A perovskite oxide, BaSnO3, has been classified as one of transparent conducting materials with high electron mobility, and its application for field-effect transistors has been the focus of recent research. Here we report transistor operation in BaSnO3-based heterostructures with atomically smooth surfaces, fabricated on SrTiO3 substrates by the (Sr,Ba)SnO3 buffer technique. Indeed, modulation of band profiles at the channel interfaces with the insertion of wide bandgap (Sr,Ba)SnO3 as a barrier layer results in a significant improvement of field-effect mobility, implying effective carrier doping at the regulated heterointerface. These results provide an important step towards realization of high-performance BaSnO3-based field-effect transistors.

Confining electrons at interfaces offers rich opportunities for exploring quantum transport phenomena as well as for producing novel device concepts based on low dimensionality and high mobility. Recent technological advances in oxide thin-film growth and atomic-scale characterization have brought various oxides into such studies that have been carried out mainly with conventional semiconductors.1 Among a wide variety of oxides, perovskite-type oxides are one of ideal platforms for such research because various fascinating physical properties including ferroelectricity and ferromagnetism can be incorporated in the common crystal framework. In this context, the recent observation of high electron mobility exceeding 300 cm2V−1s−1 at room temperature in La-doped BaSnO3 single crystals2 has provided a new arena for building oxide-based heterostructures. Distinct from transition-metal-based perovskites where strong electron correlation and spin-orbit coupling originating from d orbital character play important roles, BaSnO3 possesses a highly dispersive conduction band predominantly composed of Sn 5s orbitals, which leads to relatively small electron effective mass of 0.20m0 (where m0 is the free electron mass) and high electron mobility.3 

One of obvious research targets of the exceptionally high mobility is field-effect transistor (FET) that employs a non-doped BaSnO3 film as the channel. FET allows for electric-field control of high-mobility charge carrier transport at two-dimensional interface as well as practically useful switching function. Because of the lack of appropriate substrates for lattice-matched epitaxy, however, the growth of BaSnO3 films with bulk-like properties remains challenging.2,4–11 In BaSnO3/SrTiO3(001) with a lattice mismatch of 5.5%, an inevitably formed high density of misfit dislocations was observed by transmission electron microscopy, which is dominant scattering origin for the lower electron mobility in La-doped BaSnO3 films directly grown on SrTiO3(001) than that in single crystals.5 In those samples, it is also hard to obtain a flat surface, which is another critical issue for the realization of high-performance FET with high-quality gate dielectric/channel interfaces.5,11 Since the La-doped BaSnO3 channel prepared on BaSnO3/SrTiO3(001) was confirmed to show FET operation,12–14 the use of an insulating BaSnO3 buffer effectively enhances FET characteristics owing to reduction of dislocation density in the channel and improvement of surface morphology. In addition, we have recently succeeded in obtaining an atomically smooth BaSnO3 surface by introducing Sr0.5Ba0.5SnO3 buffer and high-temperature annealing that mitigate the large lattice mismatch.11 In this paper, we describe successful FET operation using the non-doped, insulating BaSnO3 channel and enhancement of field-effect mobility μFE by engineering the (Sr,Ba)SnO3-based heterointerface.

BaSnO3/Sr0.5Ba0.5SnO3(buffer) multilayers were grown on single-crystalline SrTiO3(001) substrates by pulsed-laser deposition. The film growth was performed using a KrF excimer laser at 900 °C with an oxygen pressure of 0.1 Torr, followed by an ex-situ post-annealing at 1200 °C in air.11 Atomic force microscopy (AFM) observation of the film surfaces, displayed in Fig. 1(a), revealed one-unit-cell height (∼4.1 Å) steps over a wide area. We fabricated a top-gate FET structure on BaSnO3/Sr0.5Ba0.5SnO3(buffer)/SrTiO3(001), schematically shown in Fig. 1(b). The film was patterned into a strip-shaped channel with a width W of 60 μm and a length L of 240 μm by photolithography and Ar ion milling. A gate dielectric layer, an organic polymer parylene-C (relative dielectric constant: εr = 3.15),15,16 with a thickness of ∼480 nm was deposited onto the channel at room temperature. Source, drain, and gate electrodes consisting of Au/Ti bilayers were prepared by electron-beam evaporation. Transport measurements were carried out at room temperature in air with a source-measure unit (Keithley 2614B).

FIG. 1.

(a) Typical AFM image and cross-sectional height profile (measured along the dotted line in the image) of atomically smooth BaSnO3 films grown on Sr0.5Ba0.5SnO3 (buffer)/SrTiO3(001). (b) Schematic device structure of BaSnO3 FET with parylene gate dielectric (Sample A). G, S, D, BSO, SBSO, and STO sub. denote gate, source, drain, BaSnO3, Sr0.5Ba0.5SnO3, and SrTiO3 substrate, respectively. The values in (b) indicate film thicknesses of BSO and SBSO. (c) Output characteristics at room temperature for Sample A.

FIG. 1.

(a) Typical AFM image and cross-sectional height profile (measured along the dotted line in the image) of atomically smooth BaSnO3 films grown on Sr0.5Ba0.5SnO3 (buffer)/SrTiO3(001). (b) Schematic device structure of BaSnO3 FET with parylene gate dielectric (Sample A). G, S, D, BSO, SBSO, and STO sub. denote gate, source, drain, BaSnO3, Sr0.5Ba0.5SnO3, and SrTiO3 substrate, respectively. The values in (b) indicate film thicknesses of BSO and SBSO. (c) Output characteristics at room temperature for Sample A.

Close modal

Output characteristics are displayed in Fig. 1(c) as drain current IDS versus drain voltage VDS measured at various gate voltage VGS for the parylene/BaSnO3 device (hereafter referred to as Sample A). The gate leakage current did not exceed 200 pA while VGS was applied. The application of a positive VGS induces an amplification of IDS from the initially insulating channel, demonstrating typical n-type FET behavior. Linear IDSVDS relation at low VDS indicates the formation of Ohmic contact between Ti and BaSnO3 under application of gate electric field. Current on/off ratio in the measured VGS range is ∼104, which is as large as the values for parylene/SrTiO316 and parylene/KTaO317 devices fabricated on the bulk single crystal surfaces. Clear saturation behavior at high VDS, as well as relatively small IDSVDS hysteresis, corroborates that parylene/BaSnO3 works well as a conducting channel. From the linear region characteristics (as presented below in Figs. 3(c) and 3(d)), μFE was estimated to be 8.2 cm2V−1s−1 using the general relation:

μ FE = I DS V GS L c i W V DS
(1)

where Ci = 5.8 nFcm−2 is the calculated capacitance per unit area of the parylene gate dielectric with assuming εr = 3.15.15,16 In contrast to previous FETs using La-doped BaSnO3 conducting channels,12–14 the present FET on non-doped, insulating ones would have an advantage for applications in that a normally-off device with higher current on/off ratio can be developed.

Previous studies aiming at improving characteristics of La-doped BaSnO3 FETs focused on selection of appropriate gate dielectrics and suitable deposition methods that could reduce trap density at the gate dielectric/channel interface.12–14 In particular, epitaxial LaInO3 barrier/La-doped BaSnO3 channel revealed high μFE of about 90 cm2V−1s−1.13 Instead of taking such approaches, we attempted to tailor interface band profiles by utilizing bandgap tunability of the (Sr,Ba)SnO3. The A-site substitution with isovalent Sr2+ in BaSnO3 has been reported to increase the bandgap from 3.50 eV (BaSnO3) to 4.27 eV (SrSnO3) due to the octahedral tilting distortion.18 In Sr0.5Ba0.5SnO3, bandgap becomes wider by about 0.4 eV than in BaSnO3, while the change in the lattice constant is less than 1%. Although the accurate band alignment of the (Sr,Ba)SnO3/BaSnO3 system remains to be determined, it is reasonable to assume that large part of the increased bandgap conforms to the conduction band offset since the energy level of the valence band composed mainly of O 2p orbitals would be unperturbed so much by cation doping. On the basis of these considerations, we inserted a thin barrier layer of Sr0.5Ba0.5SnO3 to the parylene/BaSnO3 interface. Figure 2 illustrates the cross-sectional schematic of the heterostructure and the plausible band diagram under applying positive VGS. The Sr0.5Ba0.5SnO3 barrier likely acts as a dielectric in response to the applied electric field owing to the wide bandgap and conduction band offset. As a consequence, the dominant charge carrier conduction under positive VGS is preferred in the underlying epitaxial Sr0.5Ba0.5SnO3/BaSnO3 interface. In contrast to the parylene/BaSnO3 interface, the interface scatterings, e.g., by potential fluctuations, are expected to be suppressed by the separation of the conduction channel from the parylene gate dielectric. The concept proposed here essentially resembles the modulation-doping technique used in III-V semiconductor heterostructures.19 

FIG. 2.

Schematics of the cross-sectional device structure (left) and conduction band diagram (right) for (Sr,Ba)SnO3/BaSnO3 heterostructure-based FET proposed in this work. In contrast to the prototype FET, Sample A, in which carriers are induced at the parylene/BaSnO3 interface, conceptually the carriers would be accumulated at the (Sr,Ba)SnO3/BaSnO3.

FIG. 2.

Schematics of the cross-sectional device structure (left) and conduction band diagram (right) for (Sr,Ba)SnO3/BaSnO3 heterostructure-based FET proposed in this work. In contrast to the prototype FET, Sample A, in which carriers are induced at the parylene/BaSnO3 interface, conceptually the carriers would be accumulated at the (Sr,Ba)SnO3/BaSnO3.

Close modal

Figure 3(a) depicts the device structure with a thin Sr0.5Ba0.5SnO3 barrier (Sample B). Using the deposition conditions described above, 7-nm-thick homoepitaxial BaSnO3 and 7-nm-thick Sr0.5Ba0.5SnO3 were regrown on the atomically smooth BaSnO3/(Sr,Ba)SnO3. Laue fringes up to several orders observed in X-ray diffraction measurement ensure the successful regrowth (see Fig. S1 in the supplementary material). To detect electrically the embedded interface of top Sr0.5Ba0.5SnO3/BaSnO3, source and drain electrodes were contacted to BaSnO3 by penetrating the Sr0.5Ba0.5SnO3 barrier with Ar ion milling. In Figs. 3(c) and 3(d), one can see that IDS is in fact increased by the insertion of Sr0.5Ba0.5SnO3 (Samples B and C), which yields μFE = 17 cm2V−1s−1 more than twice as high as that for Sample A. As an additional exemplification to exclude the possibility that the carrier accumulation takes place at the parylene/Sr0.5Ba0.5SnO3 interface, we performed a similar measurement on devices in which source and drain electrodes were placed directly onto the Sr0.5Ba0.5SnO3 barrier; however, the electrostatic modulation in conductance was much smaller (Fig. S2 in the supplementary material). In other words, the much larger IDS in Samples B and C than that in Sample A does not originates from the current conduction at the parylene/Sr0.5Ba0.5SnO3 but from the enhancement effect at the Sr0.5Ba0.5SnO3/BaSnO3. These results strongly support our hypothesis that the conduction channel is preferentially formed in BaSnO3 owing to the band offset effect (Fig. 2).

FIG. 3.

Schematic device structures of Sr0.5Ba0.5SnO3(barrier)/BaSnO3/Sr0.5Ba0.5SnO3(buffer)/SrTiO3(001) FETs: (a) Sample B with regrown Sr0.5Ba0.5SnO3 barrier (7 nm in thickness) and BaSnO3 (7 nm) layers and (b) Sample C with an increased BaSnO3 thickness (164 nm). To maintain a flat surface for Sample C, an in-situ annealing was conducted at 800 °C with an oxygen pressure of 1 × 10−6 Torr for 90 min after depositing half of the regrown layer (82 nm). (c) Transfer characteristics at room temperature measured at VDS = 1 V for Samples A (blue), B (green), and C (red). (d) μFE calculated from the transfer curves using the relation in the text.

FIG. 3.

Schematic device structures of Sr0.5Ba0.5SnO3(barrier)/BaSnO3/Sr0.5Ba0.5SnO3(buffer)/SrTiO3(001) FETs: (a) Sample B with regrown Sr0.5Ba0.5SnO3 barrier (7 nm in thickness) and BaSnO3 (7 nm) layers and (b) Sample C with an increased BaSnO3 thickness (164 nm). To maintain a flat surface for Sample C, an in-situ annealing was conducted at 800 °C with an oxygen pressure of 1 × 10−6 Torr for 90 min after depositing half of the regrown layer (82 nm). (c) Transfer characteristics at room temperature measured at VDS = 1 V for Samples A (blue), B (green), and C (red). (d) μFE calculated from the transfer curves using the relation in the text.

Close modal

Another factor that limits charge transport in BaSnO3 films on SrTiO3 is scattering caused by dislocations.2,5 In our recent study on La-doped BaSnO3 conducting films grown on BaSnO3/Sr0.5Ba0.5SnO3(buffer)/SrTiO3(001),11 we found that the density of misfit dislocations reaching the top La-doped BaSnO3 layer could be markedly reduced by increasing the thickness of BaSnO3. According to this observation, we prepared Sample C with an increased BaSnO3 thickness shown in Fig. 3(b), resulting in much enhanced mobility of μFE = 52 cm2V−1s−1 (Fig. 3(d)). The systematic increase in μFE of three device structures is understood to be due to the suppression of extrinsic scatterings at the gate dielectric/channel interface and misfit dislocations.

Figure 4 summarizes the dependence of μFE on carrier density n3D for our FETs, together with Hall mobility data reported for La-doped BaSnO3 bulk crystals4 and films grown on various substrates by pulsed-laser deposition or molecular-beam epitaxy.4,7,9,11 To perform a fair comparison, the n3D in FETs are estimated from the sheet carrier density given by Ci and VGS and effective channel thickness taking into account the wave function distribution.20 This analysis is usually applied for comparing 2D transport properties with bulk counterparts.21,22 The values of μFE reported for La-doped BaSnO3 FETs are also plotted against n3D calculated by the same procedure using relative permittivity of gate dielectric and sample geometry in literatures.12,14 On the basis of above assumptions, the estimated n3D corresponds to the upper limit of conducting charge carriers because the number of conducting charge carrier is usually less than the calculated value due to interface trapping and/or low dielectric constant in reality. This plot represents the significance of this work in the following points. At a doping level of n3D = 6 × 1018 cm−3, μFE of our prototype device, Sample A, is comparable to those of La-doped BaSnO3 FETs; but its much steeper rise in μFE against n3D suggests that μFE can potentially be higher than them upon further carrier doping. Although the highest value of μFE of Sample C with the (Sr,Ba)SnO3-heterostructured channel does not reach the 90 cm2V−1s−1 of LaInO3/La-doped BaSnO3,13 it is distinctly high as compared with those values in other La-doped BaSnO3 FETs,12,14 and, more importantly, seems to be exceeding some Hall mobilities for lightly La-doped BaSnO3 films. To clarify this point, we need Hall effect measurements to evaluate directly the conducting charge carrier density and mobility in future experiment. Considering the absence of ionized impurity scattering in FETs, we envisage that the μFE in our FETs may eventually reach and even surpass the high mobility observed in the bulk when further increase of charge carrier density is realized by employing high-κ dielectric materials. At least, the dependence of μFE on calculated n3D in our FETs manifests large potential of high-mobility channel in BaSnO3 in the low-doping region inaccessible by chemical doping.

FIG. 4.

μFE at room temperature as a function of n3D for BaSnO3 FETs (Samples A and C). Hall mobilities at room temperature for La-doped BaSnO3 single crystals (black closed circles), La-doped BaSnO3/SrTiO3(001) (Ref. 4) (black open circles), La-doped BaSnO3/BaSnO3(001) (Ref. 7) (light blue triangles), and La-doped BaSnO3/BaSnO3/Sr0.5Ba0.5SnO3 (buffer)/SrTiO3(001) (Ref. 11) (red open squares) prepared by pulsed-laser deposition, and La-doped BaSnO3/PrScO3(110) (Ref. 9) (green diamonds) by molecular-beam epitaxy are plotted for comparison. FETs with La-doped BaSnO3 thin-film channels on BaSnO3 (buffer)/SrTiO3(001) and Al2O3 (Ref. 12) (purple solid line) and HfO2 gate dielectrics14 (orange solid line) are also included.

FIG. 4.

μFE at room temperature as a function of n3D for BaSnO3 FETs (Samples A and C). Hall mobilities at room temperature for La-doped BaSnO3 single crystals (black closed circles), La-doped BaSnO3/SrTiO3(001) (Ref. 4) (black open circles), La-doped BaSnO3/BaSnO3(001) (Ref. 7) (light blue triangles), and La-doped BaSnO3/BaSnO3/Sr0.5Ba0.5SnO3 (buffer)/SrTiO3(001) (Ref. 11) (red open squares) prepared by pulsed-laser deposition, and La-doped BaSnO3/PrScO3(110) (Ref. 9) (green diamonds) by molecular-beam epitaxy are plotted for comparison. FETs with La-doped BaSnO3 thin-film channels on BaSnO3 (buffer)/SrTiO3(001) and Al2O3 (Ref. 12) (purple solid line) and HfO2 gate dielectrics14 (orange solid line) are also included.

Close modal

In summary, we fabricated FETs with atomically smooth non-doped BaSnO3 channels. We proposed heterointerface engineering with the wide bandgap (Sr,Ba)SnO3 for improving FET characteristics. By suppressing extrinsic scatterings due to ionized impurity, interface roughness, and dislocation, we obtained substantial μFE enhancement up to 52 cm2V−1s−1. The (Sr,Ba)SnO3/BaSnO3 heterointerface as demonstrated here is a promising system for achieving high-mobility electron transport via electrostatic doping. Enhancement in doping capability, e.g., by use of high-κ gate dielectric materials such as atomic-layer-deposited HfO2, would lead to further investigation on electronic transport of the high-mobility 5s-based electronic state in common perovskite oxide.

See Supplementary Material for results of XRD (Fig. S1) and control FET measurements (Fig. S2).

We would like to thank T. Seki and K. Takanashi for the use of lithography facilities. This work was supported by a Grant-in-Aid for Scientific Research (A) (No. 15H02022) from Japan Society for the Promotion of Science.

1.
H. Y.
Hwang
,
Y.
Iwasa
,
M.
Kawasaki
,
B.
Keimer
,
N.
Nagaosa
, and
Y.
Tokura
,
Nat. Mater.
11
,
103
(
2012
).
2.
H. J.
Kim
,
U.
Kim
,
H. M.
Kim
,
T. H.
Kim
,
H. S.
Mun
,
B.-G.
Jeon
,
K. T.
Hong
,
W.-H.
Lee
,
C.
Ju
,
K. H.
Kim
, and
K.
Char
,
Appl. Phys. Express
5
,
061102
(
2012
).
3.
H.-R.
Liu
,
J.-H.
Yang
,
H. J.
Xiang
,
X. G.
Gong
, and
S.-H.
Wei
,
Appl. Phys. Lett.
102
,
112109
(
2013
).
4.
H. J.
Kim
,
U.
Kim
,
T. H.
Kim
,
J.
Kim
,
H. M.
Kim
,
B.-G.
Jeon
,
W.-J.
Lee
,
H. S.
Mun
,
K. T.
Hong
,
J.
Yu
,
K.
Char
, and
K. H.
Kim
,
Phys. Rev. B
86
,
165205
(
2012
).
5.
H.
Mun
,
U.
Kim
,
H. M.
Kim
,
C.
Park
,
T. H.
Kim
,
H. J.
Kim
,
K. H.
Kim
, and
K.
Char
,
Appl. Phys. Lett.
102
,
252105
(
2013
).
6.
P. V.
Wadekar
,
J.
Alaria
,
M.
O’Sullivan
,
N. L. O.
Flack
,
T. D.
Manning
,
L. J.
Phillips
,
K.
Durose
,
O.
Lozano
,
S.
Lucas
,
J. B.
Claridge
, and
M. J.
Rosseinsky
,
Appl. Phys. Lett.
105
,
052104
(
2014
).
7.
W.-J.
Lee
,
H. J.
Kim
,
E.
Sohn
,
T. H.
Kim
,
J.-Y.
Park
,
W.
Park
,
H.
Jeong
,
T.
Lee
,
J. H.
Kim
,
K.-Y.
Choi
, and
K. H.
Kim
,
Appl. Phys. Lett.
108
,
082105
(
2016
).
8.
K.
Ganguly
,
P.
Ambwani
,
P.
Xu
,
J. S.
Jeong
,
K. A.
Mkhoyan
,
C.
Leighton
, and
B.
Jalan
,
APL Mater.
3
,
062509
(
2015
).
9.
S.
Raghavan
,
T.
Schumann
,
H.
Kim
,
J. Y.
Zhang
,
T. A.
Cain
, and
S.
Stemmer
,
APL Mater.
4
,
016106
(
2016
).
10.
Z.
Lebens-Higgins
,
D. O.
Scanlon
,
H.
Paik
,
S.
Sallis
,
Y.
Nie
,
M.
Uchida
,
N. F.
Quackenbush
,
M. J.
Wahila
,
G. E.
Sterbinsky
,
D. A.
Arena
,
J. C.
Woicik
,
D. G.
Schlom
, and
L. F. J.
Piper
,
Phys. Rev. Lett.
116
,
027602
(
2016
).
11.
J.
Shiogai
,
K.
Nishihara
,
K.
Sato
, and
A.
Tsukazaki
,
AIP Adv.
6
,
065305
(
2016
).
12.
C.
Park
,
U.
Kim
,
C. J.
Ju
,
J. S.
Park
,
Y. M.
Kim
, and
K.
Char
,
Appl. Phys. Lett.
105
,
203503
(
2014
).
13.
U.
Kim
,
C.
Park
,
T.
Ha
,
Y. M.
Kim
,
N.
Kim
,
C.
Ju
,
J.
Park
,
J.
Yu
,
J. H.
Kim
, and
K.
Char
,
APL Mater.
3
,
036101
(
2015
).
14.
Y. M.
Kim
,
C.
Park
,
U.
Kim
,
C.
Ju
, and
K.
Char
,
Appl. Phys. Express
9
,
011201
(
2016
).
15.
A. F.
Stassen
,
R. W. I.
de Boer
,
N. N.
Iosad
, and
A. F.
Morpurgo
,
Appl. Phys. Lett.
85
,
3899
(
2004
).
16.
H.
Nakamura
,
H.
Takagi
,
I. H.
Inoue
,
Y.
Takahashi
,
T.
Hasegawa
, and
Y.
Tokura
,
Appl. Phys. Lett.
89
,
133504
(
2006
);
H.
Nakamura
, Ph.D. thesis,
University of Tokyo
, Tokyo,
2007
.
17.
H.
Nakamura
and
T.
Kimura
,
J. Appl. Phys.
107
,
074508
(
2010
).
18.
Q.
Liu
,
B.
Li
,
J.
Liu
,
H.
Li
,
Z.
Liu
,
K.
Dai
,
G.
Zhu
,
P.
Zhang
,
F.
Chen
, and
J.
Dai
,
EPL
98
,
47010
(
2012
).
19.
T.
Mimura
,
S.
Hiyamizu
,
T.
Fujii
, and
K.
Nanbu
,
Jpn. J. Appl. Phys.
19
,
L225
(
1980
).
20.
T.
Ando
,
A. B.
Fowler
, and
F.
Stern
,
Rev. Mod. Phys.
54
,
437
(
1982
).
21.
K.
Ueno
,
S.
Nakamura
,
H.
Shimotani
,
A.
Ohtomo
,
N.
Kimura
,
T.
Nojima
,
H.
Aoki
,
Y.
Iwasa
, and
M.
Kawasaki
,
Nat. Mater.
7
,
855
(
2008
).
22.
C.
Bell
,
S.
Harashima
,
Y.
Kozuka
,
M.
Kim
,
B. G.
Kim
,
Y.
Hikita
, and
H. Y.
Hwang
,
Phys. Rev. Lett.
103
,
226802
(
2009
).

Supplementary Material