This content has been downloaded from IOPscience. Please scroll down to see the full text.

Download details:

IP Address: 133.60.10.167

This content was downloaded on 03/12/2015 at 03:00

Please note that terms and conditions apply.

BaTiO3 based relaxor ferroelectric epitaxial thin-films for high-temperature operational

capacitors

View the table of contents for this issue, or go to the journal homepage for more

2015 Jpn. J. Appl. Phys. 54 04DH02

(http://iopscience.iop.org/1347-4065/54/4S/04DH02)

Home Search Collections Journals About Contact us My IOPscience

BaTiO3 based relaxor ferroelectric epitaxial thin-films

for high-temperature operational capacitors

Somu Kumaragurubaran1, Takahiro Nagata1, Kenichiro Takahashi2, Sung-Gi Ri2,Yoshifumi Tsunekawa2, Setsu Suzuki2, and Toyohiro Chikyow1

1International Center for Materials Nanoarchitectonics (WPI-MANA), National Institute for Materials Science (NIMS),Tsukuba, Ibaraki 305-0044, Japan2COMET Inc., c/o NIMS, Tsukuba, Ibaraki 305-0044, JapanE-mail: Somu.Kumaragurubaran@nims.go.jp

Received September 9, 2014; accepted November 13, 2014; published online January 22, 2015

The epitaxial growth of 0.6[BaTiO3]–0.4[Bi(Mg2/3Nb1/3)O3] (BT–BMN) relaxor ferroelectric thin-films on (100) Nb doped SrTiO3 substrates has beenachieved and the structure is investigated for high-temperature capacitor applications. The post growth annealing decreases the oxygen vacancyand other defects in BT–BMN films, resulting in the enhancement of dielectric constant. An insertion of intermediate SrRuO3 layers as an electrodeinstead of Pt, sandwiching the film, is found to be more effective in enhancing the dielectric constant. A very high dielectric constant exceeding 400was achieved from high-temperature annealed film and the film showed an excellent dielectric constant stability of below 11% in the temperaturerange of 80–400 °C. This will enable smaller, high-temperature tolerant, monolithically integrated thin-film capacitors on power semiconductordevices. © 2015 The Japan Society of Applied Physics

1. Introduction

Power semiconductor devices based on SiC, GaN, anddiamond are needed for compact and efficient energy-savingdomestic, automotive, aviation, space, and geothermalexploration applications and power transmission. Many ofthese applications involve high-temperature and harshenvironment. In view of this, SiC based transistors andSchottky diodes capable of performing over 250 °C havebeen developed.1,2) At present, SiC based devices haveadvanced to the commercial manufacturing stage. Like theadvancement in these active devices, it is also necessary todevelop passive elements such as a capacitor, resistor andinductors that are capable of operating at temperatures over250 °C.

Currently available capacitors can function up to 175 °Conly. Moreover, they are bulky and occupy a significantspace in the module. Therefore, it is very important todevelop a thin-film capacitor that can withstand at high-temperatures and can be integrated monolithically to theactive devices. This will facilitate a reduction in parasiticinductive losses and faster switching of the device. Also it ispossible to take advantage of the high thermal conductivity ofthe substrate and develop cooling-free compact device. Thecore issue is to develop a dielectric medium for the capacitorand the requirements of a dielectric material is to have (1) ahigh dielectric constant of above 400 over a wide frequencyrange, (2) high stability of dielectric constant against tem-perature typically, below 15%, (3) low dielectric loss andleakage current, and (4) free from hazardous elements.Several lead-free relaxor ferroelectric materials have beenexplored3–8) as a dielectric medium since the temperaturedependency of dielectric constant in relaxors exhibit a verybroad peak correlated to the local polar nanometer sizeregions (PNR). As a member of relaxor family, recently,x[BaTiO3]–(1 ¹ x)[Bi(Mg2/3Nb1/3)O3] (BT–BMN) bulk ce-ramics showed a high dielectric constant and a very highstability of dielectric constant against temperature.9) How-ever, a thin-film process technology has to be developed formonolithic device integration. From the device design andoperation view, a capacitor made of epitaxial thin-film has

several advantages like control in the direction of polarizationorientation, suppression of grain boundary related leakagecurrent and clear definition of interface states. However, untilnow, the reports on the epitaxial thin-film process technologyof lead-free relaxors are very scarce in the literature.3,5,10)

In the practical capacitor design and process integrationespecially, for high-temperature applications, factors such asthe electrode structure, interfacial reaction at repeated heatcycles, adhesion of layer to the substrate, thermal stress, bandoff-set, etc. are to be considered. Among several lead-freerelaxors, BaTiO3 based compounds are preferable becauseBaTiO3 has been adopted in the semiconductor industry fora long time. Its distorted cubic structure is compatible toseveral electrodes and also to major semiconductor sub-strates. Therefore, one can think of an all-epitaxial capacitorwith a better process controllability and thereby a controlledinterface states, defects and electrical properties. In thispaper, we demonstrate the epitaxial growth of BT–BMN thin-films on (100) SrTiO3 (STO) substrate employing pulse laserdeposition (PLD) method and show its potential as adielectric material for high-temperature tolerant capacitors.The influence Pt and SrRuO3 (SRO) electrodes on theelectrical properties is also addressed.

2. Experiment

High-temperature sintered Bi-excess 0.6[BaTiO3]–0.4[Bi-(Mg2/3Nb1/3)O3] and SRO ceramic targets were preparedand pre-installed in the deposition chamber. The (100)oriented Nb 0.5wt% doped STO was used as a substrate. AKrF excimer laser with a wavelength of 248 nm, repetitionrate of 10Hz, and energy density of 1.5 J/cm2 was employedto ablate the target. The deposition was performed at specifictemperatures of 490, 510, and 530 °C and at specific oxygenpartial pressures of 0.2, 2, and 20 Pa. Prior to BT–BMNgrowth, a 50 nm SRO was deposited on some substrates at570 °C and an oxygen partial pressure of 0.13 Pa. Prior to Ptelectrode deposition, some BT–BMN films were annealed at825 and 850 °C for 5min with an oxygen flow of 0.2 L/min.X-ray diffraction (XRD) analysis was carried out usingBruker AXS D8 discover GADDS system equipped with alarge angle two-dimensional (2D) detector. Using a shadow

Japanese Journal of Applied Physics 54, 04DH02 (2015)

http://dx.doi.org/10.7567/JJAP.54.04DH02

REGULAR PAPER

04DH02-1 © 2015 The Japan Society of Applied Physics

metal mask, a 120-nm-thick Pt was deposited as topelectrode. On some films, a 50 nm SRO was deposited atroom temperature prior to Pt deposition. To aid groundcontact, a 120 nm Pt was deposited on the backside of STO.Current–voltage (I–V) and capacitance–voltage (C–V) meas-urements were conducted using source-measurement unit(Keithley 2612B) and LCR meter (Agilent E4980A)respectively in the temperature range of 25–400 °C. X-rayphotoelectron spectroscopy measurements were carried onthe samples at room temperature and UHV conditions, usingVG Thermo ScientificTheta Probe system, with Al K¡radiation (h¯ = 1486.6 eV). The X-ray beam diameter was400 µm and the pass energy was set to 30 eV. An Ar ionneutralizer gun was used to counter the sample chargingeffect. The overall energy resolution of the system is 0.7 eV.The X-ray photoelectron spectroscopy (XPS) data wascalibrated against the Au 4f7/2 core level (84.0 eV). Corelevel spectra were fitted using the Voigt function after thebackground correction employing the Shirley function.11)

3. Results and discussion

The 2D XRD pattern of BT–BMN films grown directly onSTO at 20 Pa and 510 °C showed singular spots [Fig. 1(a)]corresponding to 100 and 200 reflections at 22.3 and 44.5°respectively. The singular spot intensity decreased in filmsgrown at 2 and 0.2 Pa and exhibited poly crystallinity (notshown). BT–BMN films grown at 20 Pa but at temperatures490 and 530 °C were also polycrystalline. This suggests thatthe pressure and temperature window to realize highlyordered epitaxial film growth is rather narrow in (100)direction. Figure 1(b) is the pole figure projection of fourof [111] reflections on (100) plane, exhibiting a fourfoldsymmetry, further confirming the epitaxial relationship ofBT–BMN film with the substrate. Figure 2 shows thedielectric constant and dielectric loss, as a function of sweepfrequency, for the films deposited at temperatures of 490,510, and 530 °C and at 20 Pa. In the as-grown state, theepitaxial BT–BMN film grown at 510 °C exhibited higherdielectric constant than the poly crystalline BT–BMN filmsgrown at 490 and 530 °C, which exceeded 250 at thefrequency of 100 kHz.

In the next step, to investigate the post growth annealingeffects on the dielectric constant of BT–BMN films, BT–BMN films grown at 510 °C were annealed. Annealingprocess substantially improved the dielectric constant of the

film and the highest value is obtained on the film annealed at850 °C (Fig. 2). The dielectric constant and loss are over 350and below 0.2 even at 1MHz, respectively. To understand thereason behind the enhancement of dielectric constant of theannealed film, XRD and XPS analysis were performed.Though the XRD pattern of the annealed film did not showany remarkable difference between the as-grown and theannealed BT–BMN films, XPS measurements revealed thereduction of defects. Figure 3(a) shows Ba 3d5/2 core levelspectra of the as-grown and annealed BT–BMN films. Theas-grown BT–BMN indicated a peak around 781 eVcorresponding to defects of the B-site in perovskitestructure.12) In the annealed film, this peak drasticallydecreased, and the intensity ratio of O 1s core spectra (notshown) corresponding to BT–BMN increased from 0.3 to0.5, suggesting that the concentration of defects and oxygenvacancy were improved. The oxygen vacancy related pointdefects play a major role on the ferroelectric properties of

Fig. 1. (Color online) (a) 2D-XRD pattern of BT–BMN film depositedon (100) STO substrate at 510 °C with an oxygen partial pressure of 20 Pa.(b) Pole figure projection of four of [111] reflections of BT–BMN film on(100) STO substrate.

Fig. 2. (Color online) Dielectric constant (solid lines) and dielectric loss(broken lines) as a function of sweep frequency for the films deposited attemperatures 490, 510, and 530 °C on (100) oriented substrates at 20 Pa. Postgrowth annealing improved the dielectric constant and the highest value isobtained for 850 °C annealed film.

Fig. 3. (Color online) XPS of (a) Ba 3d5/2 and (b) Bi 4f7/2 core levels ofas-grown and 825 °C annealed BT–BMN films. Solid lines show measuredspectra (Exp.; red line) and background (BG; black line). Open circles showsum-fitted curves. Dashed lines are fitted curves for each bond: BT–BMN,defective region (filled area), and metallic Bi (filled area).

Jpn. J. Appl. Phys. 54, 04DH02 (2015) S. Kumaragurubaran et al.

04DH02-2 © 2015 The Japan Society of Applied Physics

perovskite oxides.13–15) It should be mentioned here that theNb containing perovskites tolerate a high density of oxygenvacancy defects.16) The as-grown films are deficient ofoxygen and the annealing process reduced the oxygenvacancies and thereby increased the film dielectric constant.However, the dielectric constant of the thin-film is muchlower than its bulk counterpart. The electrode-film interfaceand the effective charge distribution in this region andsimilarly the substrate-film interface affect the capacitance.Therefore, the estimated dielectric constant could be differentfrom the real film dielectric constant. It is worthy to note thatBT–BMN films were deposited directly on STO substratesfrom Bi-enriched targets and post annealed at high temper-ature. It is known that Bi incorporated STO films haveshowed several dielectric peaks which are very sensitiveto oxygen ions or oxygen vacancies.17) Figure 3(b) showsBi 4f7/2 core level spectra for the as-grown and the annealedBT–BMN films. In the as-grown state, a Bi metallic peakappeared on the lower energy end of the spectrum.18,19) Thispeak is pronounced in the annealed film surface. Further-more, the intensity ratio of Bi 4f7/2 to Ba 3d5/2 increasedfrom 0.5 to 0.8 after the annealing. These results imply thathigh-temperature annealing caused Bi out-diffusion since825 °C is close to the melting point of Bi2O3.20) In addition,many a Bi containing oxide compounds melts around900 °C.21,22)

Besides segregation of Bi on BT–BMN surface, it isalso likely that Bi affects the STO interface. We altered theSTO substrate-film interface by introducing SRO epitaxiallayer, a widely used perovskite conductive electrode in Bicontaining thin-films on STO.5,23,24) In addition, SROinterface has been found to be advantageous for BaTiO3

based thin-film capacitor structures.25–27) The 2D-XRDpattern of BT–BMN grown on SRO/STO structure is shownin Fig. 4(a). Singular spots corresponding to 100 and 200reflections are obtained without any foreign phase. The filmslattice spacings are larger than the substrate. Figure 4(b)depicts the pole figure projection of four of [111] reflectionson (100) plane evidencing a coherent epitaxial relationshipamong BT–BMN and SRO films and STO substrate. Thefilms were annealed at 850 °C under an oxygen flow of0.2 L/min. To circumvent a possible reaction of Pt and Biat high temperatures,28) a 50 nm SRO was deposited using ametal mask, before Pt electrode deposition on BT–BMMfilm. Frequency dispersion of dielectric constant (solid lines)and dielectric loss (broken lines) of 850 °C annealed epitaxial

films with the SRO/Pt top electrode is illustrated in Fig. 5.The dielectric constant remarkably increased in the SROsandwiched BT–BMN film. The films showed a highdielectric constant exceeding 550 at 100 kHz at room tem-perature measurement. This value is close to the bulkdielectric constant.

The high-temperature C–V measurements were performedon Pt/SRO/BT–BMN/SRO/STO/Pt capacitor structurefrom room temperature to 400 °C. Figure 6 represents thedielectric constant (solid symbols) and dielectric loss (opensymbols) of the film as a function of temperature in theforward (increasing temperature) and reverse (decreasingtemperature) directions at a measured frequency of 100 kHz.The BT–BMN film showed a very high dielectric constantexceeding 500 and a low dielectric loss up to 350 °C in theforward direction and it is decreased in the reverse direction.However, it remains over 400 at room temperature. Thedielectric constant stability of the film against temperature isdefined as

�C

C¼ "max � "min

"RT� 100%

where ¾max, ¾min, and ¾RT are respectively the maximum,minimum and room temperature dielectric constants. Thestability is higher in the reverse direction. Note that the

Fig. 4. (Color online) (a) 2D-XRD pattern of BT–BMN film depositedon SRO/STO and (b) pole figure projection of four of [111] reflections ofBT–BMN film on (100) STO substrate.

Fig. 5. (Color online) Frequency dispersion of dielectric constant (solidlines) and dielectric loss (broken lines) of 850 °C annealed epitaxial BT–BMN film with SRO electrodes. The inset illustrates the sample structure.

Fig. 6. (Color online) Temperature dependence of dielectric constant(solid symbols) and dielectric loss (open symbols) of 850 °C annealedepitaxial BT–BMN film with SRO electrodes.

Jpn. J. Appl. Phys. 54, 04DH02 (2015) S. Kumaragurubaran et al.

04DH02-3 © 2015 The Japan Society of Applied Physics

dielectric constant exceeds 400 and the stability is below 11%in the temperature range of 80–400 °C which are morepromising while considering recent reports.3,29,30) In thesecond heat cycle and more, the dielectric constant valueswere about the same values as the reverse direction valuesof first cycle (not shown). This should be caused by aninterfacial reaction between BT–BMN and SRO top electrodedeposited at room temperature, indicating that the interfacialreaction at high-temperature influences the dielectric constantand also the stability. Therefore, it is possible to achievehigher dielectric constant and better thermal stability ofdielectric constant from this material when the interface isproperly engineered.

4. Conclusions

Epitaxial growth of (100) oriented BT–BMN thin-films wererealized on Nb:SrTiO3 substrates by PLD. The temperatureand pressure window to realize epitaxial growth in (100)direction is rather narrow. A high dielectric constant exceed-ing 400 at 100 kHz and a high thermal stability of dielectricconstant below 11% was accomplished in the temperaturerange of 80–400 °C. The interfacial reactions influence thesuperior dielectric properties of BT–BMN. When the inter-face is appropriately tuned it is possible to gain higherdielectric constant and its stability against temperature.

Acknowledgements

Funding from Japan Science and Technology Agency(JST)—Adaptable and Seamless Technology TransferProgram through Target-driven R&D (A-STEP, No.AS2315002B). WPI-MANA was established by WorldPremier International Research Center Initiative (WPI), theMinistry of Education, Culture, Sports, Science and Tech-nology, Japan (MEXT).

1) R. Hedayati, L. Lanni, S. Rodriguez, B. G. Malm, A. Rusu, and C.-M.Zetterling, IEEE Electron Device Lett. 35, 693 (2014).

2) P. G. Neudeck, D. J. Spry, L.-Y. Chen, C. W. Chang, G. M. Beheim, R. S.Okojie, L. J. Evans, R. D. Meredith, T. L. Ferrier, M. J. Krasowski, and

N. F. Prokop, Mater. Sci. Forum 615–617, 929 (2009).3) T. M. Correia, M. McMillen, M. K. Rokosz, P. M. Weaver, J. M. Gregg, G.

Viola, and M. G. Cain, J. Am. Ceram. Soc. 96, 2699 (2013).4) V. V. Shvartsman and D. C. Lupascu, J. Am. Ceram. Soc. 95, 1 (2012).5) H. Tanaka, M. T. Chentir, T. Yamada, S. Yasui, Y. Ehara, K. Yamato, Y.

Kashiwagi, C. N. Theng, J. Wang, S. Okamura, H. Uchida, T. Iijima, S.Wada, and H. Funakubo, J. Appl. Phys. 111, 084108 (2012).

6) R. Dittmer, W. Jo, D. Damjanovic, and J. Rödel, J. Appl. Phys. 109, 034107(2011).

7) B. Xiong, H. Hao, S. Zhang, H. Liu, and M. Cao, J. Am. Ceram. Soc. 94,3412 (2011).

8) H. Ogihara, C. A. Randall, and S. Trolier-McKinstry, J. Am. Ceram. Soc.92, 1719 (2009).

9) R. K. Malhan, N. Sugiyama, Y. Noguchi, and M. Miyayama, US Patent8,194,392,B2 (2012).

10) D. S. Tinberg and S. Trolier-McKinstry, J. Appl. Phys. 101, 024112 (2007).11) D. A. Shirley, Phys. Rev. B 5, 4709 (1972).12) M. Wegmann, L. Watson, and A. Hendry, J. Am. Ceram. Soc. 87, 371

(2004).13) R.-A. Eichel, Phys. Chem. Chem. Phys. 13, 368 (2011).14) G. Y. Yang, E. C. Dickey, C. A. Randall, D. E. Barber, P. Pinceloup, M. A.

Henderson, R. A. Hill, J. J. Beeson, and D. J. Skamser, J. Appl. Phys. 96,7492 (2004).

15) G. Deng, G. Li, A. Ding, and Q. Yin, Appl. Phys. Lett. 87, 192905 (2005).16) X. Zhao, W. Qu, and X. Tan, J. Am. Ceram. Soc. 91, 3031 (2008).17) Z. Yu, C. Ang, P. M. Vilarinho, P. Q. Mantas, and J. L. Baptista, J. Appl.

Phys. 83, 4874 (1998).18) S. Ono, A. Sakakibara, T. Osaka, I. Koiwa, J. Mita, and K. Asami,

J. Electrochem. Soc. 146, 685 (1999).19) A. C. Palanduz and D. M. Smyth, J. Electroceram. 11, 191 (2003).20) P. Patnaik, Handbook of Inorganic Chemical Compounds (McGraw-Hill,

New York, 2003) p. 115.21) S. Kumaragurubaran, S. M. Babu, K. Kitamura, S. Takegawa, and C.

Subramanian, J. Cryst. Growth 229, 233 (2001).22) O. Sanz, E. Haro-Poniatowski, J. Gonzalo, and J. M. Fernandez Navarro,

J. Non-Cryst. Solids 352, 761 (2006).23) L. Gunawan, G.-Z. Zhu, Y. Shao, S. Lazar, O. Gautreau, C. Harnagea, A.

Pignolet, and G. A. Botton, J. Appl. Phys. 113, 044102 (2013).24) S. Yasui, K. Yazawa, T. Yamada, K. Nishida, H. Uchida, M. Azuma, and

H. Funakubo, Jpn. J. Appl. Phys. 49, 09MB04 (2010).25) Y. A. Boikov and T. Claeson, Appl. Phys. Lett. 80, 4603 (2002).26) Y. S. Kim, J. Y. Jo, D. J. Kim, Y. J. Chang, J. H. Lee, T. W. Noh, T. K. Song,

J.-G. Yoon, J.-S. Chung, S. I. Baik, Y.-W. Kim, and C. U. Jung, Appl. Phys.Lett. 88, 072909 (2006).

27) J. Shin, A. Y. Borisevich, V. Meunier, J. Zhou, E. W. Plummer, S. V.Kalinin, and A. P. Baddorf, ACS Nano 4, 4190 (2010).

28) Y. Shimakawa and Y. Kubo, Appl. Phys. Lett. 75, 2839 (1999).29) C.-C. Huang and D. P. Cann, J. Appl. Phys. 104, 024117 (2008).30) Y. Yuan, C. J. Zhao, X. H. Zhou, B. Tang, and S. R. Zhang, J. Electroceram.

25, 212 (2010).

Jpn. J. Appl. Phys. 54, 04DH02 (2015) S. Kumaragurubaran et al.

04DH02-4 © 2015 The Japan Society of Applied Physics