Physica C 469 (2009) 39–43

Contents lists available at ScienceDirect

Physica C

journal homepage: www.elsevier .com/ locate/physc

Sputtered magnesium diboride thin films: Growth conditionsand surface morphology

April O’Brien, Brendon Villegas, J.Y. Gu *

Department of Physics and Astronomy, California State University, Long Beach, CA 90840, United States

a r t i c l e i n f o a b s t r a c t

Article history:Received 18 July 2008Received in revised form 15 October 2008Accepted 26 October 2008Available online 5 November 2008

PACS:74.62.Bf74.70.Ad68.37.Ps81.15Cd

Keywords:Magnesium diborideSputteringSuperconducting filmsAtomic force microscopy

0921-4534/$ - see front matter � 2008 Elsevier B.V. Adoi:10.1016/j.physc.2008.10.006

* Corresponding author.E-mail address: jgu@csulb.edu (J.Y. Gu).

Magnesium diboride (MgB2) thin films were deposited on C-plane sapphire substrates by sputtering pureB and Mg targets at different substrate temperatures, and were followed by in situ annealing. A system-atic study about the effects of the various growth and annealing parameters on the physical properties ofMgB2 thin films showed that the substrate temperature is the most critical factor that determines thesuperconducting transition temperature (Tc), while annealing plays a minor role. There was no supercon-ducting transition in the thin films grown at room temperature without post-annealing. The highest Tc ofthe samples grown at room temperature after the optimized annealing was 22 K. As the temperature ofthe substrate (Ts) increased, Tc rose. However, the maximum Ts was limited due to the low magnesiumsticking coefficient and thus the Tc value was limited as well. The highest Tc, 29 K, was obtained forthe sample deposited at 180 �C, annealed at 620 �C, and was subsequently annealed a second time at800 �C. Three-dimensional (3D) AFM images clearly demonstrated that the thin films with no transition,or very low Tc, did not have the well-developed MgB2 grains while the films with higher Tc displayed thewell-developed grains and smooth surface. Although the Tc of sputtered MgB2 films in the current work islower than that for the bulk and ex situ annealed thin films, this work presents an important step towardsthe fabrication of MgB2 heterostructures using rather simple physical vapor deposition method such assputtering.

� 2008 Elsevier B.V. All rights reserved.

1. Introduction

Among non-oxide superconductors magnesium diboride(MgB2) has a record of high superconducting transition tempera-ture (Tc � 39 K). The discovery of the novel superconductor MgB2

raised great interest for its potential in both science and technol-ogy applications. MgB2 is a good candidate for superconductingelectronic application due to (1) the higher Tc compared to the con-ventional low-Tc superconductors, (2) large critical current density,and large coherence length (�5 nm) [1] compared to the high-Tc

oxide superconductors. However, to realize these advantages inthin film form, several problems related to the materials have tobe resolved. At present, the most important issue in the electronicapplication of MgB2 is to develop the technology needed to pro-duce and manipulate high-quality thin films. High-quality MgB2

thin films, preferably deposited in situ, are necessary for hybridheterostructures. It has been found that it is very difficult to main-tain an environment for the in situ formation of the stoichiometricMgB2 phase due to the extreme difference in the vapor pressures ofMg and B. Therefore, simple physical vapor deposition methods,

ll rights reserved.

such as sputtering and pulsed laser deposition, were not really suc-cessful in fabricating high-quality MgB2 thin films.

Since the discovery of superconductivity in MgB2, thin filmshave been fabricated in many different ways. In the early stages,the pulsed laser deposition method was mostly used to fabricateMgB2 films [2–6]. Subsequently, e-beam evaporation [7], molecularbeam epitaxy [8,9], sputtering [10–14], and chemical vapor depo-sition [15] methods have been explored. Since it is very difficultto maintain a stable environment for the in situ formation of thestoichiometric MgB2 phase, most of the films were made by eitherex situ diffusion of Mg in B precursor film [2,7,15] or in situ anneal-ing [3–6,9,11,14]. The thin films made by ex situ annealing showedthe Tc close to the bulk value (�39 K), but the thin films treatedin situ showed lower Tc in general. According to the pressure–tem-perature phase diagram reported by Liu et al. [16], an extremelyhigh deposition rate of Mg is necessary to get the stoichiometricMgB2 phase in situ. The hybrid physical–chemical vapor depositionmethod has produced thin films with bulk Tc values [17]. Evenhigher Tc (>40 K) than the bulk value was reported in MgB2 filmin which it had a biaxial tensile strain [18].

Sputtering is a relatively simple method for the fabrication ofthin film and has the potential for device application. The mostideal case for the application of MgB2 thin films to electronic

40 A. O’Brien et al. / Physica C 469 (2009) 39–43

devices would be that the as-grown sputtered MgB2 thin filmsdeposited at low growth temperatures have a significant Tc valuewithout a post-annealing process. However, as shown previouslyin other papers, this is still left as a very challenging problem.

To solve this problem, it is necessary to carefully investigate therole of each growth parameter and tune it to optimize the growthprocess. Although there were some previous works where thesputtering method was used, the systematic thorough study ofthe correlation between the growth conditions and physical prop-erties, including transport, magnetic, and surface properties, wasnot done.

In this work, we fabricated MgB2 thin films by a sputteringmethod and investigated the effect of various growth and anneal-ing parameters on their physical properties. We also associated thesuperconducting properties of the MgB2 thin films with surfacemorphology characterized by Atomic Force Microscopy (AFM).Our current work provides an important step towards the fabrica-tion of MgB2 thin films using a rather simple physical vapor depo-sition method, such as sputtering, by studying the correlationbetween the growth parameters and the physical properties. Wepropose a possible solution to the problem that the sputteringmethod contains for the optimized MgB2 thin film growth.

Mg

Mg

B

MgBx

Mg

[Mg/MgBx]8 multilayer

Mg capping layer

B capping layer

2. Experimental

We used a conventional multi-target sputtering system to fab-ricate the MgB2 thin films. The MgB2 thin films were depositedon C-plane sapphire substrates using DC-magnetron sputtering ofMg at 50 W and rf-magnetron sputtering of B at 150 W. The basepressure of the deposition chamber was always lower than2.0 � 10�7 Torr. Before the deposition, the substrates were chemi-cally cleaned using acetone and methanol. Argon (Ar) pressure dur-ing the sputtering was 1.5 mTorr for the deposition at roomtemperature and 0.5 mTorr for the deposition at elevated temper-atures. To improve the uniformity of the sample, the substrate wasrotated during deposition. Due to a low sticking coefficient of Mg athigh temperatures, MgB2 films were deposited either at room tem-perature or at temperatures lower than 200 �C, and were followedby in situ post-annealing at a higher temperature. For most of thesamples, a thick layer of Mg and a thin layer of B were deposited ontop of the MgB2 layer before annealing. A Mg capping layer wasused to compensate for the Mg loss during the high temperatureannealing and a B capping layer was used to prevent the Mg escap-ing from the films at high temperatures. For comparison, some ofthe samples were made without capping layers. After the deposi-tion, the films were heated in situ at a rate of 3.44 �C/s to theannealing temperature. Annealing conditions were varied: temper-ature between 500 and 800 �C, Ar pressure between 2.0 � 10�2 and0.1 Torr, and the annealing time between 5 and 60 min.

Tc was determined from a four-probe resistance measurementor an AC susceptibility measurement. Resistance of the samplewas measured using either a Quick Dipper resistance probe or aQuantum Design Physical Property Measurement System (PPMS-9T). AC susceptibility was measured using PPMS-9T. The surfacemorphology of MgB2 thin films was investigated using Multi-modeTM AFM. The images were taken for the various samples to ex-plore the effects of substrate and annealing temperatures on thesurface morphology. The RMS roughness and grain size were ob-tained and compared among the different samples.

c-Al2O3

MgBx

Sapphire substrate

Fig. 1. Sample structure of the [Mg/MgBx]8 thin films deposited at roomtemperature.

3. Results and discussion

We report the results for two different sets of data. One set con-tains the thin films deposited at room temperature, while the other

set contains the thin films deposited at elevated, but low growthtemperatures (<200 �C).

3.1. Thin films deposited at room temperature

Thin films were deposited by the sputtering pure B and Mg tar-gets at room temperature to be followed by in situ annealing. Wegrew our samples at room temperature to maximize the Mg stick-ing coefficient. In our sputtering condition, the Mg deposition ratewas about 1 Å/s. As-grown samples at room temperature weremetallic down to 1.9 K and did not show superconducting proper-ties. Only after the as-grown films were annealed at higher tem-peratures were the superconducting transitions detected. Wehave investigated the effects of the sample structure and annealingconditions on the superconducting properties. The highest Tc wasobtained for the films in situ annealed at 620 �C with Ar pressureof 100 mTorr giving a Tc of 22 K.

Several different types of structures were used to optimize thegrowth of MgB2 precursor at room temperature. We fabricated(a) the single layer of co-deposited Mg and B, (b) the bilayer ofMg and B, and (c) the multilayer in which the layer of co-depositedMgBx and the thin layer of Mg alternate. For the multilayer sampleas shown in Fig. 1, Mg and B were co-deposited for 1.5 min to cre-ate the MgBx layers and a very thin layer of Mg was deposited for30 s between each layer of MgBx so as to make up for any Mg lossduring the annealing process. All three types of structures werecovered with Mg/B capping layer for the in situ post-annealing.The thickness dependence of these capping layers will be discussedlater.

The most effective structure was a multilayer of MgBx and Mg.Referring to the Mg:B binary phase diagram [16], a Mg vapor pres-sure between 10�3 and 10 Torr is required to make the MgB2 at atemperature of 620 �C. The Mg layers sandwiched between layersof MgBx are assumed to provide this Mg vapor pressure at hightemperatures. The Tc dropped from 22 to 15 K when we did not de-posit the Mg sandwich layers (this structure is eventually the sameas the single layer of co-deposited Mg and B). AFM data, discussedin a later section, supported the notion that the thin film depositedat room temperature was a mixture of Mg and B, and MgB2 phasewas developed during the post-annealing.

Since the MgB2 phase was not formed at room temperature, werelied on a high temperature annealing process to form the phase.To optimize the annealing condition, we changed the parameters.The annealing conditions we have tested were (a) the tempera-tures between 500 and 800 �C, (b) Ar pressures between 10 mTorrand 0.1 Torr, and (c) the time varying from 5 to 60 min.

The Tc increased as the annealing temperature increased, butthen started to drop above 630 �C. The sample became insulatingwhen the annealing temperature became higher than 650 �C. Atthis temperature Mg would evaporate from the sample. It has been

0 50 75 100 125 150 175 2000

4

8

12

16

20

24

T c (K

)

Ts (ºC)

Mg+B mixture

InsulatingMg-deficient

25

Fig. 2. Tc dependence on the substrate temperature, Ts, before annealing.

A. O’Brien et al. / Physica C 469 (2009) 39–43 41

reported that once the Mg forms MgB2, Mg is not easily re-evapo-rated even at high temperatures, while the pure Mg is re-evapo-rated very quickly [16]. Tc did not show a significant dependenceon Ar pressure. For the time dependence, an annealing time of30 min provided the best quality samples. Annealing time longerthan 30 min did not help to form the MgB2 phase or improve theTc. If the MgB2 phase was not produced, this longer time seemsto only allow the left-over Mg to re-evaporate.

Overall, the annealing temperature was the most critical factorin determining the Tc for the annealing process. Since our filmsdeposited at room temperature were not superconducting at all,there was a Tc barrier of 22 K even after optimizing the annealingconditions. We found that the annealing process improves thequality of the as-grown film by only a limited amount throughdeveloping the MgB2 phase.

We used a Mg capping layer to compensate for Mg loss duringthe high temperature annealing. A very thin B layer was then pla-ted on top of that to protect the Mg from dissipating during anneal-ing. This form of double capping layers is called the CM, or ‘‘capand melt” method [19]. These capping layers are necessary to cre-ate the MgB2 and protect the sample from significant Mg loss athigh temperatures. We varied the thickness of these layers to seetheir effect on the Tc of the sample. The Tc dropped from 22 to14.5 K when we did not add the Mg capping layer. With an in-creased sputtering time of 10 min for the Mg capping layer, theTc dropped to 9 K. Increasing the thickness of the B capping layeralso had a negative effect on the Tc, but it was minimal comparedto the drop caused by excess Mg. The B capping layer with a sput-tering time of 5–10 min produced a Tc of 19 K, and with a sputter-ing time of 20 min produced a Tc of 17 K.

We found that there are several problems with room tempera-ture deposition. For the thin films grown at room temperature, themain problem is that the MgB2 phase cannot be produced, and thusthere is no superconducting transition. Although the post-anneal-ing process produced some MgB2 superconducting phase, we couldonly obtain the Tc of the sample up to 22 K. The other problem wefaced is that samples grown at room temperature are not highlyreproducible. Even when we used exactly the same growth andannealing conditions, the Tc of those samples were inconsistent.Thus, samples grown at room temperature are not ideal for deviceapplications.

We have also observed that each published group has a differ-ent ‘‘recipe” for creating high-quality MgB2 thin films and the ‘‘rec-ipe” is specific to their system. We were unable to achieve thesame result by simply duplicating their recipe. This indicates thatthe superconducting property of MgB2 thin film is very sensitiveto the local growth conditions.

3.2. Thin films deposited at elevated low substrate temperatures(<200 �C)

We grew thin films at elevated substrate temperatures in orderto create the MgB2 phase during the growth. After the deposition,the in situ annealing process was done to increase the Tc. Severalother groups using e-beam evaporation and PLD have reported thatthey could make thin films with high Tc using a low growing tem-perature of under 300 �C [20]. The main problem with elevatedtemperature deposition is that the Mg sticking coefficient de-creases dramatically as the temperature increases. Our Mg flux isabout 1 Å/s. At this flux rate, the sticking coefficient of Mg is�0.74 at 150 �C and it is only �0.17 at 200 �C [21]. This explainswhy our samples made at 200 �C were insulating.

We changed the substrate temperature (Ts) to between 140 and180 �C. Samples grown at an elevated temperature showed a dras-tic difference in Tc from the ones deposited at room temperature.The Tc of the sample is directly related to the Ts as shown in

Fig. 2. As the substrate temperature increased, the Tc increasedand the transition became sharper. An increase of 10 �C in sub-strate temperature improved the Tc by 3–5 K. Samples grown at140 �C, where the excess Mg was seen on the film surface, werehighly metallic and possessed Tc values around 7–9 K. The samplesgrown at 150 �C showed Tc values of 10–14 K. At Ts of 170 and180 �C, the Tc was around 19–20 K. The samples made above180 �C were insulating, indicating a low Mg concentration. Wehad a very narrow window for the substrate temperature change,since it was limited in both directions; when the temperaturewas too low it did not form the MgB2 phase, and when it wastoo high, there was not enough Mg.

To maximize the Mg collection during the growth we also triedto vary the Ts during the deposition. We started the deposition at140 �C to collect enough Mg and gradually increased the tempera-ture during the deposition to a different final temperature value,for example, 170, 180, and 190 �C. If the film was deposited directlyat 190 �C, it would have been insulating since the film could notcollect any Mg at that temperature. However, when we startedthe deposition at 140 �C and ended at 190 �C, the film showedthe superconducting transition. This finding confirms that the mostdecisive factor for the superconducting transition in sputteredMgB2 thin films is the substrate temperature where the phase isfirst formed. Overall, it increased the Tc slightly, but the transitionwas a lot wider than the one of the film deposited at fixed temper-ature. It is possible that these thin films were not homogeneousand had several regions with different Tc values within one sampledue to the Ts change during the growth.

After we added the post in-situ annealing process, the final Tc

improved for all the samples. The sample grown at 150 �C showeda Tc of 24 K, while the sample grown at 180 �C showed a transitiontemperature of 27 K when annealed under the same conditions(620 �C, 30 min). This points out that the Tc of annealed sampleswas primarily determined by the substrate temperature and im-proved by post-annealing.

To find out the Tc dependence on the annealing temperature(Ta), a set of samples deposited at the same temperature were an-nealed at different temperatures. Two sets of data are shown inFig. 3. Each set of the samples were deposited on one big substrateat 170 �C (�) or 180 �C (�) and cut into small pieces to vary theannealing conditions. The first set of data (�) for the thin filmsdeposited at 170 �C showed that Tc increased with an increase inannealing temperature. The Tc value for this set was saturated atthe Ta between 630 and 650 �C. The highest Tc achieved after thesingle temperature annealing was 28 K. As we increased theannealing temperature to above 650 �C, the Tc of the sample de-creased, as we have already stated for the room temperaturedeposited samples. To reach the annealing temperature higher

500 550 600 650 700 750 800

25

26

27

28

29T c

(K)

Ta (ºC)

Fig. 3. Tc dependence on the annealing temperature, Ta. Two sets of data are shownin the figure. Each set of the samples were deposited on one big substrate at 170 �C(�) or 180 �C (�) and cut into small pieces to vary the annealing conditions. Thesample annealed at 800 �C went through a two-step annealing process, 800 �C after620 �C.

42 A. O’Brien et al. / Physica C 469 (2009) 39–43

than 650 �C, we had to use a two-step annealing process. We setthe first temperature lower than 650 �C and increased it to the de-sired annealing temperature. The highest Tc value of 29 K was ob-tained for the sample deposited at 180 �C and annealed at 800 �Cafter the first annealing at 620 �C. The Tc value was increased by2 K from the first annealing.

a

b

c

y:1.00 μm

y:1.00 μm

y:1.00 μm

x:1.00 μm

x:1.00 μm

x:1.00 μm x

Fig. 4. The 3D AFM images with a scan area of a 1.0 lm � 1.0 lm of the MgB2 thin filannealing. Thin films were deposited at (a) room temperature, (b) 140 �C, and (c) 160 �annealed (d) at 600 �C after deposited at room temperature, (e) at 600 �C after deposite

In contrast to the room temperature deposited MgB2 films, theMg/B capping layer did not play a significant role in the annealingprocess for the films deposited at elevated low temperatures.

3.3. Surface morphology of the thin films

The surface morphology of the thin films was measured usingthe AFM operating in the non-contact ‘‘tapping” mode. The RMSroughness was measured with the AFM software analysis programWSxM [22], while the 3D images were generated using the Gwyd-dion software package [23]. We fabricated a set of controlled sam-ples with which we could investigate the effect of the growth andannealing conditions on the surface morphology of the thin films.

The 3D AFM images shown in Fig. 4 were captured with a1.0 lm � 1.0 lm scan area of the MgB2 thin films. The images(a)–(c) are for the as-grown MgB2 thin films without post in-situannealing. Thin films were deposited at (a) room temperature,(b) 140 �C, and (c) 160 �C. The images (d)–(f) are for the in situ an-nealed MgB2 thin films. Thin films were annealed (d) at 600 �C afterbeing deposited at room temperature, (e) at 600 �C after depositedat 160 �C, and (f) at 800 �C/600 �C after being deposited at 160 �C.When we compare the as-grown thin films, (a)–(c), with thepost-annealed thin films, (d)–(f), the post-annealed films appar-ently have well-defined and smaller size grains.

The images (a)–(c) demonstrate the surface morphology changeas a function of Ts. In Fig. 4a, for the thin films grown at room tem-perature, the image appeares to be just the mixture of Mg and B,and the well-defined MgB2 grains are not found. The RMS rough-

e

f

d

y:1.00 μm

y:1.00 μm

x:1.00 μm

x:1.00 μm

:1.00 μm

y:1.00 μm

ms. The images (a)–(c) are for the as-grown MgB2 thin films without post in-situC. The images (d)–(f) are for the in situ annealed MgB2 thin films. Thin films wered at 160 �C, and (f) at 800 �C/600 �C after deposited at 160 �C.

A. O’Brien et al. / Physica C 469 (2009) 39–43 43

ness for Fig. 4a is �0.97 nm. In Fig. 4b for the thin film grown at140 �C, the well-defined grains started to form and the RMS rough-ness increased up to �30 nm. The MgB2 grains seem to be activelydeveloped for the thin films grown at 160 �C as shown in Fig. 4c.Fig. 4c illustrates the two very distinct features. The average RMSroughness value for the whole image is �7 nm but if we excludethe larger grains the roughness is only �1.3 nm. It appears thatthe surface gets smoother as MgB2 grains evolve.

The surface morphology data are highly correlated with our re-sult about the Tc. The thin films deposited at room temperaturewithout post-annealing did not reveal the superconducting transi-tion, and well-defined MgB2 grains were not found in the AFM im-age. The thin films deposited at 140 �C had a quite low Tc,suggesting that the MgB2 phase just started to develop and thiswas seen in AFM image. Tc improved significantly when the sub-strate temperature changed from 140 to 160 �C as shown inFig. 2. This Tc change agrees with the surface morphology changeof the films in AFM images, especially that related to the grain for-mation. Figs. 4e and f show that as the Ta increases, the surface getssmoother. The RMS roughnesses in Figs. 4e and f are �2.09 and�0.97 nm, respectively.

Further research is being performed to analyze the individualgrain sizes of the thin films using a Watershed algorithm. Compo-sitional analysis utilizing Energy Dispersive X-ray analysis is alsocurrently underway to determine the ratio of the Mg and B inthe MgB2 thin films. Once we map out the composition in a smallregion we will check out the composition of the two different re-gions in Fig. 4c, namely the larger grains and the smoother back-ground, to determine how the compositions are different in thosetwo regions.

4. Conclusion

We fabricated MgB2 thin films on C-plane sapphire substratesusing magnetron sputtering. Thin films were deposited by sputter-ing pure B and Mg targets at different substrate temperatures andfollowed by in situ annealing. We have investigated the effects ofthe various growth and annealing parameters, such as depositionand annealing temperatures, on the superconducting properties.

Through the systematic study of the sputtered MgB2 thin films,we established that the substrate temperature is the most deter-mining factor of Tc. It is important to have it in the right tempera-ture region to fully form a MgB2 phase. Any post-treatment gives asecondary effect for which Tc can only increase by a limitedamount. There was no superconducting transition for the thin filmsgrown at room temperature. We had to rely on the post-annealingto form the MgB2 phase. The maximum Tc of the room temperaturegrown samples after the post-annealing was 22 K. When the thinfilms were deposited at an elevated temperature, the Tc signifi-cantly increased. However, the maximum Ts was limited by thelow magnesium sticking coefficient at high temperatures. Themagnesium sticking coefficient became almost zero above 180 �Cwith a deposition rate of �1 Å/s. The highest Tc, 29 K, was obtainedfrom the sample deposited at 180 �C and annealed at 800 �C after

the first annealing at 620 �C. The 3D AFM images clearly demon-strated the correlation between the surface morphology and thesuperconducting property.

To maintain a high sticking coefficient of Mg at high substratetemperature, a method such as thermal evaporation, which has amuch higher deposition rate (>10 Å/s), should be combined withthe regular sputtering method. Although the Tc of the sputteredand in situ annealed MgB2 films is lower than that for the bulkor ex situ annealed thin films, this work presents a significant steptowards the fabrication of MgB2 heterostructures using rather sim-ple physical vapor deposition method such as sputtering.

Acknowledgements

This research was supported by the Research Corporation underContract No. CC6756 and NSF MRI Grant under Contract No.0619909.

References

[1] K. Gorny, O.M. Vyasilev, J.A. Marindale, V.A. Nandor, C.H. Pennington, P.C.Hammel, W.L. Hults, J.L. Smith, P.L. Kuhns, A.P. Reyes, W.G. Moulton, Phys. Rev.Lett. 82 (1999) 177.

[2] W.N. Kang, H.-J. Kim, E.-M. Choi, C.U. Jung, S.I. Lee, Science 292 (2001) 1521.[3] S.R. Shinde, S.B. Ogale, R.L. Greene, T. Venkatesan, P.C. Canfield, S.L. Bud’ko,

G. Lapertot, C. Petrovic, Appl. Phys. Lett. 79 (2001) 227.[4] Dave H.A. Blank, Hans Hilgenkamp, Alexander Brinkman, Dragana Mijatovic,

Guus Rijnders, Horst Rogalla, Appl. Phys. Lett. 79 (2001) 394.[5] X.H. Zeng, A. Sukiasyan, X.X. Xi, Y.F. Hu, E. Wertz, Qi. Li, W. Tian, H.P. Sun, X.Q.

Pan, J. Lettieri, D.G. Schlom, C.O. Brubaker, Zi-Kui Liu, Qiang Li, Appl. Phys. Lett.79 (2001) 1840.

[6] H.Y. Zhai, H.M. Christen, L. Zhang, C. Cantoni, M. Paranthaman, B.C. Sales, D.K.Christen, D.H. Lowndes, Appl. Phys. Lett. 79 (2001) 2603.

[7] S.H. Moon, J.H. Yun, H.N. Lee, J.I. Kye, H.G. Kim, W. Chung, B. Oh, Appl. Phys.Lett. 79 (2001) 2429.

[8] Kenji Ueda, Michio Naito, Appl. Phys. Lett. 79 (2001) 2046.[9] W. Jo, J.-U. Huh, T. Ohnishi, A.F. Marshall, M.R. Beasley, R.H. Hammond, Appl.

Phys. Lett. 80 (2002) 3563.[10] S.D. Bu, D.M. Kim, J.H. Choi, J. Giencke, E.E. Hellstrom, D.C. Larbalestier,

S. Patnaik, L. Cooley, C.B. Eom, J. Lettieri, D.G. Schlom, W. Tian, X.Q. Pan, Appl.Phys. Lett. 81 (2002) 1851.

[11] J.R. Ahn, S.G. Lee, Y. Hwang, G.Y. Sung, D.K. Kim, Physica C 388–389 (2002) 127.[12] H. Ohkubo, M. Akinaga, Physica C 408–410 (2004) 898.[13] M. Akinaga, S. Umeda, H. Hasegawa, T. Shirasawa, Physica C 388–389 (2003)

119.[14] Soon-Gul Lee, Jong-Rok Ahn, Yunwon Kim, Seung-Hyun Moon, Kyu Won Lee,

In-Seon Kim, Yong Ki Park, Supercond. Sci. Technol. 16 (2003) 1550.[15] X.H. Fu, D.S. Wang, Z. P Zhang, J. Yang, Physica C 377 (2002) 407.[16] Z.K. Liu, D.G. Scholm, Q. Li, X.X. Xi, Appl. Phys. Lett. 78 (2001) 3678.[17] A.V. Pogrebnyhalov, J.M. Redwing, S. Raghavan, V. Vaithyanathan, D.G. Schlom,

S.Y. Xu, Q. Li, D.A. Tenne, A. Soukiassian, X.X. Xi, M.D. Johannes, D. Kasinathan,W.E. Paickett, J.S. Wu, J.C.H. Spence, Phys. Rev. Lett. 93 (2004) 147006.

[18] Xianghui Zeng, Alexej V. Pogrebnyakov, Armen Kotcharov, James E. Jones, X.X.Xi, Eric M. Lysczek, Joan M. Redwing, Shengyong Xu, Qi Li, James Lettieri,Darrell G. Schlom, Wei Tian, Xiaoqing Pan, Zi-Kui Liu, Nat. Mater. 1 (2002) 35–38.

[19] T. Uchiyama, H. Koga, T. Iguchi, Jpn. J. Appl. Phys. 43 (2004) 121.[20] H. Shimakage, A. Saito, A. Kawakami, Z. Wang, IEEE Trans. Appl. Supercon. 13

(2003) 3309.[21] Jihoon Kim, R.J. Singh, J.M. Rowell, N. Newman, Kin Gu, David J. Smith, J. Crystal

Growth 270 (2004) 107–112.[22] I. Horcas, R. Fernandez, J.M. Gomez-Rodriguez, J. Colchero, J. Gomez-Herrero,

A.M. Baro, Rev. Sci. Instrum. 78 (2007) 013705.[23] AFM free data analysis program. <http://gwyddion.net>.