Carrier doping into two dimensional (2D) Mott insulators is one of the prospective strategies for exploring exotic quantum phenomena. Although ultra-thin oxide films are one such target, it is vitally important to fabricate well-defined and clean samples to extract intrinsic properties. In this study, we start from establishing the growth of clean SrVO3 films with a low residual resistivity (∼4 × 10−7 Ω cm) and a high mobility (∼103 cm2/V s). By confining them with SrTiO3 barrier layers, the Mott insulator state appears at the thickness below 3 unit cells (u.c.). By the electron doping in the form of LaxSr1−xVO3 for such two dimensional systems (2 and 3 u.c), metallic-like phases appear in a narrow x region around x = 0.17, indicating a collapse of the Mott insulator state. This study demonstrates that artificial 2D systems of clean oxides are a promising playground for exploring novel Mott physics in confined systems.

Dimensionality is one of the key parameters to control quantum critical phenomena in condensed matter physics. For example, exotic electronic phases have been discovered in two dimensional (2D) electronic systems such as layered perovskite oxides, graphene, and transition metal dichalcogenides.1–3 In particular, an unconventional superconductivity often emerges in carrier doped 2D Mott insulators as exemplified by layered cuprates/nickelates as well as magic-angle twisted bilayer graphene.4–6 Among transition-metal oxides, research targets include not only compounds with natural 2D crystal structures but also artificial 2D structures that are grown in a thin film form controlling the layer thickness on an atomic scale, owing to the recent progress in the epitaxy technique.7,8

Here, we study SrVO3 (SVO), which is one of the suitable candidate materials to study the artificial 2D Mott physics. SVO with an electron occupation of 3d1 is a Pauli paramagnetic metal in bulk samples due to the relatively large electron transfer interaction mediated with strong hybridization between V 3d and O 2p orbitals. By substituting La for Sr as LaxSr1−xVO3 [L(x)SVO], the occupation of 3d orbitals can be increased to induce phase transition toward insulator around x = 0.8 in bulk crystals.9,10 The counter end compound LaVO3 is one of the prototypical Mott insulators with the electron occupation of 3d2 that undergoes a transition to antiferromagnetic and orbital-ordered insulating phase below T = 140 K.10 Another controlling parameter for inducing the Mott insulator state in SVO is the 2D confinement in ultra-thin films, where the effective bandwidth is reduced to drive the system into an insulating state. Indeed, photoemission measurements revealed the existence of Mott–Hubbard gaps in SVO films with thicknesses t = 2 and 3 unit cells (u.c.), while showing the evidence of a dimensional crossover around 3–5 u.c.11,12 In an example of a quasi-2D system with a layered-perovskite (K2NiF4) structure, the LaxSr2−xVO4 system has previously been investigated.13,14 The end compound, Sr2VO4, is known as a typical Mott insulator with a 3d1 configuration. Interestingly, it has been found to be metallic for x = 0.15–0.20 in LaxSr2−xVO4.14 

Our motivation is that carrier doping into the ultra-thin Mott insulator state of SVO may lead to interesting electronic phases such as exotic metallic or even superconducting states. However, the transport properties of such systems have not been reported so far. Recently, high crystalline quality complex oxide films have been grown by metalorganic gas source molecular beam epitaxy (MOMBE).15–17 For example, the quantum Hall effect was successfully observed in δ-doped SrTiO3 (STO) grown at a high temperature by MOMBE.8 Also reported is the much lower residual resistivity (0.13 μΩ cm) in high crystalline quality SVO films grown by MOMBE17 compared to the typical values (0.46–8.7 μΩ cm) of bulk single crystals.18,19

In this study, we have fabricated 2D confined L(x)SVO(t) thin films with various thicknesses t and La doping concentrations x by MOMBE. The transport properties have been analyzed to elucidate the electronic phase diagram. A metal–insulator transition (MIT) is confirmed by the reduction of SVO layer thickness below 3 u.c. in accord with previous reports.11,12 Starting from the insulating 2D SVO with thicknesses of 2 and 3 u.c., the effect of La doping is analyzed to reveal the emergence of metallic phase in a narrow x region around x = 0.17. Such a phase diagram is quite contrastive with that of 3D L(x)SVO, where the La substitution is found to reduce the mobility monotonically due to the enhanced randomness. This is the first demonstration to dope electron carriers into artificial 2D Mott insulators of SVO to cause the filling-controlled Mott transition.

The heterostructure studied here is schematically shown in Fig. 1(b). Thin films of the STO buffer layer (5 u.c.), L(x)SVO layer (t), and STO capping layer (5 u.c.) were successively grown on (001) (LaAlO3)0.3(Sr2AlTaO6)0.7 (LSAT) single-crystal substrates. In our MOMBE, Sr and La were provided from conventional effusion cells with respective pure elemental sources, while Ti and V were supplied as MO precursors of titanium tetraisopropoxide (TTIP, 99.9999%) and vanadium triisopropoxide (VTIP, 99.9999%), respectively. These MO containers were kept around 100 °C for evaporation, and the precursor vapor was fed without any carrier gas. The beam equivalent pressure (BEP) for Sr was set at 8 × 10−8 Torr throughout the growth. The MO BEP and substrate temperature for the buffer STO layer were set at 1.7 × 10−6 Torr and 900 °C, those for the L(x)SVO layer were 2.4 × 10−6 Torr and 800 °C, and those for capping STO were 1.7 × 10−6 Torr and 800 °C, respectively. The La flux was controlled by the effusion cell temperature calibrated by a quartz crystal microbalance flux monitor. We supply no oxidation gas since TTIP and VTIP precursor molecules contain four oxygen atoms per Ti atom and V atom, respectively, which are sufficient to stabilize perovskite oxides. After deposition, the heterostructures were ex situ annealed at 250 °C in air to fully oxidize the samples.

Figure 1(a) shows a 2θθ scan around the (002) peak of the LSAT substrate for the SVO heterostructures with various thicknesses. For the 3D SVO (190 u.c.) heterostructure, a sharp (002) peak of SVO layer and clear Laue’s fringes can be seen. With a reduction in the SVO thickness, the intensity of SVO peaks gradually reduces, but the peak position remains the same. In addition, a broad peak at around 46° can be assigned to the (002) diffraction from the buffer and capping STO layers. The reciprocal space mapping of x-ray diffraction (XRD) indicates that the in-plane lattice constant is locked to that of the LSAT substrate for all the films (not shown). Compared with a = 3.842 Å in a cubic bulk SVO crystal, the SVO layer is under tensile strain with in-plane and out-of-plane lattices of a = 3.868 Å and c = 3.825 Å. An atomic force microscopy image (AFM) of a L(0.13)SVO (2 u.c.) heterostructure is shown in Fig. 1(c), which exhibits step-and-terrace structures with a step height of 4 Å.

Figure 1(d) shows a cross-sectional image of the L(0.13)SVO (2 u.c.) heterostructure obtained using a high-angle annular dark-field scanning transmission electron microscope (HAADF-STEM). The incident beam is set along the [100] direction, which corresponds to the vertical direction in the AFM image in Fig. 1(c). The arrows denote the interface between the heterostructure and the substrate. Figure 1(e) shows the energy dispersive x-ray spectroscopy (EDX) intensity map for the K-edge of a V atom. Three atomic layers composed of a brightest center atomic layer sandwiched between two atomic layers with a lower intensity can be seen. Although the interface between L(x)SVO and STO layers is expected to be atomically sharp, one can recognize small islands and meandered steps in the AFM image in Fig. 1(c). Because of such imperfections, the cross-sectional atomic image of a STEM sample taken with a thickness of 100 nm is thought to result in a rather indistinct interface as shown Fig. 1(e).

Figure 2 shows the temperature dependence of resistivity of SVO heterostructures with various thicknesses. A MIT is observed between 5 and 3 u.c. as has been observed by photoemission spectroscopy experiments.11,12 Interestingly, the critical thickness of the MIT appeared to be much higher (around 17 u.c.) for SVO thin films grown by pulsed laser deposition (PLD) in previous reports.20 In view of the much larger residual resistivity (74 μΩ cm) in thick SVO films grown by PLD, the MIT phenomenon may come not only from intrinsic 2D confinement but also from disorder potentials due to possible defects.

Now, we discuss the transport properties of electron doped 2D Mott insulators, L(x)SVO(t) with t = 2 and 3 u.c., in comparison with the doping effect for the 3D metal case with t = 190 u.c. Figure 3(a) shows the temperature dependence of resistivity of various L(x)SVO (190 u.c.) heterostructures. As explained above, due to the high crystalline quality of the present MOMBE films, the thick non-doped SVO shows an excellent metallicity with a very low residual resistivity (4.1 × 10−7 Ω cm) at 2 K. The resistivity monotonically increases with La substitution presumably due to the enhancement of randomness. In contrast, distinct behaviors were found in the Mott insulating samples with t = 3 and 2 u.c., as shown in Figs. 3(b) and 3(c), respectively. In the case of t = 2 u.c. samples, the resistivity drops by more than two orders of magnitude from x = 0 to 0.17, followed by an increase with a larger x. Although the reduction is smaller for t = 3 u.c. samples, a similar trend is also observed. In Fig. 4, we summarize the x dependence of resistivity [(a)–(c)], carrier density [(d)–(f)], and mobility [(g)–(i)] for the three heterostructures. As mentioned above, the resistivity of 190 u.c. samples increases monotonically with x. The mobility at a low temperature for x = 0 samples is over 1000 cm2/V s, and it quickly decreases below 10 cm2/V s for x > 0.13. This is straightforwardly understood that La doping into clean SVO has a considerable impact on the scattering of carrier transport. Interestingly, the electron carrier density increases more rapidly from the nominal value (dotted line), the reason for which is unclear yet. In the cases of 3 and 2 u.c. samples, the behavior of the MIT becomes more obvious in these plots. Both carrier density and mobility have maximum peaks around x = 0.17, resulting in the region of minimum resistivity there.

The transport data can be used to visualize phase diagrams in temperature–x planes and compare those with various confinement configurations. Since the MIT in correlated electron oxides generally locates at critical resistivity values around 1 × 10−3 Ω cm, the color scale is fixed in the range between 10−3.5 (=3.2 × 10−4) and 10−2 Ω cm to highlight the MIT boundary. The contour mapping of the resistivity of the 3 u.c. (left) and 2 u.c. (middle) films is depicted with the color scale in Fig. 5. The mapping data for both the 3 and 2 u.c. samples clearly indicate that metallic-like conductive states emerge at around x = 0.17. Note that such a phase diagram is completely different from that of bulk crystals for LaxSr1−xVO3, showing the transition from paramagnetic metal to antiferromagnetic insulator at around x = 0.8.10 The insulator state at a smaller x side is considered as the two dimensional Mott insulating state of d1 orbital bands, which was assigned by previous photoemission experiments.11 Thus, the important issue to address is the magnetic nature of this 2D Mott insulator state. Due to the very small volume in such ultra-thin films and rather small spin moment of 3d1 system, however, it is challenging to experimentally elucidate the magnetic ground state, especially the antiferromagnetic one, by neutron diffraction and superconducting quantum interference device (SQUID) magnetometer measurements. So far, we could not find any sign of magnetic orders from the SQUID magnetometer measurements. In the case of conductive samples of L(x)SVO (2 and 3 u.c.) heterostructures, we measured the magnetoresistance and the Hall effect at low temperatures. As shown in Fig. S2, the typical magnetotransport data for L(0.20)SVO (2 u.c.) indicate the indiscernible sign of anomalous Hall effect and a negative magnetoresistance plausibly due to weak localization that may cause the resistivity upturns observed in Figs. 3(b) and 3(c).

Here, we refer to the results in the LaxSr2−xVO4 system, which is viewed as a quasi-2D system with a layered-perovskite (K2NiF4) structure,14 to discuss the possible relation with Mott insulators in our system. The end compound, Sr2VO4, is known as a typical Mott insulator with 3d1 configuration. Sr2VO4 was reported as an antiferromagnet below a Néel temperature of 45 K.21 Later, however, there have been several reports claiming various other magnetic states, causing controversy over the magnetic ground state of Sr2VO4.22,23 As for the orbital ordering, a low temperature x-ray diffraction clearly revealed that there exists dyz/dzx orbital ordering below 97 K.24 In our tensile strained thin films (c/a = 0.989), the orbital may be polarized to dxy. However, it is left for a future study to elucidate the spin/orbital nature associated with the confinement induced Mott insulator state of SrVO3.

The origin of the insulator state above x = 0.2 is not evident yet. One possible scenario is that the increasing electron correlation with x toward the d2 Mott insulating state leads to the suppressed kinetic energy and hence to the charge localization in conjunction with the confinement effect and the enhanced electron–phonon effect. Another possibility is that the charge and orbital ordering characteristic of the partial band filling regime might be stabilized in this confined structure.25 In a previous report on layered-perovskite LaxSr2−xVO4 films, similar MITs were revealed to give metallic state only around x = 0.15–0.20 for the natural 2D system14 as shown in the right panel of Fig. 5. Interestingly, the region of metallic state is also quite narrow and the insulating state extends from x = 0.3 to 1. Such phase diagrams in 3d1+x systems (0 ≤ x ≤ 1) in (quasi-)2D structures seem to be quite similar.

In summary, we have investigated the MIT of 2D confined LaxSr1−xVO3 films grown on LSAT substrates by MOMBE. The SVO (x = 0) samples exhibit a clear transition from clean metal to insulator with a reduction in the thickness from 190 u.c. to around 3 u.c., indicating the emergence of 2D Mott insulator states. By the La doping in these 2D systems, electron carriers are successfully doped, inducing a metallic-like state with a sharp suppression in the resistivity around x = 0.17. This study clearly indicates that the quantum-critical Mott transition can be studied in artificial 2D correlated electron systems.

See the supplementary material for the possibility of conduction in buffer and/or capping SrTiO3 layers and the magnetotransport properties of La doped SrVO3 heterostructures.

We are grateful to T. Arima for fruitful discussions. This work was partly supported by JSPS KAKENHI (Grant No. 22H04958).

The authors have no conflicts to disclose.

K. S. Takahashi: Conceptualization (lead); Data curation (lead); Formal analysis (lead); Writing – original draft (lead). Y. Tokura: Project administration (equal); Writing – original draft (supporting). M. Kawasaki: Project administration (lead); Supervision (equal); Writing – original draft (supporting).

The data that support the findings of this study are available within the article and its supplementary material.

1.
A. P.
Mackenzie
and
Y.
Maeno
,
Rev. Mod. Phys.
75
,
657
(
2003
).
2.
K. S.
Novoselov
,
A. K.
Geim
,
S. V.
Morozov
,
D.
Jiang
,
M. I.
Katsnelson
,
I. V.
Grigorieva
,
S. V.
Dubonos
, and
A. A.
Firsov
,
Nature
438
,
197
(
2005
).
3.
V.
Fatemi
,
S.
Wu
,
Y.
Cao
,
L.
Bretheau
,
Q. D.
Gibson
,
K.
Watanabe
,
T.
Taniguchi
,
R. J.
Cava
, and
P.
Jarillo-Herrero
,
Science
362
,
926
(
2018
).
4.
J. G.
Bednorz
and
K. A.
Müller
,
Z. Phys. B: Condens. Matter
64
,
189
(
1986
).
5.
D.
Li
,
K.
Lee
,
B. Y.
Wang
,
M.
Osada
,
S.
Crossley
,
H. R.
Lee
,
Y.
Cui
,
Y.
Hikita
, and
H. Y.
Hwang
,
Nature
572
,
624
(
2019
).
6.
Y.
Cao
,
V.
Fatemi
,
S.
Fang
,
K.
Watanabe
,
T.
Taniguchi
,
E.
Kaxiras
, and
P.
Jarillo-Herrero
,
Nature
556
,
43
(
2018
).
7.
Y.
Kozuka
,
M.
Kim
,
C.
Bell
,
B. G.
Kim
,
Y.
Hikita
, and
H. Y.
Hwang
,
Nature
462
,
487
(
2009
).
8.
Y.
Matsubara
,
K. S.
Takahashi
,
M. S.
Bahramy
,
Y.
Kozuka
,
D.
Maryenko
,
J.
Falson
,
A.
Tsukazaki
,
Y.
Tokura
, and
M.
Kawasaki
,
Nat. Commun.
7
,
11631
(
2016
).
9.
F.
Inaba
,
T.
Arima
,
T.
Ishikawa
,
T.
Katsufuji
, and
Y.
Tokura
,
Phys. Rev. B
52
,
R2221
(
1995
).
10.
S.
Miyasaka
,
T.
Okuda
, and
Y.
Tokura
,
Phys. Rev. Lett.
85
,
5388
(
2000
).
11.
K.
Yoshimatsu
,
T.
Okabe
,
H.
Kumigashira
,
S.
Okamoto
,
S.
Aizaki
,
A.
Fujimori
, and
M.
Oshima
,
Phys. Rev. Lett.
104
,
147601
(
2010
).
12.
M.
Kobayashi
,
K.
Yoshimatsu
,
T.
Mitsuhashi
,
M.
Kitamura
,
E.
Sakai
,
R.
Yukawa
,
M.
Minohara
,
A.
Fujimori
,
K.
Horiba
, and
H.
Kumigashira
,
Sci. Rep.
7
,
16621
(
2017
).
13.
J.
Matsuno
,
Y.
Okimoto
,
M.
Kawasaki
, and
Y.
Tokura
,
Phys. Rev. Lett.
95
,
176404
(
2005
).
14.
J.
Matsuno
,
Y.
Okimoto
,
M.
Kawasaki
, and
Y.
Tokura
,
Appl. Phys. Lett.
82
,
194
(
2003
).
15.
B.
Jalan
,
R.
Engel-Herbert
,
N. J.
Wright
, and
S.
Stemmer
,
J. Vac. Sci. Technol. A
27
,
461
(
2009
).
16.
M.
Brahlek
,
A. S.
Gupta
,
J.
Lapano
,
J.
Roth
,
H. T.
Zhang
,
L.
Zhang
,
R.
Haislmaier
, and
R.
Engel‐Herbert
,
Adv. Funct. Mater.
28
,
1702772
(
2017
).
17.
J. A.
Moyer
,
C.
Eaton
, and
R.
Engel-Herbert
,
Adv. Mater.
25
,
3578
(
2013
).
18.
B. L.
Chamberland
and
P. S.
Danielson
,
J. Solid State Chem.
3
,
243
(
1971
).
19.
I. H.
Inoue
,
O.
Goto
,
H.
Makino
,
N. E.
Hussey
, and
M.
Ishikawa
,
Phys. Rev. B
58
,
4372
(
1998
).
20.
M.
Gu
,
S. A.
Wolf
, and
J.
Lu
,
Adv. Mater. Interfaces
1
,
1300126
(
2014
).
21.
A.
Nozaki
,
H.
Yoshikawa
,
T.
Wada
,
H.
Yamauchi
, and
S.
Tanaka
,
Phys. Rev. B
43
,
181
(
1991
).
22.
J.
Sugiyama
,
H.
Nozaki
,
I.
Umegaki
,
W.
Higemoto
,
E. J.
Ansaldo
,
J. H.
Brewer
,
H.
Sakurai
,
T.-H.
Kao
,
H.-D.
Yang
, and
M.
Månsson
,
Phys. Rev. B
89
,
020402(R)
(
2014
).
23.
R.
Viennois
,
E.
Giannini
,
J.
Teyssier
,
J.
Elia
,
J.
Deisenhofer
, and
D.
Van der Marel
,
J. Phys.: Conf. Ser.
200
,
012219
(
2010
).
24.
H. D.
Zhou
,
B. S.
Conner
,
L.
Balicas
, and
C. R.
Wiebe
,
Phys. Rev. Lett.
99
,
136403
(
2007
).
25.
M.
Pickem
,
J.
Kaufmann
,
K.
Held
, and
J. M.
Tomczak
,
Phys. Rev. B
104
,
024307
(
2021
).

Supplementary Material