UDK 621.3:(53+54+621 +66), ISSN 0352-9045 Informacije MIDEM 22(1992)3, Ljubljana THERMALL Y TREATED COLD SPUTTERED YBaCuO FILMS BY TRIODE SPUTTERING A.Cvelbar, A.Žabkar, P.Panjan, S.Navinšek, M.Ambrožič , E.Karič, J.Gasperič KEY WORDS: superconductive films. Y8aCuO films, MgO substrates, amorphous films, cold sputtering, triode sputtering, thermal treatment, experimental research, experimental results ABSTRACT: Superconductive Y8aCuO films were sputtered onto (100) MgO substrates in a triode system. Stoichiome!ric target was used in reactive DC mode. Low temperature depositions were followed by various heat treatments in nitragen and/or oxygen atmosphere. In the as deposited state the film s were amorphous. Fast heating in nitrogen caused grains to grow up to 2 flm while oxygen would reduce the grain growth to average size below 0.4 flm. Raman spectra as well as XRD confirm the presence of orthorombic Y8a2CU307 phase but indicate also !race s of BaCu01. The transition temperature to superconducting state (R=O) was 80 K. Films annealed in nitrogen were porous. TOPLOTNO OBDELANE HLADNO TRIODNO NAPRŠENE YBaCuO PLASTI KLJUČNE BESEDE: superprevodne plasti, Y8aCuO plasti, MgO substrati, amorfne plasti, hladno naprševanje, triodno naprševanje, toplotna obdelava, eksperimentalne raziskave, eksperimentalni rezultati POVZETEK: Superprevodne Y8aCuO plasti smo napršili na (100) MgO podlage v triodnem sistemu. Uporabili smo stehiometrično tarčo in enosmeren način naprševanja. Nizkotemperaturni depoziciji se sledile različne toplotne obdelave v dušiku in/ali kisiku. Napršene plasti so bile amorfne. Hitro segrevanje v dušiku je povzročilo rast zrn do velikosti 2flm, v kisiku pa je znašala povprečna velikost zrn le do O.4flm. Ramanski spektri, kot tudi rentgenski spektri potrjujejo prisotnost ortorombske Y8a2Cu3D7 faze ob sledovih 8aCu01. Temperatura prehoda v superprevodno stanje (R=O) je znašala 80 K. Plasti pregrete v dušiku, so bile porazne. INTRODUCTION Since the discovery of the superconductivity at tempera- tures above liquid nitrogen boiling point in some ceramic materials in 1986/11 an enormous amount of research efforts was put into both the scientific understanding and application of the phenomenon. Several thin film deposi- tion methods for high temperature superconducting (HTSC) materials have been established, including PVD techniques (sputtering, evaporation, laser ablation), all giving films with superconductive transition temperature around 90 K. The superconducting quality of films made by different methods is similar because the postdeposi- tional annealings are usually similar. However, the sig- nificance of annealing details (temperature, time, heat- ing and cooling rates, atmosphere, etc.) has been gener- aIly accepted although not well understood. Usually, the deposition method is chosen depending on the available facility and then optimized. The triode sputtering system is known for its ability of transferring multicomponent materials from the target to the substrate without changing the composition. There- fore it should be adequate for sputtering from the stoi- chiometric YBa2Cu307 target. Additionally, due to large target to substrate distance, such arrangement is effi- cient when one wants to avoid the detrimental effect of 180 negative ions du ring the deposition of YBaCuO films. Triode sputtering system has already been used by the Phillips group 12,31 who successfully prepa red Hall structures with 0.25 ~m YBa2Cu307 thin films on SrTi03, while their films on other substrates were reported to be of inferior quality. Several materials have been considered for potential use as a substrate. The most important properties in- clude the structure (lattice constants), temperature coef- ficient of expansion, dielectric constant and the sensitiv- ity to film-substrate interface reactions. So far, the best results have been obtained on expensive substrates such as SrTi03 or LaAI03. Somewhat less succesful were films on cheaper MgO substrates. Regarding the temperature expansion, MgO substrate may be an ap- propriate choice with its temperature coefficient of 13*10-6 K1 while YBa2Cu307 is anisotropic having the corresponding temperature coefficients between 10 and 15*10-6 K1 for a and b directions of the unit celi, respec- tively 141. At present the majority of research groups are dealing with in situ PVD thin film preparation methods. These methods give better film quality (controlled orientation, epitaxy, high Je and Teo, weaker film-substrate reac- tion) and are faster than two step ex situ methods. Their ~ rJO C aJ C Informacije MIDEM 22(1992)3, str. 180-184 drawback is, on the other hand, that they are limited to: rather small areas with homogenous temperature. Ex situ methods will offer potentially larger deposition areas and larger substrates if adequate deposition and anneal- ing conditions are defined and fulfilled. ln 1989 Nagata and his group /5/ have noticed that rapid heating of YBCO thin film in nitrogen atmosphere acts in favour of crystal grain growth. The aim of this study was to confirm this occurance and determine the in- fluence of the atmosphere, heating rate and temperature during film annealing and oxidation after deposition onto (100) MgO substrates. EXPERIMENT AL DET AILS All depositions were performed in a plasma beam ap- paratus Sputron (Baizers, Liechtenstein) which permits DC as well as RF operation in inert or re active atmos- phere. The system geometry with its 15 cm target to substrate distance is similar to that during evaporation and minimizes the detrimental effects of negative oxygen ions /6/. Both home made /7/ and original Balzers targets with nominal stoichiometric composition were used in our experiments. According to Balzers specification, the source contains 99.9% of the superconducting phase and very little impurities (less than 250 ppm Sr, 30 ppm Na, 25 ppm K, 10 ppm Ca; Al, Fe, As, Ni, Zn below 5 ppm). Control Raman spectra revealed excellent target homogeneity and possible presence of BaCu02 phase. Fig.1 shows Raman spectra from Balzers and home made targets. The oscillations are due to surface rough- ness, nevertheless cie ar peaks at 342 and 505 cm-1 can be found. The home made target was less homogene- ous than Balzers target and gave poorer deposition repeatability. Its Raman spectrum is presented on figure 1b. 120 I a 110 100 SO~) 90 "" f~"'" ~ ~ 80 IV ~\\ r~~rll ~~ 70 !\ 1~\ I ,! I~ 1 60 ~~~WII 50 T ·-------r-- ------, ---"-.-"~ - r----- ! --T--- ,~ rJO c aJ C A. Cvelbar, A. Zabkar, P. Panjan et al.: Thermally treated cold sputtered YBaCuO films by triade sputtering Stoichiometric targets were sputtered in argon DC plas- ma (800 V, 0.65 A, 5*10-3 mbar). To obtain reproducible results, we cleaned the targets through presputtering for 10 min before each deposition. One possible source of compositional chang es in the film can be in the segre- gation on the target surface. To stabilize the target surface, 5*10-5 mb ar of oxygen was introduced into the working chamber. This can hardly be considered as re active deposition since "ordinary" re active process would require about ten times higher oxygen partial pressure /8/. 1 J..lm thick films were sputtered at 2.5 nm/min deposition rate onto 100°C substrates. Sapphire, Zr02, SrTi03 and MgO substrates were used. However, we present here only the results obtained with (100) MgO crystals. Both polished and air cleaved sur- faces were used. After polishing the substrates were cleaned in mild detergent, distilled water and etanol. Several postdepositional annealing procedures were tested. The samples were treated in 1 atm flowing nitrogen and/or oxygen following the experience of Na- gata et al. /5/. The treatment can be divided into three (or four) stages: heating up, annealing at high tempera- ture and cooling down (with or without additional oxida- tion above 450°C).ln the warm up stage, heating rates between 1 and 10°C/min were used. Annealing tem- perature was varied between 750 and 850°C. Tests were made with annealing times between 1 and 60 min and 10 min annealing was chosen as optimum. Slow cooling down rates of 1°C/min in oxygen were used to improve the oxidation process and the transition from tetragonal to orthorombic phase. In early runs the cool down stage included also the 7 h oxidation at 450°C, whereas later this step was omitted. Gold contacts were sputter deposited onto the heat treated films and wired with the conductive paste (Epo- tek, Poliscience AG). Conventional DC four point probe was used for measuring electrical characteristics. The temperature was measured with platinum thermometer in the range 300 to 25 K and varied slowly (0.5 to 1 160 ---- 150 140 130 120 110 100 90 80 70 60 SOl b sos A'(~' o'"~ J ~y ~~I 1 ~ V1J~ ltl~t\V vj\~ ~ ~ l' li ) ~ ! li: \ " I . 1'1\1' I ! 111\vV , " , , I o 200 400 600 800 1000 1200 1400 o 200 400 600 800 1000 1200 1400 Energy (cm-1) Energy (cm-1) Fig. 1: Raman spectra of Balzers (a) and home made (b) target. 181 q vl c ?!. A. Cvelbar, A. Zabkar, P. Panjan et al.: Thermally treated cold sputtered YBaCuO films by triade sputtering KImin) to compensate the influence of poor thermal contact between the sample and the thermometer and reduce the hysteresis around the critical temperature. Films were analyzed before and after heat treatment using several techniques. Raman light scattering was measured in triple spectrometer with carefully nor- malized response of the detector. Argon ion laser (A = 514.5 nm) was used with 40 mW beam (30 f.lm in diameter). X-ray diffraction was made in the Enraf Nonius 591 apparatus with 40 kV / 26 mA generator and Huber thin film diffractometer having a Seemann-Bohlin geometry. The wavelength of the CuKa source was 0.154 nm). 450 400 350 300 250 200 150 100 50 o O Fig. 2: ~)I 1400 Ener- - - ' : ,-' ~ (v') ..- o '(V')' §s;2 I <:t i NIO I o', o,, lil bru, \'.'1/' " .... .-' ; .. 1 1, <:tU') ..-0 I J oo I. li o N l o <.o' o II o X-ray diffraetion of 111m thin yaCO film on MgO, heated up and annealed for 10 minutes in nitrogen and eooled in oxygen at a rate 1°C/min. Peaks marked by "a" are aaCu~ peaks. 182 Informacije MI DEM 22(1992)3, str. 180-184 Re.l.tlvlty (srb.unlt.) 6 3 o -' 70 100 Fig. 4: 130 150 190 220 Temper at ure (K) 250 280 Resistivity vs. temperature for 11lm YBCO film onMgO cording to X-ray diHraction peaks (O 13), (103) and (110) the films are randomly oriented. Peaks around t} = 13° belong to BaCu02 phase. In some cases poor correla- tion was found between the Raman spectrum and the resistance characteristic. The reason for this lies in the diHerence of the signalorigin. Light is scattered from the Fig. 5: Eleetron mierographs of surfaee replieas from 1 mm thiek films completely treated ln oxygen (a), and heated in nitrogen, then treated in oxygen (b). 183 A. Cvelbar, A. Zabkar, P. Panjan et al.: Thermally treated cold sputtered YBaCuO films by triode sputtering surface while the, whole thickness is important for resis- tivity. Various heat treatments result in very different tempera- ture characteristics. Annealing in oxygen with additional oxidation at 450°C resulted in relatively broad transition as measured through temperature dependence of resis- tivity. The resistivity began to drop at 80 K with R = O below 25 K for the sample which was heated at 2°C/min in oxygen. Negative temperature coefficient of resistivity in the normal state indicated the presence of tetragonal phase. Rapid heating (10°C/min) of the sample in ni- trogen moved the transition start pOint between 85 and 90 K and end point (R = O) to 60 K. Also, the temperature coefficient became positive in the normal state. In con- trast to many authors who reported the benefits of oxidation above 400°C for several hours, our experience Fig. 6: SEM mierograph of YBCO film on MgO, annealed at 8scfc in nitorogen and eooled down in oxygen. Intenslty (srb.unlta) 115 --------.-.. --.-------.- .. ----- 14 13 12 11 10 9 8 7 6 5 -4 !~~~~~~~~~~~~~~-~;;~=~~~J 78 Fig, 7: 88 98 108 118 128 138 Etching time (min.) _ ... Mg -+- Cu .... -. Ba --U- O -~- Y AES film depth profile of YBCOIMgO junetion after 8scf C annealing in nitrogen and cool down in oxygen. A. Cvelbar, A. Zabkar, P. Panjan et al.: Thermally treated cold sputtered YBaCuO films by triode sputtering suggests to abandon this step - the slow cooling (1 DC/min in oxygen) alone. moved the transition end point to 80 K. Fig.4 shows temperature dependence of resistivity for different annealing procedures. Surface replicas were taken to look at the grain size. For the films that were heat treated in oxygen only the grain size was in the range around O.4!-tm, while fast heating in nitrogen caused the grains to grow up to 2!-tm, on the average, as illustrated in Figs. 5a, 5b and 6. Optical inspection of films uder microscope revealled, that films heated in nitrogen were porous. In some cases the nucleation process was so strong, that a 1!-tm film thickness was not satisfactory to enable full substrate coverage after the furnace annealing process. On the other hand, films full time annealed in oxygen were homogenous what is in agreement with above men- tioned differencies in grain sizes. CONCLUSIONS Various heat treatments were tested for Y-Ba-Cu-O films sputter deposited onto (100) MgO substrates. Raman light scattering spectra and X-ray diffraction patterns were taken for heat treated films together with temperature dependence of resistivity. As deposited films were amorphous and insulating (at room tempera- ture their resistivity was in MD range). Rapid heating (10°C/min) in nitrogen enhances the grain growth to about 2 !-tm compared to 0.4 !-tm when heated up in oxygen. The taken spectra indicate the dominance of orthorombic YBa2Cu307±O.1 phase but also the presence of BaCu02 phase. The resulting resistivity charac- teristics were improved considerably. Our best films had critical temperature around 85 K and transition width of 184 Informacije MIDEM 22(1992)3, str. 180-184 5 K. The temperature coefficient of resistivity in normal state was positive. REFERENCES 1) J.G.Sednorz and KA Mueller, Z.Phys. a.2.4 (1986), 189 2) S,Dam, HAM.van Hal and C,Langereis, Europhysics Leti. ~ (5) (1988),455 3) JW.C,de Vries, S,Dam, M.G.J.Heijman, G.M.Stollman and MAM.Gijs, AppI.Phys.Lett. 52. (22) 1988, 1904 4) T.Aida, T.Fukazawa, T.Takagi and K.Miyauchi, Jap.J.AppI.Phys. .2&(1987),1489 5) H.Nagata, E.Min, M.Aihara, T.ltoh and H.Takai, Physica C ili (1989),66 6) S.M.Rossnagel and J.J.Cuomo, (unpublished) 7) J.Gasperič, R.Slinc, S.Sernik, P.Panjan and E.Karič, 2nd Polar Solids Ceramic Superconductors Conference, 1988, Imperial Col- lege, London 8) A.Žabkar, S.Navinšek and P,Panjan, IJS Report DP-5408, Ljubljana, 1989 9) R.M.MacFarlaane, H.J.Rosen, E,M.Engler, R.D.Jacowitz and V.Y.Lee, Phys.Rev . .8.JJl. (1988),284 Prispelo: 02.09.92 A. Cvelbar, dipl. ing. el. A.Žabkar, dipl. ing. fiz. P. Panjan, dipl. ing. fiz. B.Navinšek, prof. dr. dipl. ing. el. M.Ambrožič, dipl. ing. fiz. E.Karič, dipl. ing. fiz. J. Gasperič dr. dipl. ing. el. Jožef Stefan Institute, Jamova 39, 61111 Ljubljana, Slovenia Sprejeto: 21.09.92