High critical current density and enhanced irreversibility ®eld in superconducting MgB2 thin ®lms

The discovery of superconductivity at 39 K in magnesium diboride offers the possibility of a new class of low-cost, high-performance superconducting materials for magnets and electronic applications. This compound has twice the transition temperature of Nb3Sn and four times that of Nb-Ti alloy, and the vital prerequisite of strongly linked current ̄ow has already been demonstrated. One possible drawback, however, is that the magnetic ®eld at which superconductivity is destroyed is modest. Furthermore, the ®eld which limits the range of practical applicationsÐthe irreversibility ®eld H*(T)Ðis approximately 7 T at liquid helium temperature (4.2 K), signi®cantly lower than about 10 T for Nb-Ti (ref. 6) and ,20 T for Nb3Sn (ref. 7). Here we show that MgB2 thin ®lms that are alloyed with oxygen can exhibit a much steeper temperature dependence of H*(T) than is observed in bulk materials, yielding an H* value at 4.2 K greater than 14 T. In addition, very high critical current densities at 4.2 K are achieved: 1 MA cm at 1 T and 10 A cm at 10 T. These results demonstrate that MgB2 has potential for high-®eld superconducting applications. Three MgB2 ®lms were prepared by pulsed laser deposition at room temperature onto (111) oriented single crystal SrTiO3 substrates. The base pressure before deposition was 3 3 10 2 6 Pa, and deposition took place under 0.3 Pa of argon. The MgB2 target was prepared by pressing and sintering, as described previously. After deposition, ®lm 1 was annealed in an evacuated niobium tube at 850 8C for 15 minutes. Film 2 was annealed in an evacuated quartz tube at 750 8C for 30 minutes and quenched to room temperature. Film 3 was annealed in a tantalum envelope inside an evacuated niobium tube at 750 8C for 30 minutes. Magnesium pellets were included in each tube to prevent magnesium loss. Film 1 showed visible evidence of inhomogeneity in the centre of the ®lm, which could not be excluded from samples used for electromagnetic measurements. The thickness of each ®lm was approximately 500 nm. The chemical compositions of the ®lms were determined by wavelength dispersive X-ray spectroscopy using an electron microprobe at 3 keV to avoid X-ray generation in the substrate. Before annealing, the ®lms had a Mg:B:O atomic ratio of about 1.0:1.0:0.17, the oxygen probably coming from the target or background gas during deposition. The primary differences in the ®lms are the measured ratios of Mg:B:O obtained after annealing. Film 2 had a ratio of ,1.0:0.9:0.7, which corresponds to a relative oxygen concentration approximately twice that of ®lm 1 and 3, as well as higher Mg content. By contrast, ®lm 1 and ®lm 3 show 1.0:1.0:0.4 and 1.0:1.2:0.3 ratios, respectively. We believe that the higher oxygen of post-annealed ®lm 2 came from the relatively poor vacuum (approximately 150 Pa) and the larger volume of the tube used for annealing. X-ray diffraction also revealed that the MgB2 lattice parameter c is larger in ®lm 2 than in ®lm 1, ®lm 3, or bulk samples. Figure 1 presents v±2v scans for ®lms 1, 2 and 3, which clearly show that the (002) MgB2 peak of ®lm 2 is shifted to lower angle. We believe the expansion of the c-lattice parameter in MgB2 is due to oxygen alloying of the boron layer. The transition temperature Tc was determined by both inductive and resistive measurements. The resistive transitions, shown in Fig. 2, have onset values of 36, 31 and 35 K for ®lms 1, 2 and 3 respectively, and fall to zero resistance just above the principal inductive transitions. The room temperature resistivity r(300 K) of ®lm 2, 390 mQ cm, is clearly much higher than that of ®lm 1 (125 mQ cm) or ®lm 3 (60 mQ cm). High-resistivity ®lm 2 also exhibits a slight resistivity increase with decreasing temperature, indicating a resistivity ratio, RR ˆ r…300 K†=r…40 K†, less than 1. By contrast, the RR of ®lm 1 is about 1.5, whereas that of ®lm 3 is 1.3. For comparison, the RR of Nb-Ti alloys (Ginzburg±Landau parameter k < 40) is seldom higher than 1.4 (ref. 6), whereas that of Nb3Sn (k < 30) is typically between 2 and 5 (ref. 7). Good bulk samples of MgB2 can attain RR values of up to 25 and have k < 25 (refs 4 and 5). The magnetization of 3 mm 3 3 mm pieces of each ®lm was measured from 4.2 to 30 K and in perpendicular ®elds up to 14 T using a vibrating sample magnetometer (VSM). Hysteresis curves at 4.2 K (Fig. 3a) close at ®elds very much higher for ®lm 2, at about 14 T, than for ®lm 1 (,6 T) or ®lm 3 (,6.5 T). We were not able to unambiguously determine Hc2 owing to the smooth blending of the superconducting magnetization into the background magnetization of substrate and sample holder. This is in marked contrast to the behaviour seen for bulk samples; these have a signi®cant reversible contribution to the remaining hysteretic magnetic moment above the primary loop closure ®eld. The oxygenrich ®lm 2 has a much higher irreversibility ®eld than ®lms 1 and 3 up to 25 K. Magneto-optical examination of the ®lms showed strong ̄ux shielding with only slightly distorted roof patterns, indicative of macroscopic supercurrents. This validates use of the standard expression for the magnetization of a thin ®lm, Jc ˆ 30DmV 2 r 2 , to extract critical current densities Jc(H,T), where Dm is the width of the hysteresis loop, V < 4:5 3 10 2 6 cm is the ®lm volume, and r ˆ 0:15 cm is the sample half-width. Figure 3b shows that the highest Jc value at 1 T, 4.2 K exceeds 3 MA cm -2 for the low-resistivity ®lm 3. This result is comparable to that of ref. 13. letters to nature