Metallization of Epitaxial VO2 Films by Ionic Liquid ... - ACS Publications

Subscriber access provided by CORNELL UNIVERSITY LIBRARY

Communication

Metallization of epitaxial VO2 films by ionic liquid gating through intially insulating TiO2 layers. Donata Passarello, Simone G. Altendorf, Jaewoo Jeong, Mahesh G. Samant, and Stuart S. P. Parkin Nano Lett., Just Accepted Manuscript • DOI: 10.1021/acs.nanolett.6b01882 • Publication Date (Web): 01 Aug 2016 Downloaded from http://pubs.acs.org on August 2, 2016

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Nano Letters is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

Metallization of epitaxial VO2 films by ionic liquid gating through initially insulating TiO2 layers Donata Passarello §,†, Simone G. Altendorf †, Jaewoo Jeong§, Mahesh G. Samant§, and Stuart S. P. Parkin*,§,† §

IBM Research - Almaden, San Jose, California 95120, USA.

†

Max Plank Institute for Microstructure Physics, Weinberg 2, 06120 Halle (Saale), Germany

Ionic liquid gating has been shown to metallize initially insulating layers formed from several different oxide materials. Of these vanadium dioxide (VO2) is of especial interest because it itself is metallic at temperatures above its metal-insulator transition. Recent studies have shown that the mechanism of ionic liquid gated induced metallization is entirely distinct from that of the thermally driven metal-insulator transition and is derived from oxygen migration through volume channels along the (001) direction of the rutile structure of VO2. Here we show that it is possible to metallize the entire volume of 10 nm thick layers of VO2 buried under layers of rutile titanium dioxide (TiO2) up to 10 nm thick. Key to this process is the alignment of volume channels in the respective oxide layers which have the same rutile structure with clamped in-plane lattice constants. The metallization of the VO2 layers is accompanied by large structural expansions of up to ~6.5% in the out-of-plane direction but the structure of the TiO2 layer is hardly affected by gating. The TiO2 layers become weakly conducting during the gating process but, in contrast to the VO2 layers, the conductivity disappears on exposure to air. Indeed, even after air exposure, x-ray photoelectron spectroscopy studies show that the VO2 films have a reduced oxygen content after metallization.

Ionic liquid gating of the VO2 films through initially insulating TiO2 layers is

not consistent with conventional models that have assumed the gate induced carriers are of electrostatic origin.

KEYWORDS: oxygen migration; metal insulator transition; ionic liquid gating; epitaxial oxide film; oxide heterostructure; metallization

1

ACS Paragon Plus Environment

Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 20

Ionitronics relies on the controlled motion of ions rather than electrons, as in conventional electronic devices, to manipulate the properties of various materials but, most particularly, oxide thin films. Much work has focused on resistive switching wherein atomic size conducting paths are formed by breakdown of oxide thin films from intense electric fields provided across metallic contacts1, 2. Sometimes conducting filaments are formed by electric field induced diffusion of metal atoms from the contacts into the oxide film. In this approach, since only a small portion of the film between the contacts is metallized, potential applications are limited to two terminal devices3, 4. By contrast using polarized ionic liquids (IL) which provide intense electric fields one can transform the entire volume of initially insulating oxide layers into metals5, 6. This allows for three terminal transistor devices where the ionic liquid forms a gate dielectric which is polarized by a gate voltage7-11. In this context, the recent observation of the volume metallization of VO2

5, 6, 12

and WO3

13

, with

thicknesses of up to 120 nm using gate polarized ionic liquids is of considerable interest. The proposed mechanism of the volume metallization involves the migration of oxygen ions from the body of the oxide film to the ionic liquid driven by the intense electric fields created within the electric double layer14,

15

at the ionic liquid - oxide interface5.

Studies of the crystal facet

dependence of the volume metallization in VO2 and WO3 shows the importance of volume channels that are an innate part of the crystal structures of these oxides, to facilitate the oxygen ion migration6, 13

. Here, we show, in support of this mechanism, that epitaxial single-crystalline films of (001)

oriented VO2 can be volume metallized even when covered with insulating layers of (001) TiO2 with thicknesses of up to at least 10 nm.

TiO2 / VO2 heterostructures were grown onto TiO2 (001) single crystal substrates by pulsed layer deposition (PLD). The TiO2 substrates were sonicated in de-ionized (DI) water for 10 minutes and then etched in 7:1 HF:NH4F buffered oxide etch solution for 50 seconds16. After etching, they were rinsed in a stream of running DI water and finally were dried with nitrogen gas gun. The substrates were immediately transferred to the PLD chamber where they were annealed in 10 mTorr oxygen for 30 minutes at 450 °C. The 10 nm VO2 films were deposited, after the substrate was cooled to 400 °C, in 10 mTorr oxygen pressure; these growth conditions were optimal as the change in resistance at the VO2 metal to insulator transition (MIT) was between three and four orders of magnitude. TiO2 (001) films of various thicknesses, 2.5 nm, 5 nm, 10 nm, 17 nm and 30 nm, were grown on top of the 10 nm thick VO2 (001) films at the same growth temperature but in higher oxygen pressures (15

2


Page 3 of 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

mTorr) to prevent any degradation of the MIT of the VO2 films. The deposition of TiO2 cap layers at oxygen pressures of less than 10 mTorr likely leads to oxygen deficient VO2 films. The laser (wavelength = 248 nm) energy density on the oxide target (sintered from stoichiometric VO2 or TiO2) was attenuated to ~1 J/ cm2 per pulse, and the films were grown at a laser pulse rate of 2 Hz. The target to substrate distance was fixed at ~7 cm. The substrates were rotated during growth and the laser beam was rastered across a portion of the 25 mm diameter target which was also rotating continuously. In-situ reflection high energy electron diffraction (RHEED) before and after the film growth shows the high quality epitaxial growth (Fig. S1a,b) and atomic force microscopy measurements shows the films are very smooth (Fig. S1c).

High resolution x-ray diffraction (XRD) measurements were carried out with a Bruker D8 Discover diffractometer using monochromatic Cu Kα radiation. Detailed studies were carried out on several series of TiO2 / VO2 heterostructures where the thickness of the VO2 layer was fixed at 10 nm and the TiO2 thickness varied from 2.5 nm to 30 nm. These studies show that both the VO2 and TiO2 layers take up the rutile structure where the TiO2 lattice constants are close to bulk rutile TiO2 but the VO2 in plane lattice constant is clamped to the TiO2 substrate, as discussed later. Typical XRD θ-2θ diffraction patterns for a series of four heterostructures with TiO2 film thicknesses, tcap = 2.5 nm, 5 nm, 10 nm and 17 nm, grown on 10 nm VO2, are shown in Figure 1a. These measurements show that both layers exhibit the same crystal orientation with the rutile c-axis out-of-plane. The TiO2 substrate peaks (at 2θ =~62.8°) are sharp and intense while the VO2 film peaks (at 2θ =~66°) are broadened due to finite layer thickness effects. For the heterostructures, the TiO2 film peaks are not easy to identify as they overlap with the substrate peaks, but from the Kiessig fringes the thickness of both the TiO2 and VO2 films was determined (see Fig. S2 and S3). A cross-sectional scanning transmission electron microscopy (STEM) image of a 10 nm TiO2 / 10 nm VO2 on TiO2 (001) is shown in Figure 1b. The image was taken using a JEOL ARM 200F TEM with a Cold-FEG source that was operated at 200 keV. The heterostructure films are epitaxial with perfect alignment of the columns of TiO2 and VO2. At the interface between the substrate and the VO2 film we observe a thin dark layer which we attribute to surface damage of the substrate during the mechanical polishing of the single crystals. The interface between the VO2 and the TiO2 film, however, is abrupt and is of high quality.

3


Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 20

IL gating experiments were performed on 10 x 10 mm2 area samples in a lateral field-effect transistor geometry. The source and drain contacts were two ~1 mm wide 65 nm thick ion beam sputter deposited gold stripes and 5 nm Ru adhesion layers deposited on two opposing sides of the sample using an aluminum foil shadow mask. These stripes were deposited directly onto the VO2 layer by milling away the top TiO2 layer at the edges of the sample. This was accomplished using argon ion milling with secondary ion mass spectrometry end point detection.

For the gating

experiments the sample was placed in a Teflon sample tray together with an independent 10 x 10 mm2 gold plate (0.5 mm thick) that is the gate electrode, as shown in Figure 1c. The sample tray was filled with an IL which covers the sample and the gate electrode. The IL used in these experiments was 1-Hexyl-3-methylimidazolium bis(trifluormethylsulfonyl)imide (HMIM-TFSI). Prior to the gating experiments, the sample and the IL were baked in an external high vacuum chamber. The Teflon tray was attached to a cold head of a cryogen-free cryostat which is placed inside a stainless steel vacuum shroud. During operation the vacuum shroud is continuously pumped to maintain a vacuum of ~10-5 Torr to eliminate contamination of ionic liquid by water. The sample was gated at a temperature of 270 K where the pristine VO2 is in an initially insulating state below its MIT temperature. After gating, the IL was cleaned off with isopropyl alcohol (IPA) and ex-situ transport and x-ray diffraction (XRD) measurements were performed. The transport measurements were then performed ex-situ in a Quantum Design PPMS DynaCool at a temperature sweep rate of 3 K/min. A 4 contact in-line geometry was used for the transport measurements with contacts, spaced ~0.5 mm apart, formed by Al wire ultrasonic bonding in the middle of the sample. After these measurements the sample was placed back in the vacuum system with new ionic liquid for the next gating experiment.

Figure 1d compares ex-situ electrical transport measurements on 10 nm thick VO2 (001) with 2.5 nm, 5 nm and 10 nm TiO2 capping layers. Results are shown for the pristine state, after gating at gate voltage, VG = 2.6 V and after reverse gating at VG = -2.6 V. The 2.5 nm, 5 nm and 10 nm devices were gated for 21 hours, 15 hours and 3 days respectively.

The reverse gate voltage was

applied for ~20 hours under 250 mbar oxygen. In each case there is no evidence for any MIT after gating at VG = 2.6 V. Thus, an important result is that the VO2 layer is fully volume metallized for each of the TiO2 capping layers.

Moreover, the resistance of the high temperature metallic state

4


Page 5 of 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

above the MIT is increased after gating in each case. The sheet resistance increases from 212, 303 and 218 ohm to 427, 599 and 785 ohm respectively. This behavior is very similar to that previously found for uncapped VO2 layers5. The increased resistance is not compatible with ionic liquid electrostatic gate induced carrier injection but is compatible with, as discussed further below, the formation of oxygen vacancies by IL gating.

Moreover, this is self-evident from the fact that the

MIT remains suppressed even after the ionic liquid is completely removed. After reverse gating (VG = -2.6 V), the sheet resistance versus temperature characteristics are almost identical to those of the pristine (ungated) films, showing that ionic liquid gate induced volume metallization is reversible. In situ x-ray diffraction measurements were performed at the Stanford Synchrotron Radiation Laboratory (SSRL) in order to study structural changes of the heterostructure while gating. XRD data were collected with a monochromatic x-ray beam focused to a spot size of ~120 x 375 µm2 with an energy of 12.0004 keV. A Si(111) analyzer crystal with a scintillation counter was used to detect the diffracted x-rays. The IL experiments were carried out in a specially designed x-ray cell17 with devices that were fabricated in exactly the same manner as those used for the transport experiments, discussed above but here the gate electrode is formed from a 1 mm diameter, 21 cm long, 99.99% pure gold wire that is wrapped around the device three times. The sample and the gold wire were enclosed in a chamber that had a O-ring sealed lid formed from a thin 7.5 µm thick Kapton sheet to allow for transmission of the x-ray beam. The chamber was pump and purged with nitrogen several times at room temperature and then filled with ~2 ml of the IL which had been previously vacuum baked at 120 °C for ~24 hours and then transferred whilst still hot into the chamber using an injection syringe and Teflon tubes. After injection the chamber is pumped to remove any remaining nitrogen pockets and excess ionic liquid and to minimize the thickness of ionic liquid through which the x-rays pass. Integrated within the cell is a Peltier cooler that allows for operation at temperatures as low as 250 K but here all the experiments were carried out at 270 K. Figure 2 shows XRD θ-2θ scans for tcap = 2.5 nm, 5 nm, 10 nm and 17 nm. The XRD scans were taken in an out-of-plane geometry (the diffraction plane is perpendicular to the sample surface). The devices were gated at the same gate voltage of VG= +2.6 V but for varying times. Figure 2a shows the XRD pattern for the sample with tcap = 2.5 nm. After gating for ~4 hours the VO2 (002) peak shifts to lower 2θ angles, corresponding to an expansion of ~2.5% of the out-of-plane lattice parameter. After VG is reduced to 0, the VO2 (002) peak starts to shift back towards higher 2θ

5


Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 20

angles, indicating a partial recovery towards its original state. But, only after after reverse gating at VG= -1.5 V for 90 minutes does the structure of the film revert completely back to its pristine state. A smaller structural change is observed for tcap = 5 nm, even though it was gated for a much longer time of ~18 hours, as shown in Figure 2b. In this case, the VO2 film structure expands by only ~1.7%. After a gate voltage VG= -1 V applied for 1 hour and a VG= -1.5 V applied for 1 hour and 38 minutes the structure again reverts back to its pristine state. Measurements for tcap = 10 nm are shown in Figure 2c. In this case the structure expands by an even smaller amount of ~1.3% after ~18 hours gating, and reverted its pristine state once the gate voltage was reduced to 0 V for 3 hours and 40 minutes. Finally, for the thickest capping layers considered (17 nm and 21 nm) no evidence for any structural change was found (see data for tcap = 17 nm in Figure 2d). The temporal response of the source-drain current for the same gate voltage measured in similar devices in the in situ XRD setup and in the ex situ transport studies is the same within experimental error. The in situ XRD measurements were performed in a nitrogen environment whereas for the ex situ experiments the samples were gated in vacuum and the chamber was continuously pumped during the gating process, allowing for a more oxygen free environment. To check whether this had limited the magnitude of the gate induced structural changes we carried out a series of ex-situ XRD experiments using samples that were gated in the same setup and under identical conditions to those used for the transport measurements in Figure 1d. Figure 3 shows the resulting XRD θ-2θ scans. For tcap = 2.5 nm we find a lattice expansion along the out-of-plane component of ~6% (Figure 3a), which is more than twice as high as that found for the same capping layer thickness in in situ measurements, though the gating time is longer here (21 hours). An even larger expansion of ~6.5% is observed for the sample with tcap = 5 nm (Figure 3b, gating time = ~ 15 hours). A smaller but still substantial lattice expansion of ~2.8% is observed for tcap =10 nm (Figure 3c, gating time = ~3 days). Finally, Figure 3d shows results for a 10 nm thick VO2 (001) film without any TiO2 capping layer. This film was also gated for a considerable time of ~18 hours, but the VO2 film undergoes a smaller expansion, upon gating, compared to the films capped with 2.5 and 5 nm TiO2. Reciprocal space mapping (RSM) measurements were performed before and after gating for all the samples. Results are shown in Figure 4 for tcap = 10 nm.

The RSM data clearly show that the

intense TiO2(202) substrate peaks and the weaker VO2(202) peaks (at h = 2 and l = ~2.07) and their associated Kiessig fringes in both samples which move only along the (20l) direction after gating.

6


Page 7 of 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

These data thus show that the VO2 film structure is clamped in-plane to the TiO2 substrate and has the same in-plane lattice parameter, irrespective gating. This result is in agreement with the STEM data shown in Figure 1b.

Figure 5a shows transport measurements for tcap = 2.5 nm, 5 nm and 10 nm: pristine state (black) and after gating for 3, 6 and 12 hours at VG = 2.6 V. The MIT is already fully suppressed after 3 hours for tcap = 2.5 nm and 5 nm but 6 hours are needed to suppress the MIT for tcap = 10 nm. Note that the electrical data shown in Fig. 5a are for gating after up to 12 hours but the structural data shown in Fig 3c are after gating for 3 days. Long after the MIT is suppressed and the VO2 is metallized there continue to be both changes in structure and conductivity.

The conductivity worsens with

longer gating times as shown in Fig. S5. Although, at first sight this might seem surprising this is consistent with the decrease in conductivity in the metallic state above the MIT that gating causes. The decreased conductivity is consistent with the formation of oxygen vacancies – effectively defects that give rise to additional resistance. Fig. S5 shows that with gating the conductivity initially increases and then eventually decreases with gate times of many hours (or days). At the same time the structure continues to evolve slowly even after several days. Thus we can identify 2 regimes: at short gating times the structure evolves first before the MIT is suppressed6 and at longer times the structure and conductivity both evolve more slowly with increasing oxygen vacancy levels.

In a previous study we showed that ionic liquid gating of a TiO2 (001) single crystal leads to metallization of only a very thin region at the surface, ~2-3 nm thick18. Here we have shown that we are able to gate VO2 layer through intervening TiO2 insulating layers with thicknesses of up to 10 nm. To understand the role of the TiO2 capping layers we prepared thin layers of TiO2 on TiO2 (001) single crystal substrates. The substrates were first annealed in 15 mTorr oxygen for 30 minutes at 450 °C and then TiO2 layers 2.5 nm, 5 nm, 10 nm and 20 nm thick were deposited at 400 C in the same oxygen pressure. A control sample that was otherwise treated identically was without any TiO2 layer. Three terminal transistor devices were fabricated with channel sizes of 100 µm x 20 µm, that were defined by patterning, using standard UV lithography, a 60 nm thick SiO2 dielectric layer deposited on top of the TiO2 layers by ion beam deposition. The TiO2 layers were unpatterned. A gate voltage

7


Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

of VG = 2.6 V was applied for ~ 27 hours till the source drain current reached a steady value and Hall measurements were then performed. All devices with TiO2 layers up to 20 nm thick could be metallized. Figure 5b shows the sheet carrier density derived from the Hall measurements using a single carrier model as a function of TiO2 layer thickness. The carriers are electrons and the carrier density increases almost linearly with the TiO2 layer thickness at a given temperature (data for 100 K and 150 K are shown in the Figure). This shows that the entire thickness of the TiO2 layer is gated to the same metallicity and that only a small portion of the substrate is gated. (Note that results for the 20 nm thick TiO2 device were limited to 2 terminal resistance measurements due to contact problems (Fig. S4b)). By extrapolating to zero carrier density we estimate that a thickness of the substrate of ~2 nm at 100 K and ~1.2 nm at 150 K is metallized in good agreement with ref. 18. We find no evidence for any structural changes on gating the TiO2 layers. In-situ x-ray data for an exemplary pristine and a fully gated 10 nm thick TiO2 layer show negligible changes (see Fig. S6). A STEM of the 5 nm thick TiO2 layer is shown in Figure 5c. The TiO2 films are expitaxial and of high crystalline quality but the micrograph shows similar damage at the interface between the TiO2 layer and the substrate to that which we found for the VO2 layers. We suppose that this damage prevents oxygen migration from within the volume of the TiO2 substrate due to damaged or filled volume channels that are needed for oxygen migration.

We note that the transport studies shown in Fig. 1d and 5a were carried out ex-situ such that the TiO2 layer would have reverted to its non-conducting state, based on the gating measurements discussed above. Thus, there is no contribution from this layer to these ex-situ transport studies. In any case the sheet resistance of the gated 10 nm thick TiO2 film (on TiO2) is ~104 Ω/sq. is much higher than that of the gated 10 nm thick VO2 ( ~102 Ω/sq.) at room temperature. As the temperature is lowered the carriers within the TiO2 layer become localized18 and below the localization temperature the films become insulating (see Fig. S4a and S5).

X-ray photoelectron spectroscopy (XPS) measurements were performed on samples with tcap = 2.5 nm and 5 nm, in the pristine state and after gating. The spectrum of the gated samples is shown in Figure 6a. The V 2p3/2 core-level peaks become broader and shift to lower binding energies indicating a reduction of the oxidation state from V4+ to V3+. On the other end the TiO2 oxidation

8


Page 8 of 20

Page 9 of 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

state does not change upon gating in agreement with our previous studies18, where we show that TiO2 gating effect is volatile, contrary to that of VO2. Evidence of oxygen vacancy formation is provided by comparison of the reverse gating process in vacuum and in the presence of 250 mbar O2 (Fig. 6b). When a negative gate voltage, VG = - 2.6 V, is applied to a device that has previously been fully metallized, the conductivity decreases slowly in vacuum, but rather drops abruptly in the presence of oxygen, confirming the key role of oxygen vacancies in the metallization and reverse metallization of the device. We also note that any electrostatically induced surface charge accumulation effects by gating can be ruled out by these oxygen studies both for VO2 and also for the single TiO2 layers. In the latter case this confirms that the offset in the thickness dependence of the charge carrier density with TiO2 thickness, discussed above, arises from the TiO2 substrate. In summary, we have demonstrated that thin layers of VO2 can be metallized by ionic liquid gating even when the ionic liquid is separated from the VO2 layer by thin layers of insulating TiO2 up to 10 nm thick. Moreover, the VO2 metallization is accompanied by a massive expansion in thickness of the VO2 layer by up to 6.5 %. The metallization and layer expansion is non-volatile but can be completely reversed by applying an opposite gate voltage. These results cannot be accounted for by an electrostatic gating effect but rather are due to the formation of oxygen vacancies in the VO2 layer. Evidence for metallization of the TiO2 overlayer is found from independent experiments on thin layers of TiO2 grown directly on the TiO2 single crystalline substrate. These layers become metallic on gating but the effect is volatile and the layers immediately revert to their insulating state when the gate voltage is reduced. Moreover, these layers are volume metallized since the number of carriers induced by gating increases linearly with the thickness of the TiO2 layer, with an additional contribution presumed to be due to the formation of a surface electron gas in the TiO2 substrate. Our studies point to the important role played by the volume channels in the respective VO2 and TiO2 layers that are aligned with each other due to the similarity in structure of these layers and the high quality epitaxy of these layers, that allow for the migration of oxygen from the VO2 layer through the TiO2 layer induced by surface electric fields in the electric double layer provided by the remote ionic liquid. It is fascinating that oxygen vacancies can be induced remotely through an insulating

9


Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 20

layer and our results provide important new insights into the formation of novel conducting states. These concepts may allow for novel membrane structures for oxygen separation.

ASSOCIATED CONTENT Supporting information RHEED, AFM measurement, XRD fits, and gating response of TiO2 films on TiO2 (001) single crystals. This material is available free of charge via the internet at http://pubs.acs.org.

AUTHOR INFORMATION Corresponding Author *E-mail: [email protected] Notes The authors declare no competing financial interest.

ACKNOWLEDGEMENTS The authors thank T. Topuria and P. Rice for TEM sample characterization. Part of this research was carried out at the Stanford Synchrotron Radiation Lightsource, a Directorate of SLAC National Accelerator Laboratory and an Office of Science User Facility operated for the U.S. Department of Energy Office of Science by Stanford University.

REFERENCES (1) Waser, R.; Aono, M. Nat. Mater. 2007, 6, 833-840. (2) Yang, J. J.; Feng, M.; Matthew, D. P.; Douglas, A. A. O.; Duncan, R. S.; Chun Ning, L.; Williams, R. S. Nanotechnology 2009, 20, 215201. (3) Pickett, M. D.; Medeiros-Ribeiro, G.; Williams, R. S. Nat. Mater. 2013, 12, 114-117. (4) Borghetti, J.; Snider, G. S.; Kuekes, P. J.; Yang, J. J.; Stewart, D. R.; Williams, R. S. Nature 2010, 464, 873-876. (5) Jeong, J.; Aetukuri, N.; Graf, T.; Schladt, T. D.; Samant, M. G.; Parkin, S. S. P. Science 2013, 339, 1402-1405. (6) Jeong, J.; Aetukuri, N. B.; Passarello, D.; Conradson, S. D.; Samant, M. G.; Parkin, S. S. P. Proc. Natl. Acad. Sci. 2015, 112, 1013-1018.

10


Page 11 of 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

(7) Lee, J.; Panzer, M. J.; He, Y.; Lodge, T. P.; Frisbie, C. D. J. Amer. Chem. Soc. 2007, 129, 4532-4533. (8) Cho, J. H.; Lee, J.; He, Y.; Kim, B.; Lodge, T. P.; Frisbie, C. D. Adv. Mater. 2008, 20, 686690. (9) Dhoot, A. S.; Wimbush, S. C.; Benseman, T.; MacManus-Driscoll, J. L.; Cooper, J. R.; Friend, R. H. Adv. Mater. 2010, 22, 2529-2533. (10) Ye, J. T.; Inoue, S.; Kobayashi, K.; Kasahara, Y.; Yuan, H. T.; Shimotani, H.; Iwasa, Y. Nat. Mater. 2010, 9, 125-128. (11) Ueno, K.; Nakamura, S.; Shimotani, H.; Yuan, H. T.; Kimura, N.; Nojima, T.; Aoki, H.; Iwasa, Y.; Kawasaki, M. Nat. Nano. 2011, 6, 408-412. (12) Passarello, D.; Jeong, J.; Samant, M. G.; Parkin, S. S. P. Appl. Phys. Lett. 2015, 107, 201906. (13) Altendorf, S. G.; Jeong, J.; Passarello, D.; Aetukuri, N. B.; Samant, M. G.; Parkin, S. S. P. Adv. Mater. 2016. (14) Fujimoto, T.; Awaga, K. Physical Chemistry Chemical Physics 2013, 15, 8983-9006. (15) Yuyama, K.; Masuda, G.; Yoshida, H.; Sato, T. J. Power Sources 2006, 162, 1401-1408. (16) Martens, K.; Aetukuri, N.; Jeong, J.; Samant, M. G.; Parkin, S. S. P. Appl. Phys. Lett. 2014, 104, 081918. (17) Samant, M. G.; Toney, M. F.; Borges, G. L.; Blum, L.; Melroy, O. R. J. Phys. Chem. 1988, 92, 220-225. (18) Schladt, T. D.; Graf, T.; Jeong, J.; Aetukuri, N.; Li, M.; Fantini, A.; Jiang, X.; Samant, M.; Parkin, S. S. P. ACS Nano 2013, 7, 8074–8081.

Figure Captions Figure 1. Metallization by ionic liquid gating of VO2 films through TiO2 (001) films. (a) High resolution Cu Kα θ-2θ x-ray diffraction patterns of 2.5 nm, 5 nm, 10 nm and 17 nm thick TiO2 (001) films deposited on 10 nm VO2 (001), and a bare 10 nm VO2 film. The films are grown on TiO2 (001) single crystals and are tensile strained with the c-axis oriented out of plane. (b) Crosssectional high angle annular dark field (HAADF) STEM image of a 10 nm TiO2 / 10 nm VO2 deposited on a TiO2 (001) single crystal. (c) Schematic of an electrical double layer transistor (EDLT) device and it`s working principle upon ionic liquid gating. (d) Resistivity versus temperature measurements for 2.5 nm, 5 nm and 10 nm thick TiO2 films grown on top of 10 nm thick VO2 films tensile strained on TiO2 (001) single crystals. The curves show the films in their pristine state (black), after gating at VG = 2.6 V (red) and after reverse gating at VG = -2.6 V (blue).

Figure 2. In situ structural characterization of IL gated TiO2/VO2 films.

11


Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 20

In situ θ-2θ x-ray diffraction patterns for (a) 2.5 nm, (b) 5 nm, (c) 10 nm and (d) 17 nm thick TiO2 on 10 nm VO2 films grown on TiO2 (001) single crystals. XRD measurements were performed at Stanford Synchrotron Radiation Laboratory (SSRL). XRD was performed in pristine state, after gating at VG = 2.6 V, at 0 V and after reverse gating at VG = -2.6 V.

Figure 3. Structural changes of VO2 / TiO2 (001) thin films buried under TiO2 (001) films upon ionic liquid gating. Ex situ high resolution Cu Kα θ-2θ x-ray diffraction patterns (XRD) for (a) 2.5 nm (b) 5 nm and (c) 10 nm thick TiO2 (001) films deposited on 10 nm VO2 (001). (d) High resolution Cu Kα θ-2θ x-ray diffraction patterns for 10 nm thick VO2 film grown on a TiO2 (001) single crystal. For all the samples XRD was measured in pristine state (black), after gating at VG = 2.6 V (red) and after reverse gating at VG = -2.6 V (blue). Gating and reverse gating were performed at 270 K and the ionic liquid was washed off before performing XRD measurements. Fits to these data are shown in Figs. S2 and S3 from which the lattice expansion values were determined.

Figure 4. Reciprocal space maps of pristine and gated TiO2/VO2 films. Reciprocal space maps for a 10 nm TiO2 film grown on 10nm VO2 / TiO2 (001) in pristine state (a) and after gating (b). The TiO2 substrate reciprocal lattice units (r. l. u.) are used.

Figure 5. Transport measurements of TiO2 / VO2 and TiO2 films and structure of TiO2 film. (a) Sheet resistance vs temperature for 2.5 nm, 5 nm and 10 nm thick TiO2 films on 10 nm VO2 (001) after gating for 3, 6 and 12 hours. Ionic liquid gating was performed at 270 K and the ionic liquid was washed off prior transport measurements. The lattice expansion for the same films versus time is included in Fig. S5a. (b) Electron carrier densities at 100 K and 150 K from Hall measurements for devices created from 2.5 nm, 5 nm and 10 nm TiO2 films deposited on TiO2 (001) single crystal substrates and substrate annealed in oxygen for 30 minutes at 450 °C. (c) Cross-section HAADF-STEM image of a 5 nm TiO2 deposited on a TiO2 (001) single crystal.

12


Page 13 of 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

Figure 6. The influence of oxygen on ionic liquid gating. (a) X-ray Photoelectron Spectroscopy for 2.5 nm and 5 nm TiO2 films on 10 nm VO2 (001). V 2p and Ti 2p core-level spectra for pristine and gated states. Gating was performed at 270 K and a gate voltage of VG = 2.6 V was applied. The ionic liquid was washed off prior performing the XPS measurements. (b) Influence of oxygen on ISD during the reverse gating process at VG = -2.6 V and VDS = 100 mV, for a 5 nm TiO2 on 10 VO2 thin film grown on TiO2 (001) single crystals. Note that the measurements are two wires and they were performed on two different devices. The reverse gating process is compared in vacuum and in 250 mbar of oxygen.

13


Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 1.

14


Page 14 of 20

Page 15 of 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

Figure 2.

15


Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 3.

16


Page 16 of 20

Page 17 of 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

Figure 4

17


Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 5.

18


Page 18 of 20

Page 19 of 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

Figure 6.

19


Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

34x24mm (300 x 300 DPI)


Page 20 of 20

Metallization of Epitaxial VO2 Films by Ionic Liquid ... - ACS Publications

Recommend Documents