Critical current characteristics of composite thin films of Au and YBa2Cu3O7

Critical current characteristics of composite thin films of Au and YBa2Cu3O7

Physica C 313 Ž1999. 11–20 Critical current characteristics of composite thin films of Au and YBa 2 Cu 3 O 7 E.J. Cukauskas ) , Laura H. Allen NaÕal ...

363KB Sizes 0 Downloads 17 Views

Physica C 313 Ž1999. 11–20

Critical current characteristics of composite thin films of Au and YBa 2 Cu 3 O 7 E.J. Cukauskas ) , Laura H. Allen NaÕal Research Laboratory, Electronics Science and Technology DiÕision, Washington, DC 20375, USA Received 28 September 1998; accepted 16 December 1998

Abstract Thin film composites of YBa 2 Cu 3 O 7 ŽYBCO. with gold were deposited by inverted cylindrical magnetron sputtering on SrTiO 3 ŽSTO., MgO, and yttrium stabilized zirconium ŽYSZ. substrates using a bilayer deposition process. The material and electrical properties of crystal structure, surface morphology, resistivity, transition temperature ŽTc ., resistance ratio, and critical current Ž Ic . as a function of temperature and small magnetic fields were investigated. Films deposited on STO ˚ of gold with 1 kA˚ of YBCO; for YSZ and MgO, the showed little degradation in superconducting properties up to 500 A ˚ of gold caused the films to become insulating; films showed a rapid fall-off in Tc with gold thickness. For YSZ, 3–10 A ˚ the films were superconducting but Tc rapidly fell to zero by 500 A˚ of gold. however, for gold greater than 25 A, ˚ of gold. The temperature dependence of Ic Composites on MgO showed the most rapid fall-off in Tc going to zero by 200 A was predominately quadratic. However, some films showed a crossover between quadratic and linear temperature dependence in Ic . The Ic of all the composites showed a response to small magnetic fields with the films on STO being the least sensitive and those on MgO the most. There was a trade-off between Tc and field response with the film having the greatest response having the lowest Tc . These composite films show potential for vortex flow device development. q 1999 Published by Elsevier Science B.V. All rights reserved. Keywords: Critical current; Au; YBa 2 Cu 3 O 7

1. Introduction High-temperature superconducting ŽHTS. electronics is in need of a reproducible and environmentally-stable active three-terminal device with gain greater than unity. There are a number of transistorlike superconducting devices that use the controlled motion of magnetic vortices as the basis of their )

Corresponding author. Naval Research Laboratory, Code 6863, Electronics Science and Technology Division, Washington, DC 20375, USA. Tel.: q1-202-767-3247; Fax: q1-202-76784290; E-mail: [email protected]

operation w1,2x. In 1981, Rajeevakumar demonstrated a three terminal vortex flow device with a current gain of 5 using a long Josephson junction w3x. The propagation of fluxons in a parallel array of weak-link Josephson junctions first reported by Hontsu and Ishii, the fluxonic junction transistor proposed by Kadin, and the long Josephson junction biased in the flux flow mode reported by McGinnis et al. are three candidate devices w4–9x. The principle of operation for these devices is based on the controlled motion of magnetic vortices: Abrikosov vortices in a singlelayer thin film Žthe first two. or Josephson vortices in a multiple-layer tunnel junction Žthe latter.. Each

0921-4534r99r$ - see front matter q 1999 Published by Elsevier Science B.V. All rights reserved. PII: S 0 9 2 1 - 4 5 3 4 Ž 9 8 . 0 0 7 1 0 - 2


E.J. Cukauskas, L.H. Allen r Physica C 313 (1999) 11–20

approach has advantages. A single-layer device offers ease and simplicity of fabrication. However, Abrikosov vortices have relatively low velocities, about 10 3 mrs w10x. Josephson vortices offer higher device speeds, perhaps as much as 10 4 times greater because they have no core drag w11x. We have been investigating the material and electrical properties of superconducting composite materials that combine the best of both approaches: a single-layer film which has the properties of a random two-dimensional array of Josephson-coupled grains. We have been developing several thin film composites of YBCO with either an immiscible insulator or noble metal. Our goal has been to make thin films with superconducting grains of YBCO embedded within a background matrix of an insulating or noble-metal material. By adjusting the thickness of the material between the superconducting grains, the coupling between the grains can be changed. Substrate choice will significantly affect the grain boundary morphology because of differences in the lattice match to the components of the composite film. Dimos et al. w12x used artificial grain boundaries to show that critical currents in YBCO films are diminished for high-angle grain boundaries. Using high-resolution STEM, Browning et al. w13x have shown that these high-angle grain boundaries are also non-uniform. By using substrates with various degrees of lattice matches to YBCO, one can affect the critical currents by incorporating material within these narrow and wide angle grain boundaries and thus forming a composite thin film. We have had some encouraging results in our thin film composite materials research. For example, cosputtered films of YBCO with CeO 2 have transport properties which are consistent with the presence of Josephson vortices w14x. We have also used a bilayer deposition technique to grow single-layer, composite films of YBCO with noble metals of Ag or Au. For the Ag composites, we found that increasing the Ag content decreased the critical current density by nearly 98%, while still maintaining a high transition temperature w15,16x. Our initial studies of AurYBCO composite films showed that substrate choice significantly affected the transport properties and magnetic field response of the films grown on them w17,18x. In this paper, we report new transport and magnetic field response of the AurYBCO composite films

with greater Au content and weaker coupling between the superconducting grains for several substrates.

2. Film growth The AurYBCO composite films were grown using a bilayer technique, depicted in Fig. 1. First, a layer of Au was deposited at ambient temperature onto cleaned substrates ŽFig. 1a.. Next, without exposure to the ambient environment, the gold coated substrates were transferred to another chamber for the YBCO deposition. The substrates were heated to 790 or 8008C, which was sufficiently hot for the Au film to coalesce into ‘island-like’ regions on the substrate surface ŽFig. 1b. Then the YBCO layer was deposited, filling in between the Au ‘islands’ which extend above the YBCO surface, as islands rise above the ocean’s surface ŽFig. 1c.. The Au layer was sputtered using conventional ˚ on-axis sputtering in 1.5 Pa of Argon at 3.6 Ars. The YBCO was reactively sputtered with an Inverted Cylindrical Magnetron sputtergun in a 1:1 oxygen: argon gas mixture and a total pressure of 53 Pa. More details of the deposition system and techniques are given elsewhere w17,18x. Except for a few samples discussed in the morphology section, the films ˚ were grown at 7908C, the YBCO layers were 1 kA thick, and the Au layer thickness Ž dAu . was between

Fig. 1. The schematic representation of the bilayer process used for the growth of the composite thin films. Ža. A layer of Au is deposited at ambient temperature. Žb. The substrate is heated to 8008C, and the Au coalesces into ‘puddles’. Žc. The YBCO is deposited, filling in the area between the puddles.

E.J. Cukauskas, L.H. Allen r Physica C 313 (1999) 11–20

˚ The other films, referred to in this paper 0 and 1 kA. as the ‘thicker’ films, were grown at 8008C, having 2 ˚ layers of YBCO, and dAu between 500 A˚ and 2 kA ˚ kA. All the films were grown on 6.4 mm square substrates of Ž100.-oriented STO, MgO, and YSZ. With STO, the lattice match to the YBCO Ž100. and Ž010. directions is very close, within 2%. With MgO, the lattice mismatch for these directions is about 8 or 9%. So, YBCO films on STO will have significantly less strain than those on MgO. With YSZ, the situation is more complicated, and the alignment of the YBCO crystal directions would be difficult to predict. The Ž100. direction of YSZ is a 25% mismatch for these YBCO directions, but the Ž110. direction is a y5% mismatch. Some experimental work using X-ray diffractometry ŽXRD. has been done by Zheng et al. w19x in which they found more rotated YBCO grains on YSZ substrates than on MgO, suggesting that YSZ films had more strain. 3. Film morphology To study the microstructure of the composite films, u –2 u scans were taken between 5 and 958 using copper K a radiation from an automated X-ray

˚ Fig. 2. u –2 u XRD scans for composite films on STO with 2 kA ˚ of gold for the upper trace and no gold for the YBCO and 1 kA lower trace. The open circles are impurity peaks and the stars are substrate reflections described in the text.


Table 1 ˚ for the YBCO grains in the c-axis lattice parameter in A YBCOrAu films Ts Ž8C.

dAu ˚. ŽA








0 500 1000 2000

11.695 11.689 11.696 11.693

11.708 11.697 11.696 11.703

11.701 11.703


6 25 75 125

11.713 11.709


11.675 11.678


diffractometer. Fig. 2 is a logarithmic plot of u –2 u ˚ thick YBCO film Žbottom trace. scans for a 2 kA ˚ Au with 2 kA˚ YBCO composite film and a 1 kA Župper trace., both grown at 8008C on STO substrates. The intensity scale is the same for each scan, and the upper trace is offset for clarity. Also, for clarity, the large substrate K a peaks are not shown, but other substrate reflections from Kb radiation and the tungsten filter are marked by stars on the upper trace. No peaks were observed that indicate a mixing or compounding of the YBCO with the Au. The only YBCO peaks present are Ž00 l . reflections, which are marked by their order number on the lower trace. ŽThe l s 3, 6, and 9 peaks are under the STO substrate peaks.. Au peaks are seen in scans of the composite films. As indicated on the upper trace, these are predominantly the Ž111. and Ž222. reflections, which are much larger than the YBCO Ž005. and Ž0010. peaks and slightly shifted to smaller 2 u values. A few small impurity peaks occasionally are seen in the spectra. They can be identified as Y2 O 3 Ž100. lines and Y2 BaCuO5 Ž001. or Ž113. lines. The upper trace in Fig. 2 contains two or three impurity peaks, which are marked with open circles. The X-ray traces for films on MgO and YSZ are similar. As dAu increased, the size of the YBCO peaks diminished for all the substrates. C-axis peaks were still found for the thickest Au layers where there was too much Au to coalesce and expose the substrate for the YBCO to nucleate. Also, in addition to the Ž111. Au reflections, other Au reflections become significant for these thick gold films: the Ž220., Ž311., and possibly the Ž200..


E.J. Cukauskas, L.H. Allen r Physica C 313 (1999) 11–20

When bulk composites of Au and YBCO were grown, the c-axis lattice parameter Ž c 0 . was ob˚ w20,21x. served to expand from 11.68 to 11.72 A Cieplak et al. w20x explained this by using neutron diffraction to show that small amounts of Au were incorporated into the YBCO crystal structure. To determine if this was occurring in our composite films, we located the centers of as many c-axis peaks as possible in the u –2 u scans, and used a least

squares fit to extrapolate the calculated parameters against a cos 2 ursin u error function w22x. Table 1 summarizes the c 0 results determined for a selection of composite films on the different substrates. The ˚ so only variations error in the data is "0.005 A, ˚ are deemed significant. A sysgreater than 0.01 A tematic study of the effects of dAu on the crystal structure was made for a number of films grown at ˚ layer of YBCO on STO, MgO, 8008C with a 2 kA

˚ on STO, Žb. 500 A˚ on MgO, and 700 Fig. 3. SEM micrographs of three YBCOrAu composite films for three thicknesses of gold, Ža. 300 A ˚ on MgO. A

E.J. Cukauskas, L.H. Allen r Physica C 313 (1999) 11–20

and YSZ substrates. A significant expansion in c 0 was not seen for any of the substrates. Additional ˚ YBCO layer films were grown at 7908C with a 1 kA and thinner Au layers. Although only a few of these films were selected to be studied, however, a few conclusions can be drawn. First, these films also do not appear to have incorporated Au into the YBCO lattice. The composites grown on STO at 7908C do ˚ larger than for have a c 0 that is about 0.015 A composites grown on STO at 8008C, but this value is still smaller than what was reported for the bulk composites w20,21x. This was unexpected, but perhaps may be explained by the lower deposition temperature. The trend is opposite for YSZ with c0 ˚ smaller for the YSZ composites grown about 0.025 A at the lower temperature than for those grown at 8008C. Perhaps YBCO reacts with YSZ at 8008C or the 25% Ž100. mismatch and the more rotated YBCO grains on YSZ offset the O 2 incorporation into the YBCO grains for composites on YSZ. A Hitachi scanning electron microscope ŽSEM. was used to take micrographs of the surface of several of the composite films. Fig. 3 is a series of these micrographs showing what happens as dAu is increased. Using the attached energy dispersive X-ray analyzer, the white areas of the micrographs were identified as Au and the dark areas as YBCO. For ˚ the Au coalesced into round regions, as dAu - 300 A, ˚ Au composite on seen in Fig. 3a, which is a 200 A STO. The YBCO grains between the Au islands look like the c-axis grains of our plain YBCO films. ˚ of Au, the islands became Between 300 and 500 A ˚ Au composite on connected, as in Fig. 3b, a 500 A ˚ the Au is MgO. With dAu ’s of 700 to 1000 A, completely interconnected, and the YBCO filled the ˚ Au pockets in between, as is seen for the 700 A composite on MgO shown in Fig. 3c. The films of Fig. 3b,c appear to be random 2-dimensional arrays of superconducting regions with channels of normal metal ŽAu. crossing them. At first, this appeared to be exactly what we wanted to achieve; normal-metal channels where Josephson vortices could easily move, but superconducting paths connected by tunneling transport currents. Occasionally, unusually large-scale features were seen in SEM micrographs of some composites which could have been superconducting areas channeled by non-conducting lines, similar to the Au and YBCO


regions of Fig. 3b, but on a much larger scale. This feature could be associated with the extra peaks observed in the X-ray scans. These impurity peaks could be from the non-conducting material between the superconducting YBCO grains. A composite film was cracked and a cross-sectional SEM was attempted to see how the noble metal structures developed within the thickness of the film. It was too difficult to get a good view of how slanted the sides of the Au islands were, but it was discovered that the Au islands were several times higher than the YBCO grains in the background.

4. Transport measurements The transport properties of the composite films were measured in a closed cycle helium refrigeration system which had been converted from a 500 lrs cryopump. The samples were heat sunk to an aluminum block attached to the cold stage of the refrigerator by pressing them into a 250 mm thick piece of indium foil. The aluminum block was cylindrically symmetric and small enough to warm or cool rapidly to quickly stabilize the temperature. Four series connected resistors of equal value were heat sunk by indium foil to the underside of the aluminum block to serve as heaters for a temperature controller. Mounted symmetrically to the sample was a silicon diode thermometer, used to monitor the block temperature. Accuracy of the thermometer and transport data was insured by checking for agreement with other measurements made in a liquid nitrogen cooled cryostat and in another refrigerator. Agreement within the tolerance of the silicon diode thermometers was observed. 4.1. As-grown film characterization The thin film composites were characterized by measurements of several of their superconducting and normal state properties. Silver contact pads were evaporated in a square array through a shadow mask onto the films. The square-probe array geometry technique was used to obtain the resistance Ž R . as a function of temperature ŽT . for the as-grown films. This method allows for the accurate determination of

E.J. Cukauskas, L.H. Allen r Physica C 313 (1999) 11–20


Fig. 4. The as-sputtered zero resistance Tc as a function of gold thickness for the three indicated substrates. The YBCO thickness ˚ is 1 kA.

the film resistivity where probe spacing error exists w23x. From the R–T traces, we obtained the resistance ratio ŽRR., defined as the ratio of the room temperature resistance at 295 K to the resistance at 100 K, the 50% Tc which is the temperature where the resistance is half its normal-state value, 0% Tc which is the temperature where resistance becomes - 0.1 m V, and DTc which is the temperature difference between the 90% and 10% values of the normal-state resistance. The zero resistance Tc of a number of films grown on the various substrates is illustrated in Fig. 4 as a function of gold thickness. The insulating films on YSZ made with small amounts of gold are not included for clarity. Some relevant film parameters for the as-sputtered films are tabulated in Table 2.

Table 2 lists several of these properties for a representative selection of the films for each sub˚ thick YBCO layer. strate. All the films had a 1 kA For all of the substrates, as dAu is increased, the RR falls from about 3 to nearly 1 because the normalstate conduction is dominated by scattering from the grain boundaries. Table 2 also shows the effects of increasing dAu on the superconducting transition. The 50% and 0% Tc falls from near 88 K, and DTc broadens. Also, the 50% Tc drops much slower than the 0% Tc , indicating that the Tc of the grains remains high, and the resistance of the film develops a tail that elongates with increasing dAu . We believe that the tail results from a random distribution of coupling strengths between the superconducting regions of the composites. This is due to the randomness of the gold islands forming a 2D Swiss cheeselike morphology where the gold islands correspond to the holes or normal regions in the background YBCO. The scatter in the data, for example the variation in the RR’s with dAu for each substrate, may also be related to this randomness in coupling strength. It can be observed from Table 2 that each substrate has a different dAu where this change in transport properties occurs. For example, composites grown on STO, maintain a high RR and a high, ˚ However, the sharp Tc out to a dAu of 500 A. properties of composites on MgO fell-off immedi˚ producing a noticeable ately, with even a dAu of 3 A effect. The trends for the composites on YSZ were between the STO and MgO results. As mentioned

Table 2 Transport properties of selected as-sputtered AurYBCO composite films grown on STO, MgO, and YSZ Substrate

˚. dAu ŽA


50% Tc ŽK.

0% Tc ŽK.


r 300 K ŽmV cm.


0 12.5 50 125 200 300 0 6 25 50 0 100 200

2.94 2.90 2.62 2.77 2.84 2.74 3.13 1.85 2.10 1.91 3.00 2.37 1.90

86.3 87.7 87.1 87.9 87.2 87.6 86.2 83.7 82.2 84.5 88.0 88.7 85.9

85.4 87.0 85.5 87.3 86.7 87.3 85.0 54.0 55.0 52.0 87.5 55.0 45.0

0.8 0.3 0.7 0.7 0.5 0.6 0.9 16 16 13 0.8 3.4 2.1

477 406 1210 510 698 884 450 807 1120 783 1110 935 2450

E.J. Cukauskas, L.H. Allen r Physica C 313 (1999) 11–20

above and not included in the table, there is a region ˚ - dAu - 25 A˚ ., for which of small Au thickness Ž6 A the films were insulating. As Table 2 shows, how˚ the transport properties of ever, for dAu s 100 A, composite films on YSZ recover somewhat and then fall-off again with greater dAu as do the films on STO and MgO. 4.2. Critical current measurements A selected number of films were patterned into I-beam shaped bridges 1.5 mm in length and a width of either 50 mm, 400 mm, or 1 mm. Standard photolithographic techniques and Argon ion beam milling were used in defining the bridges. All the patterned films experienced some degree of degradation to their superconducting properties as a result of the processing. The resistivity Ž r 300 . at 300 K was calculated from the bridge resistance and dimen˚ . was sions. The thickness of the YBCO layer Ž1 kA used for the film thickness in calculating r 300 . This is a simplification, but it provides a measure of the normal state transport for comparison among the composite films. The current vs. voltage characteristics were measured at various temperatures in determining the temperature dependence of Ic for the Au composite films. The criterion used to define Ic was the current which produced a detectable Ž- 0.1 mV. voltage drop across the patterned sample. This is equivalent to an electric field criterion of E - 7 = 10y7 Vrcm. To obtain the critical current density Ž Jc . from the


critical current, we assumed a homogeneous material and divided by the cross-sectional area. As for the ˚ YBCO thickness was resistivity calculation, the 1 kA used in calculating the Jc of the films. The temperature dependence of the critical current can indicate the nature of the intergranular coupling of polycrystalline films w24–28x. Near Tc , the critical current follows a power-law dependence of Ic A Ž1 y t . n , where the reduced temperature Ž t . is defined as TrTc . For the composite films, we expect to see a departure from the 3r2 power-law Ginzburg– Landau ŽG–L. dependence observed for plain YBCO films to either the linear dependence which characterizes superconductor–insulator–superconductor ŽSIS. coupling or the quadratic dependence of SNS coupling, where N is a normal metal w28x. The critical current was determined as a function of temperature for a number of the composite films. A least squares fit to the data was used in determining the value of n and Tc for the power law dependence of Ic . Table 3 lists the values of n and Tc which produced the best fit to the data and the temperature interval, t Žfit., used in the analysis. Values of n for the films on STO ranged from 1.42 for ˚ Au the plain YBCO to nearly 2.5 for the 12.5 A composite. An additional approach was used to analyze the temperature dependence of the critical current. Using the value of Tc at which the bridge Ic goes to zero to calculate the reduced temperature Ž t ., the log Ž Ic . was plotted vs. log Ž1 y t . from which a value of n was determined from the slope. Here, changes in intergranular coupling with temperature

Table 3 Summary of the Ic ŽT . dependence and the magnetic field response of the YBCOrAu composites Substrate

˚. dAu ŽA


t Žfit. Žlowest t .

Tc Žfit. ŽK.

nŽlow T .

t cross Žcrossover.

t Ž20% M. Ž25 G.


0 12.5 50 125 200 300 0 6 25 50 0 200

1.45 1.72 1.71 1.73 1.88 1.76 2.06 2.06 1.92 1.80 0.94 1.54

0.95 0.93 0.88 0.95 0.83 0.96 0.96 0.85 0.81 0.42 0.95 0.68

84.66 84.77 70.0 87.0 77.47 86.73 84.77 55.26 47.22 42.83 82.64 74.26

1.42 2.46 2.14 1.72 1.95 1.95 1.83 2.08 1.84 1.87 1.0 1.62

None 0.95 0.915 0.997 0.975 0.995 None None None 0.846 None None

0.99 0.954 0.948 0.98 0.95 0.99 0.993 0.69 - 0.4 - 0.4 ; 0.99 0.7


E.J. Cukauskas, L.H. Allen r Physica C 313 (1999) 11–20

show up as changes in the slope at various segments of the plot. Using this technique, the plain YBCO film on STO showed a uniform slope corresponding to a value of n equal to 1.42. On the other hand, all the composite films on STO showed a break in the slope which is labeled as t cross in Table 3. The ˚ of Au is illustrated composite film made with 50 A in Fig. 5 and has two distinct breaks. Only the position of the lower temperature break is listed in Table 3. At the lower temperatures, n is equal to 2.14 corresponding to SNS intergranular coupling and very close to Tc has a value of 0.94 corresponding to SIS coupling. There exist a transition region between these two coupling schemes which is characterized by a value of n equal to 4.0. All the samples which showed the crossover had values of n of the order of 4 in the crossover region which we believe to be a region of transition to a lower n value. Closer to Tc and beyond the range of our measurement capability, the slope may change to a more SIS-like behavior. For the composite films on MgO, the Ic ŽT . data showed a predominantly quadratic dependence on temperature for all gold compositions. This can be associated with the ease in which gold incorporates into the grain boundaries due to the existence of many high-angle grain boundaries for YBCO grown on MgO. As with the other transport properties, the critical current measurements reflected the random variations in the coupling strength which results in a decreased current density for increased gold compo˚ of gold sition. Only the composite made with 50 A

Fig. 5. The temperature dependence of the critical current for a ˚ of gold and YBCOrAu composite film on STO having 50 A showing three distinct slopes. The crossover to the linear region occurs at a reduced temperature of 0.967.

exhibited a crossover or change in coupling from SNS to SIS as Tc was approached. There was no intermediate or transition region as was observed for the films on STO Žsee Fig. 5.. This coupling change for the composites on MgO occurred at a reduced temperature of 0.846 which is significantly removed from the zero resistance temperature. Critical current measurements above t cross are difficult because of the small currents and often these composites can be noisy in this region as Tc is approached. Critical current measurements were also taken on two of the YSZ gold composites. The properties of these composites are between those on STO and MgO as illustrated in Fig. 4 above. Interestingly, the plain YBCO on YSZ has a SIS Ic ŽT . dependence over the range of measurement to below 0.6 Tc . The ˚ gold composite was also characterized and 200 A showed a G–L-like Ic ŽT . dependence having n equal to 1.54. Neither of these films on YSZ showed any sign of a crossover in their intergranular coupling characteristics. These two films were the only two measured because of the difficulty in maintaining sample integrity throughout the ion-milling process. Unlike MgO and STO, YSZ is fragile and susceptible to easy breakage. 4.3. Magnetic field characteristics To study the effects of small magnetic fields on the critical currents of the samples, a solenoid magnet was inserted into the refrigerator vacuum space between the cold stage and the vacuum sidewall. The magnet had a length to diameter ratio of two and was positioned such that the sample was centered in the magnet and the field was applied perpendicular to the surface of the film. Magnetic fields up to 170 G could be applied to the sample before thermal radiation from the copper magnet wire affected the temperature stability of the cold stage. The magnetic field response of selected samples was characterized by making Ic measurements for specific magnetic field values and temperatures. Measurements were taken at applied magnetic fields of 0, 25, 50, 100, and 150 G and over the temperature range used to analyze the Ic ŽT . data. To quantify the field response of the composite films, the critical current modulation Ž M . due to the applied magnetic field was defined and determined

E.J. Cukauskas, L.H. Allen r Physica C 313 (1999) 11–20

for each film measured. The critical current modulation at a given temperature is defined as the difference of the critical currents in zero applied field and a field Ž B . divided by the zero field current. Modulation values ranged from zero to one, where larger values correspond to a greater magnetic field sensitivity of the film. In order to make a comparison of the patterned composite films, the reduced temperature at which the modulation fell to 0.2 for a 25 G applied field is listed in Table 3. This represents the temperature at which the critical current can be changed by 20% by a 25 G applied field. A glance at the table shows that the composites on STO are the least sensitive, those on MgO have good sensitivity, and the YSZ results are between STO and MgO, but incomplete. However, there is also a trade-off between field sensitivity and Tc , with the most sensitive film having the lowest Tc . In spite of their reduced sensitivity, the composites on STO show promise for device development. The composite ˚ of gold shows 20% modulation at made with 12.5 A a temperature of 80.9 K. Sufficient modulation for vortex flow devices by small magnetic fields can be realized at higher temperature for the composites on STO. Further work at reducing the randomness of the gold islands in these composites may lead to more reproducible properties for the development of vortex flow devices. In characterizing the magnetic field properties of the composite films, the modulation as a function of field can also give insight as to the nature of the intergranular coupling. If the modulation of the composites were due only to the motion of Abrikosov vortices, it would quickly go to zero just below Tc . Such is the case for the plain YBCO, however, we observe significant modulation Ž) 0.2. down to t 0.4 for some of the composite films suggesting the existence of other vortex flow processes such as the motion of Josephson vortices. The modulation of the ˚ gold composite on YSZ was investigated in 200 A sufficient detail as a function of magnetic field. Fig. 6 illustrates the normalized critical current, Ic Ž B .rIc Ž0., as a function of magnetic field, where the normalized critical current is equal to Ž1 y M .. The envelop of the data mimics the sinŽ x .rx magnetic field dependence of a Josephson junction critical current. The randomness of the gold islands and the formation of the gold along the grain boundaries


Fig. 6. The magnetic field dependence of the critical current of the ˚ gold composite on YSZ. The reduced temperature is 0.67. 200 A

results in a two dimensional array of Josephson coupled grains which should influence the temperature and field dependence of the critical current. If one takes the apparent minimum at approximately B s 50 G as corresponding to the first zero in the sin Ž x .rx, an estimate can be made for the YBCO grain size. Taking the penetration depth Ž l. of the YBCO ˚ and the first zero corresponding to grains as 2 kA F 0 s 2 L l B, the grain size Ž L. is calculated to be approximately 1 mm, which is not unreasonable for these films. One does not expect to see the true envelop of the Josephson Ic Ž B . because of the randomness in the gold islands and grain sizes in the film but it does suggest the existence of Josephson vortices in addition to Abrikosov vortices. At the lower temperatures where a modulation is still present, the Abrikosov vortices should be pinned and only the Josephson vortices or weakly pinned Abrikosov vortices contribute to the critical current modulation. The nature of the vortex flow characteristics of these composite films is in need of further understanding in order to optimize their properties for device development.

5. Summary and conclusions We have used a bilayer deposition technique to grow composite thin films of YBCO with Au on STO, MgO, and YSZ substrates. X-ray diffractometry showed that the YBCO is well oxygenated, c-axis orientated, and has no gold incorporated into the lattice. SEM micrographs have shown that the gold


E.J. Cukauskas, L.H. Allen r Physica C 313 (1999) 11–20

forms segregated regions or islands in a background of YBCO. The superconducting properties of the as-deposited films show there is a substrate dependent gold thickness beyond which the properties begin to change. For STO this threshold thickness is ˚ and for MgO, very little gold causes beyond 500 A, a decrease in Tc . For YSZ, very small amounts of gold produce insulating films for gold thickness up ˚ after which Tc recovers at about 100 A˚ to about 25 A of gold and begins to decline again with increasing thickness. Processing the composite films into Ic bridges using standard photolithography and ion beam milling resulted in some degradation to Tc . Investigation of the temperature dependence of the critical current showed most films had a quadratic SNS-like dependence. Only the plain YBCO on STO ˚ gold composite on YSZ showed the and the 200 A G–L temperature dependence. Interestingly, the plain YBCO on YSZ showed a linear SIS behavior. One ˚ of gold on STO showed two film made with 50 A distinct breaks in Ic ŽT . with a transition region between them. At low temperature, the transport was SNS-like and SIS-like near Tc . The response of the films to small magnetic fields was characterized as due to Josephson vortices and weakly pinned Abrikosov vortices. The most sensitive films were those with the lowest Tc and greatest lattice mismatch between the substrate and the YBCO. The STO films showed reasonable sensitivity Ž20% Ic modulation. with high Tc but only a limited useful temperature range for device applications. The composites on MgO were sensitive over the entire superconducting range but had Tc below about 55 K. Perhaps the most promising films were those deposited on YSZ which had good sensitivity with a Tc greater than 74 K. The YSZ substrates proved to be fragile and resulted in the partial characterization of only two films. More work remains to be undertaken to gain a better understanding of the transport in these composite films on YSZ. The thin film composites of YBCOrAu show potential for vortex flow device development.

Acknowledgements The authors thank Bill DeSisto for his comments, suggestions, and critical reading of the manuscript

and Bob Gorman for his help in the preparation of the figures. This research was supported by the Office of Naval Research.

References w1x w2x w4x w3x w5x w6x w7x w8x w9x w10x

w11x w12x w13x w14x w15x w16x w17x w18x w19x w20x

w21x w22x w23x w24x w25x w26x w27x w28x

J.E. Nordman, Semicond. Sci. Technol. 8 Ž1995. 681. J. Mannhart, Supercond. Sci. Technol. 9 Ž1996. 49. S. Hontsu, J. Ishii, J. Appl. Phys. 63 Ž1988. 2021. T.V. Rajeevakumar, Appl. Phys. Lett. 39 Ž1981. 439. A.M. Kadin, J. Appl. Phys. 66 Ž1990. 5741. D.P. McGinnis, J.B. Beyer, J.E. Nordman, J. Appl. Phys. 59 Ž1986. 3917. R. Gerdemann, L. Alff, A. Beck, O.M. Froehlich, B. Mayer, R. Gross, IEEE Trans. Appl. Supercond. 5 Ž1995. 3292. Y.M. Zhang, D. Winkler, P.A. Nilsson, T. Claeson, Appl. Phys. Lett. 64 Ž1994. 1153. K. Miyahara, S. Kubo, M. Suzuki, J. Appl. Phys. 76 Ž1994. 4772. S.G. Doettinger, R.P. Huebener, R. Gerdemann, A. Kuhle, S. Anders, T.G. Trauble, J.C. Villegier, Phys. Rev. Lett. 73 Ž1994. 1691. D. Winkler, Y.M. Zhang, P.A. Nilsson, E.A. Stepantsov, T. Claeson, Phys. Rev. Lett. 72 Ž1994. 1260. D. Dimos, P. Chaudhari, J. Mannhart, F.K. LeGoues, Phys. Rev. Lett. 61 Ž1988. 219. N.D. Browning, M.F. Chisholm, S.J. Pennycook, Interface Science 1 Ž1993. 309. E.J. Cukauskas, L.H. Allen, J. Appl. Phys. 80 Ž1996. 5843. M.A. Fisher, L.H. Allen, E.J. Cukauskas, Appl. Supercond. 3 Ž1995. 607. M.A. Fisher, E.J. Cukauskas, L.H. Allen, IEEE Trans. Appl. Supercond. 7 Ž1997. 1. L.H. Allen, E.J. Cukauskas, M.A. Fisher, Appl. Phys. Lett. 66 Ž1995. 1003. L.H. Allen, E.J. Cukauskas, IEEE Trans. Appl. Supercond. 7 Ž1997. 1650. J.P. Zheng, S.Y. Dong, H.S. Kwok, Appl. Phys. Lett. 58 Ž1991. 540. M.Z. Cieplak, G. Xiao, C.L. Chien, A. Bakhshai, D. Artymosicz, W. Bryden, J.K. Stalick, J.J. Rhyne, Phys. Rev. B 42 Ž1990. 6200. D. Veretnik, S. Reich, J. Appl. Phys. 73 Ž1993. 8429. B.D. Cullity, Elements of X-ray Diffraction, 2nd edn., Addison-Wesley, MA, 1978, p. 359. D.R. Zrudsky, H.D. Bush, J.R. Fassett, Rev. Sci. Instrum. 37 Ž1966. 885. S. Greenspoon, H.J.T. Smith, Can. J. Phys. 49 Ž1971. 1350. N.L. Rowell, H.J.T. Smith, Can. J. Phys. 54 Ž1976. 223. J. Clarke, Proc. R. Soc. A 308 Ž1969. 447. J.W.C. de Vries, M.A.M. Gijs, G.M. Stollman, T.S. Baller, G.N.A. van Veen, J. Appl. Phys. 64 Ž1988. 426. A. Barone, G. Paterno, Physics and Applications of the Josephson Effect, Wiley, New York, 1982.