. The buffer layers should also be biaxially textured for obtaining an ab-axis orientation in the HTS YBCO film. Using PLD, a 40 nm thick CeO2 film was deposited on the composite substrate and the texture was measured by EBSD. The {111} pole figure calculated from the EBSD data is shown in Fig. 9(a). It shows a strong biaxial (45° rotated cube) texture. The misorientation distribution in the buffer layer shows a maximum close to 4° with virtually negligible fraction of high angle (⬎12°) boundaries (Fig. 9(b)). Interestingly the maximum of the grain boundary distribution function is significantly lower than that for the sub-
Fig. 8. (a) Mass magnetisation of the recrystallised Ni-4.5% W/Ni-15% Cr composite, Ni, Ni-5% W and Ni-12% Cr as a function of temperature and, (b) magnetisation loops at 77 K for Ni, Ni-5% W and the composite.
4926
V. Subramanya Sarma et al. / Acta Materialia 51 (2003) 4919–4927
Fig. 9. (a) {111} pole figure and, (b) misorientation distribution in the 40 nm CeO2 buffer layer deposited by pulsed laser deposition on the Ni-4.5% W/Ni-15% Cr composite substrate.
strate (~7°) (Fig. 9(b)), This can also be seen from the FWHMs in X-ray f and w scans on the substrate (Fig. 5) and buffer layer (Fig. 10). The improvement in the in-plane texture is ~2° and the out of plane texture ~1°. The above results indicate a self-texturing mechanism during CeO2 film growth, which was up to now only shown during growth of YBCO film [39]. Further investigations are necessary to understand the exact growth mechanism. The texture quality in the buffer layer is excellent for further deposition of YSZ and YBCO films. From the EBSD data of the substrate and the buffer layer, current limiting path simulations [39] were performed to estimate the likely enhancement of critical current density (Jc) due to the texture improvement in the buffer layer. The
Fig. 10. (a) X-ray f scan and (b) w scan of the pulsed laser deposited 40 nm CeO2 buffer layer on the Ni-4.5% W/Ni-15% Cr composite substrate of 80 µm thickness.
results showed significant improvement in the buffer layer i.e., the simulated current density was 68% of single crystal Jc when compared with the 36% of single crystal Jc in the substrate. The experimental results of YBCO deposition and the current transport measurements will be reported separately.
4. Conclusions Strongly cube textured and high strength (yield strength ~200 MPa) composite (Ni-4.5% W/Ni15% Cr) substrates of 80 and 40 µm thickness with reduced magnetisation losses were developed for coated conductor application. The strong cube tex-
V. Subramanya Sarma et al. / Acta Materialia 51 (2003) 4919–4927
ture was obtained by an optimized two-step recrystallisation annealing treatment. The presence of non-cube forming alloy (Ni-15% Cr) in the core of the composite had no adverse affect on the growth of cube textured grains on the surface even at low substrate thickness of 40 µm. A significant improvement in the texture/misorientation distribution was observed in the CeO2 buffer layer deposited on the composite substrate. Acknowledgements The authors thank Mr. Frey, Mr. Kuszinski, Mr. Opitz, Mr. Seifert, Mr. Wolf, Mr. Neumann, Mr. Trinks, Mr. Klauss, and Mrs. Grundlich for their help in the experimental work. This work is supported by the Federal Ministry of Education, Science and Technology, Germany under contract No. 13N7267A. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13]
Norton DP, et al. Science 1996;274:755. Goyal A, et al. Appl Phys Lett 1996;69:1795. Norton DP, et al. Mater Sci Eng B 1998;56:86. Ceresara S, Boffa V, Petrisor T, Fabbri F, Scardi P. Int J Modern Physics B 1999;13:1035. Eickemeyer J, Selbmann D, Opitz R, maher E, Prusseit W. Physica C 2000;341–348:2425. Reger N, deBoer B, Eickemeyer J, Opitz R, Holzapfel B, Schultz L. Inst Phys Conf Ser 167:2000;331. Tuissi A, Masini A, Villa E, Evetts JE. Int J Modern Physics B, 2000;14:3139. de Boer B, et al. Acta Mater 2001;49:1421. de Boer B, et al. IEEE Trans App Supercond 2001;11:3477. de Boer B, et al. Physica C 2001;351:38. Eickemeyer J, et al. Supercond Sci Technol 2001;14:152. Knauf J, Semerad R, Prusseit W, de Boer B, Eickemeyer J. IEEE Trans on Applied Superconductivity, 2001;11:2885. Gladstone TA, Moore JC, Wilkinson AJ, Governor CRM, IEEE Trans Appl Supercond 2001;11:2923.
4927
[14] Nekkanti RM, et al. IEEE Trans Appl Supercond 2001;11:3321. [15] Thieme CLH, et al. IEEE Trans Appl Supercond 2001;11:3329. [16] Rodionov DP, Gervaseva IV, Khlebnikova YV, Sokolov BK. Physic of Metals and Metallography 2001;92:232. [17] Glowacki BA, et al. J Mater Sci 2002;37:157. [18] Tomov RI, Kursumovic A, Majoros M, Kang DJ, Glowacki BA, Evetts JE. Supercond Sci Technol 2002;15:598. [19] Satyamurthy, et al. J Mater Res 2002;17:1543. [20] Thompson JR, Goyal A, Christen DK, Kroeger DM. Physica C 2002;370:169 [21] Eickemeyer J, et al. Physica C 2002;372–376:814. [22] Sarma VS, de Boer B, Reger N, Eickemeyer J, Opitz R, Holzapfel B. Mater Sci Forum 2002;408–412:1561. [23] Varesi E, et al. Physica C 2002;372–376:763. [24] Tuissi A, Villa E, Zamboni M, Evetts JE, Tomov RI. Physica C 2002;372–376:759. [25] Goyal A, et al. Physica C;2002:251. [26] Nast R, Obst B, Goldacker W. Physica C 2002;372– 376:733. [27] Tomov R, et al. Physica C 2003;383:323. [28] Varesi E, et al. Supercond Sci Technol 2003;16:498. [29] de Boer B, Sarma VS, Reger N Eickemeyer J, Holzapfel B. Physica C 2002;372–376:798. [30] Goyal A. US Patent No. 6,180,570; 2001. [31] Watanabe T, et al. HTS conductors, processing and applications, The 2002 International Workshop on Superconductivity (The 5th joint ISTEC/MRS Workshop), Honolulu, Hawai, USA; June 24–27, 2001, pp. 21–24. [32] Subramanya Sarma V, de Boer B, Eickemeyer J, Holzapfel B. Scripta Mater 2003;48:1167. [33] Subramanya Sarma V, de Boer B, Eickemeyer J, Holzapfel B. Acta mater 2003;51:3769. [34] de Boer B, Reger N, Holzapfel B. Proc. First Joint Int Conf On Recrystallisation and Grain Growth, Gottstein G, Moldov DA, editors. Berlin Heildelberg: Springer-Verlag; 2001, p. 1355. [35] Fernandez L, Holzapfel B, Schindler F, de Boer B, Schultz L. Physica C 2002;372–376:656. [36] de Boer B, Reger N, Opitz R, Eickemeyer J, Holzapfel B, Schultz L. Proc ICOTOM 12, Szupnar JA, editor. Ottawa: NRC Research Press; 1999, p. 944. [37] Goyal A, et al. Mater Res 1997;12:2924. [38] Arzt E. Acta mater 1998;46:5611. [39] Holzapfel B, et al. IEEE Trans on Applied Superconductivity 2001;11:3872.