Characterization of atomic force microscopy written conducting nanowires at LaAlO<sub>3</sub>/SrTiO<sub>3</sub> interfaces

(1)

Article

Reference

Characterization of atomic force microscopy written conducting nanowires at LaAlO

₃

/SrTiO

₃

interfaces

BOSELLI, Margherita, et al.

Abstract

The realization of conducting nanostructures at the interface between LaAlO3 and SrTiO3 is an important step towards the realization of devices and the investigation of exotic physical regimes. We present here a detailed study of the conducting nanowires realized using the atomic force microscopy writing technique. By comparing experiments with numerical simulations, we show that these wires reproduce the ideal case of nanoconducting channels defined in an insulating background very well and that the tip bias is a powerful knob to modulate the size of these structures. We also discuss the role of the air humidity that is found to be a crucial parameter to set the size of the tip-sample effective interaction area.

BOSELLI, Margherita, et al . Characterization of atomic force microscopy written conducting nanowires at LaAlO

₃

/SrTiO

₃

interfaces. Applied Physics Letters , 2016, vol. 108, no. 6, p.

061604

DOI : 10.1063/1.4941817

Available at:

http://archive-ouverte.unige.ch/unige:84238

Disclaimer: layout of this document may differ from the published version.

(2)

Characterization of atomic force microscopy written conducting nanowires at LaAlO3/SrTiO3 interfaces

M. Boselli, D. Li, W. Liu, A. Fête, S. Gariglio, and J.-M. Triscone

Citation: Applied Physics Letters 108, 061604 (2016); doi: 10.1063/1.4941817 View online: http://dx.doi.org/10.1063/1.4941817

View Table of Contents: http://scitation.aip.org/content/aip/journal/apl/108/6?ver=pdfcov Published by the AIP Publishing

Articles you may be interested in

Photoconductive response of a single Au nanorod coupled to LaAlO3/SrTiO3 nanowires Appl. Phys. Lett. 106, 211101 (2015); 10.1063/1.4921750

Fabricating superconducting interfaces between artificially grown LaAlO3 and SrTiO3 thin films APL Mater. 2, 012102 (2014); 10.1063/1.4854335

Direct imaging of LaAlO3/SrTiO3 nanostructures using piezoresponse force microscopy APL Mater. 1, 052110 (2013); 10.1063/1.4831855

Creation of a two-dimensional electron gas and conductivity switching of nanowires at the LaAlO3/SrTiO3 interface grown by 90o off-axis sputtering

Appl. Phys. Lett. 103, 071604 (2013); 10.1063/1.4817921

Nano-patterning of the electron gas at the LaAlO3/SrTiO3 interface using low-energy ion beam irradiation Appl. Phys. Lett. 102, 201610 (2013); 10.1063/1.4807785

Reuse of AIP Publishing content is subject to the terms at: https://publishing.aip.org/authors/rights-and-permissions. Download to IP: 129.194.8.73 On: Fri, 03 Jun 2016

(3)

Characterization of atomic force microscopy written conducting nanowires at LaAlO

₃

/SrTiO

₃

interfaces

M.Boselli,^a)D.Li,W.Liu,A.F^ete,S.Gariglio,and J.-M.Triscone

Department of Quantum Matter Physics, University of Geneva, 24 Quai Ernest-Ansermet, 1211 Geneva, Switzerland

(Received 14 December 2015; accepted 1 February 2016; published online 11 February 2016) The realization of conducting nanostructures at the interface between LaAlO₃ and SrTiO₃is an important step towards the realization of devices and the investigation of exotic physical regimes.

We present here a detailed study of the conducting nanowires realized using the atomic force microscopy writing technique. By comparing experiments with numerical simulations, we show that these wires reproduce the ideal case of nanoconducting channels defined in an insulating background very well and that the tip bias is a powerful knob to modulate the size of these structures. We also discuss the role of the air humidity that is found to be a crucial parameter to set the size of the tip-sample effective interaction area.V^C 2016 AIP Publishing LLC.

[http://dx.doi.org/10.1063/1.4941817]

The conducting interface between LaAlO₃and SrTiO₃¹ is a fascinating system which combines the properties emerging at low dimensions with the physics of complex oxides.^2,3This interfacial electronic system is characterized by a strong spin-orbit coupling related to the breaking of inversion symmetry^4–6 and it undergoes a superconducting transition below300 mK.⁷Its electronic properties are also strongly tunable by electric field effect.^4,8,9 Realizing one- dimensional (super)conducting structures opens the way to the investigation of exotic regimes. Such nanowires could, for instance, be suitable for detecting Majorana particles.^10,11 Recently, several techniques have been used to create nanostructures at this interface: electron-beam lithography,^12–14 field-effect gating,¹⁵ growth in tilted structures,¹⁶ and atomic force microscopy (“AFM-writing technique”).^17,18 Among them, the AFM-writing technique is a versatile tool to create conducting channels down to few nanometers in size and has been used to perform a series of remarkable experiments.^19–23By applying a positive electric bias to the tip of an AFM, the conductance of an insulating heterostructure with a LaAlO₃ layer thickness of 3 u.c. can be switched ON locally.¹⁷ By reversing the bias polarity, the conductance is switched OFF.

The microscopic mechanism responsible for the change of the conducting state with this approach is still unclear. It could be related to the formation/replenishment of oxygen vacancies at the LaAlO₃ surface during the writing¹⁷ or to the screening effect due to the ionization of surface adsorbates.²⁴ Experimentally, it has been shown that the AFM- writing succeeds only in air.²⁴The role of the writing parameters and the effect of the relative humidity have however not been fully explored yet. A further study of these quanti- ties is important for the realization of customized nano- devices and for the understanding of the writing mechanism.

Here, we present a detailed characterization of the effects of the voltage bias applied to the AFM tip, and we discuss the role of the relative air humidity during the writing process.

Numerical simulations show that the evolution of the total conductance during the displacement of the tip between two electrodes is a good quantity to characterize the properties of the conducting wires and the sample background. From this analysis, we deduce the width of the nanowire for different tip biases. Finally, in order to investigate the effect of the relative humidity, we model the water meniscus present between the AFM tip and the sample surface and we compare it with the experiments. We show that the meniscus is a key element for the efficiency of the writing mechanism and that the relative humidity strongly modulates the wire size.

LaAlO₃ films are deposited on TiO₂-terminated (001) SrTiO₃single crystals by pulsed laser deposition. The growth is performed at 800C in an oxygen pressure of 10⁴mbar and annealed for 1 h at 520C in an atmosphere of 200 mbar of oxygen. The fluence of the laser pulses is0.7 J/cm², and the repetition rate is 1 Hz. The thickness of LaAlO₃is monitored during the growth using reflection high-energy electron diffraction (RHEED). We found that the optimal thickness for the writing experiments is 2.8 u.c.

Figure 1 shows a device used for the writing experiments. Metallic contacts are defined around the writing area by patterning the surface using optical lithography. We then use Ar ion milling to etch the electrodes area (14 nm deep) before the in-situ deposition of Ti/Au (11 nm/3 nm) by e- beam evaporation; in this way, the height of the electrodes matches the LaAlO₃ surface within a few nanometers. A constant current of 100 nA is used for the measurements.

The evolution of the conductance is monitored in real time by measuring the voltage drop between the contacts in a 2- point configuration.³³

The writing experiments are performed using aMultimode DI3 AFM equipped with aNanonis Kolibri Sensor.²⁵ In this system, a metallic tip, mounted on a quartz resonator, oscillates 1.5 nm above the sample surface with an amplitude of 0.3 nm.

The tip is made of tungsten and has a radius of110 nm. This sensor does not entail the presence of a laser to track the tip movement, hence avoiding changes in conductivity due to pho- tocarriers excitation in SrTiO₃.

a)Electronic address: margherita.boselli@unige.ch

0003-6951/2016/108(6)/061604/5/$30.00 108, 061604-1 VC2016 AIP Publishing LLC

APPLIED PHYSICS LETTERS108, 061604 (2016)

(4)

As a preliminary step, two conducting pads are defined around the electrodes scanning a region of 1010lm², at a speed of 25lm/s with 8 V applied to the tip. This procedure improves the electrical contact between the interface and the electrodes. The first parameter we focus on is the electric bias applied to the tip,Vb, keeping constant the relative humidity and the scanning speed. Figure2(a)shows the evolution of the conductance upon writing for different values of Vb. For a relative humidity of 56%, we have found that the writing is possible forVb2 V. The conductance increases abruptly when the tip approaches the electrode (at position 12lm in Figure 2(a)). We observe that the conductance variation evolves non-linearly with the tip bias, changing from a few nS atVb¼3 V up to 0.2lS at Vb¼8 V. The effect of the relative humidity (RH) of the environment where the experiments are performed is shown in Figure 2(b). We clearly observe that when moving from RH56%

to RH31%, the effect of writing is strongly reduced.

In order to have a better understanding of the nanowire writing process, we performed a series of simulations using COMSOL MULTIPHYSICS, a software based on the finite element numerical method. The model reproduces the geometry of the device with six electrodes surrounding the writing area (see Figure3(a)). The interface is modeled as a slab of insulating material with a thickness of 10 nm and a dielectric constant of 300. Although the SrTiO₃ dielectric constant is electric field dependent (also at room temperature), the model is not very sensitive to the value of the dielectric constant, and there is almost no perceptible difference between

conductance curves from simulations run with a value of 25 or 300 for the dielectric constant using 0.5lm steps. The writing process is described as a local switching of the electrical resistance from the background value to the interface conducting value (sheet resistance in the order of 20 kX).

In the calculations, we set constant the current applied between the source and drain contacts, as in the experiments.

The total conductance between the two main electrodes, Gtot, is the sum of the background conductance, the conductance of the wire, Gw¼^w_l_R¹_w (wherelis the wire length, w the wire width, andRwis the interface metallic sheet resistance) and the conductance of any other element present in the system, for instance, additional wires.

We first study the effect of the background sheet conductance (Gb¼_R¹_b) on the shape of the conductance jump when the tip approaches the second electrode. Figure 3(b) plots the change in total conductance (䉭Gtot¼GtotG0, where G0 is the total conductance at the beginning of the writing) for different values of the sheet resistance of the background. We note that the higher Rb, the sharper the change in conductance. Comparing the conductance jumps observed in the experiments with the results of the simulation shows that the conducting wire is surrounded by a very resistive background. Therefore, the experimental situation is rather close to the ideal case of a one-dimensional conducting channel in an insulating background.

FIG. 1. Experimental details, optical images of 5 mm, 1 mm, and 100lm scale are shown. 9 devices are realized on a 55 mm²LaAlO3/SrTiO3sam- ple. The writing area, 2020lm²large, is surrounded by metallic electrodes in contact with the interface. The green dashed line between the electrodes shows the trajectory of the AFM-tip. In the upper part of the figure, an AFM image of the region around the electrode is shown. The AFM tip is sketched in grey and connected to the voltage source through a 1 GX resistance to prevent large currents. The close-up of the LaAlO3 surface (AFM image 55 lm²) reveals the step and terrace topography of the layer.

FIG. 2. The total conductance,Gtot, is plotted as a function of the position of the AFM tip between the electrodes. The speed of the tip is 200 nm/s. In (a), the curves are taken using a different tip bias,Vb, at RH¼56%. In the inset is shown a zoom of the curves taken withVb¼3 V andVb¼2 V. In (b), the conductance for lines written atVb¼8 V is shown for two values of RH. In the inset, a zoom of the total conductance for RH¼31% is shown. In all the graphs, the value of the background conductance is set equal to 0.12lS, the background value of the curve taken withVb¼8 V and RH¼56%, to better compare the conductance jumps.

061604-2 Boselliet al. Appl. Phys. Lett.108, 061604 (2016)

(5)

In the limit of a very conducting wire with respect to the background, the value ofDG_tot provides direct information on the ratio_R^w

w. In order to determinew, one notices that the sheet resistance of the metallic region is in the range of 10–50 kX.²⁶Several experimental works have indeed shown that the interface sheet resistance does not depend markedly on the LaAlO₃surface conditions.^27,28

In Figure 4(a), the experimental behavior of the total conductance as a function of the tip position for a wire written using Vb¼8 V and RH¼56% is compared with the simulated one. By setting Rw equal to 18 kX, the experimental curve is very well reproduced considering a wire 46 nm wide. Experimentally, an estimation of the wire width can be obtained by “cutting” the nanowire. As shown in Figure4(b), scanning the tip perpendicular to the channel and applying a negative bias (Vb¼ 5 V in our experiment) switches locally OFF the conductivity.¹⁷The change in the wire conductance during cutting allows the wire width to be estimated; here, 60 nm, in good agreement with the simulation (Figure 4(b)). We performed the same simulation for a wire written applyingVb¼5 V to the AFM tip.

Considering again Rw¼18 kX, the experimental curve is nicely reproduced with a wire 10.5 nm wide. For wires written with Vb¼2 V and Vb¼3 V, the simulations yield a wire’s width below 5 nm.

Figure4(c)shows the width of the wires, extracted from this analysis, as a function of the tip bias. Looking at the data, one observes that the wire width does not evolve linearly with the tip bias. For tip voltages below 4 V, the wire is just few nm wide; atVb¼5 V, its width increases to 10 nm and reaches 50 nm for Vb¼8 V. The behavior observed

here is in agreement with the results published by Cen et al.¹⁷ The modulation of the tip bias is thus a powerful knob to control the dimension of the written nanostructures.

The simulations we just discussed describe the propaga- tion of a conducting region in an insulating background, inde- pendent of the microscopic process that generates the local conductivity. The experiments in different relative humidities have shown (see Figure2(b)), however, that the switching is also affected by the environmental conditions that can

FIG. 3. (a) Simulation of the electrical potential (color scale on the side) during the writing of a 20 nm wide wire with a sheet resistance of 10 kXin a background withRb¼100 MX. The size of the area is 6060lm². (b) Behavior ofDGtotas a function of the position of the AFM tip between the two electrodes for differentRbvalues. The width and the sheet resistance of the wire are 20 nm and 10 kX, respectively.

FIG. 4. In panel (a), the experimental behavior of the total conductance as a function of the tip position between the electrodes is compared with the simulation. The parameters used in the experiment areVb¼8 V and RH¼56%.

In the simulation, a wire 46 nm wide withRw¼18 kXis defined on a background withRb¼800 MX. Using this value forRb, the shape of the curve is perfectly reproduced. The experimental value ofG0is however not reproduced by the simulation possibly because of the presence of parallel conducting paths created during the etching procedure. In order to match the experimental value ofG0, the simulated curve is vertically shifted. Note that this shift is not changing the conclusions we get. In (b), behavior of the conductance as a function of the tip position during a cut, the value is normal- ized with respect to the maximum conductance. The wire is erased using Vb¼ 5 V. The curve is fitted using a sigmoidal function, and the experimental wire width is estimated from the full width at half maximum of the fit derivative. In (c), the wire width extracted from the simulations is plotted as a function of the tip bias.

(6)

strongly modulate the dimensions of the written area for a given set of writing parameters. Previous studies have also revealed that an atmosphere lacking water vapor prevents the writing process from happening.²⁴In the following, we investigate the role of the relative humidity on the tip-sample interaction area.

In the narrow gap between the tip and the sample surface, the water present in the air condenses by capillarity.

The relationship between the relative vapor pressure of the environment and the size of the water meniscus is described by the Kelvin equation²⁹

RH¼exp kk

1 r1

l

:

In this equation,kkis the Kelvin length (for water at 25C kk¼0.52 nm), andrandlare the principal radii of curvature of the meniscus (see the inset of Figure5(a)). Following the work of Butt and Kappl³⁰ for the condensation of water between a sphere and a plane,rand lcan be expressed as a function of the same quantity and depend on the tip radius,R, the tip-sample distance,D, and the thickness of the water layer present on the surface of LaAlO₃,d.

In Figure5(a)is illustrated the evolution of the meniscus diameter, 2l, as a function of RH for different values of D.

We observe that an RH threshold for the formation of the water meniscus exists and that lifting the tip from the surface

(i.e., increasing D) pushes this value to higher RH.

Considering the tip-sample geometry of our AFM,³⁴ this threshold is 30% in agreement with the experiments (see Figure2(b)).

We also performed a set of measurements at different relative humidities writing several wires at Vb¼8 V in contact mode and measuring their width using the “cutting”

technique.³⁵ The results are shown in Figure 5(b) together with calculations of the meniscus size for the case of a hard contact (D¼0 nm). We find that the wire width is always larger than the meniscus size. Since it is found that at low tip voltage the wire width is comparable with the meniscus size,¹⁷ one suspects that the differences observed here are related to the large writing voltage. One can also notice that the evolution of the wire width with RH is rather different from one of the meniscus size, the former increasing faster than the latter. Future studies, although experimentally rather challenging, could shed more light on this behavior.

From this, we can conclude that the presence of the water meniscus is a necessary element for the efficiency of the writing process. On the one hand, the water focalizes the electric field present between the tip and the surface,³¹ and on the other hand, it may play a role in the surface adsorbates ionization involved in the microscopic mechanism of the AFM-writing, the so-called “water cycle mechanism.”²⁴

In summary, we have presented a study of the properties of the conducting nanostructures defined at the interface between LaAlO₃and SrTiO₃ using the AFM writing technique. By comparing the experiments with numerical simulations, we show that the written wires behave very much like the ideal case of conducting channels in an insulating background. We use simulations and cutting experiments to determine the size of the wires whose dimension can be in the nanometer range for low positive bias applied to the AFM tip. We also discuss the role of the relative humidity in the definition of the tip effective dimension and its conse- quences on the efficiency of the writing process.

We acknowledge Jeremy Levy and his research group at the University of Pittsburgh for enlightening discussions. We thank Dr. Cedric Blaser and Iaroslav Giaponenko for experimental advices and Marco Lopes and Sebastien Muller for technical support. This research was supported by the Swiss National Science Foundation and by the European Research Council under the European Unions Seventh Framework Programme (No. FP7/20072013)/ERC Grant Agreement No. 319286 (Q-MAC).

1A. Ohtomo and H. Y. Hwang, “A high-mobility electron gas at the LaAlO3/SrTiO3heterointerface,”Nature427(6973), 423–426 (2004).

2P. Zubko, S. Gariglio, M. Gabay, P. Ghosez, and J.-M. Triscone,

“Interface physics in complex oxide heterostructures,” Annu. Rev.

Condens. Matter Phys.2(1), 141–165 (2011).

3H. Y. Hwang, Y. Iwasa, M. Kawasaki, B. Keimer, N. Nagaosa, and Y.

Tokura, “Emergent phenomena at oxide interfaces,” Nat. Mater. 11(2), 103–113 (2012).

4A. D. Caviglia, M. Gabay, S. Gariglio, N. Reyren, C. Cancellieri, and J.- M. Triscone, “Tunable rashba spin-orbit interaction at oxide interfaces,”

Phys. Rev. Lett.104(12), 126803 (2010).

5M. Ben Shalom, M. Sachs, D. Rakhmilevitch, A. Palevski, and Y. Dagan,

“Tuning spin-orbit coupling and superconductivity at the SrTiO3/LaAlO3

interface: A magnetotransport study,” Phys. Rev. Lett.104(12), 126802 (2010).

FIG. 5. In (a), the meniscus diameter, 2l, is plotted as a function of the relative humidity of the environment. Each curve was computed using a different tip-sample distance (D). In the inset, the geometry of the water meniscus is sketched. The tip is modelled as a sphere of radiusR¼40 nm; on the surface of LaAlO3, we consider a layer of water 1 nm thick.³²(c) The evolution of the wire width with RH for a wire written in contact mode using Vb¼8 V. In the same plot are reported the calculations of the meniscus size for two tip radii, 15 nm and 40 nm, atD¼0 nm.

(7)

6A. Fete, S. Gariglio, A. D. Caviglia, J.-M. Triscone, and M. Gabay,

“Rashba induced magnetoconductance oscillations in the LaAlO3-SrTiO3

heterostructure,”Phys. Rev. B86(20), 201105 (2012).

7N. Reyren, S. Thiel, A. D. Caviglia, L. Fitting Kourkoutis, G. Hammerl, C. Richter, C. W. Schneider, T. Kopp, A.-S. Ruetschi, D. Jaccard, M.

Gabay, D. A. Muller, J.-M. Triscone, and J. Mannhart, “Superconducting interfaces between insulating oxides,” Science 317(5842), 1196–1199 (2007).

8S. Thiel, G. Hammerl, A. Schmehl, C. W. Schneider, and J. Mannhart,

“Tunable quasi-two-dimensional electron gases in oxide heterostructures,”

Science313(5795), 1942–1945 (2006).

9A. D. Caviglia, S. Gariglio, N. Reyren, D. Jaccard, T. Schneider, M.

Gabay, S. Thiel, G. Hammerl, J. Mannhart, and J.-M. Triscone, “Electric field control of the LaAlO3/SrTiO3 interface ground state,” Nature 456(7222), 624–627 (2008).

10R. M. Lutchyn, J. D. Sau, and S. Das Sarma, “Majorana fermions and a topological phase transition in semiconductor-superconductor heterostructures,”Phys. Rev. Lett.105(7), 077001 (2010).

11Y. Oreg, G. Refael, and F. von Oppen, “Helical liquids and majorana bound states in quantum wires,”Phys. Rev. Lett.105(17), 177002 (2010).

12D. Stornaiuolo, S. Gariglio, N. J. G. Couto, A. Fete, A. D. Caviglia, G.

Seyfarth, D. Jaccard, A. F. Morpurgo, and J.-M. Triscone, “In-plane electronic confinement in superconducting LaAlO3/SrTiO3 nanostructures,”

Appl. Phys. Lett.101(22), 222601 (2012).

13J.-W. Chang, J. Song, J. S. Lee, H. Noh, S. K. Seung, L. Baasandorj, S. G.

Lee, Y.-J. Doh, and J. Kim, “Quantum electrical transport in mesoscopic LaAlO3/SrTiO3 heterostructures,” Appl. Phys. Express 6(8), 085201 (2013).

14P. P. Aurino, A. Kalabukhov, N. Tuzla, E. Olsson, T. Claeson, and D.

Winkler, “Nano-patterning of the electron gas at the LaAlO3/SrTiO3inter- face using low-energy ion beam irradiation,” Appl. Phys. Lett.102(20), 201610 (2013).

15S. Goswami, E. Mulazimoglu, L. M. K. Vandersypen, and A. D. Caviglia,

“Nanoscale electrostatic control of oxide interfaces,” Nano Lett. 15(4), 2627–2632 (2015).

16A. Ron and Y. Dagan, “One-dimensional quantum wire formed at the boundary between two insulating LaAlO3/SrTiO3interfaces,”Phys. Rev.

Lett.112(13), 136801 (2014).

17C. Cen, S. Thiel, G. Hammerl, C. W. Schneider, K. E. Andersen, C. S.

Hellberg, J. Mannhart, and J. Levy, “Nanoscale control of an interfacial metal-insulator transition at room temperature,”Nat. Mater.7(4), 298–302 (2008).

18C. Cen, S. Thiel, J. Mannhart, and J. Levy, “Oxide nanoelectronics on demand,”Science323(5917), 1026–1030 (2009).

19J. P. Veazey, G. Cheng, P. Irvin, C. Cen, D. F. Bogorin, F. Bi, M. Huang, C.-W. Bark, S. Ryu, K.-H. Cho, C.-B. Eom, and J. Levy, “Oxide-based

platform for reconfigurable superconducting nanoelectronics,”

Nanotechnology24(37), 375201 (2013).

20D. F. Bogorin, C. W. Bark, H. W. Jang, C. Cen, C. M. Folkman, C.-B.

Eom, and J. Levy, “Nanoscale rectification at the LaAlO3/SrTiO3inter- face,”Appl. Phys. Lett.97(1), 013102 (2010).

21G. Cheng, P. F. Siles, F. Bi, C. Cen, D. F. Bogorin, C. W. Bark, C. M.

Folkman, J.-W. Park, C.-B. Eom, G. Medeiros-Ribeiro, and J. Levy,

“Sketched oxide single-electron transistor,” Nat. Nanotechnol. 6(6), 343–347 (2011).

22Y. Xie, C. Bell, T. Yajima, Y. Hikita, and H. Y. Hwang, “Charge writing at the LaAlO3/SrTiO3surface,”Nano Lett.10(7), 2588–2591 (2010).

23G. Cheng, M. Tomczyk, S. Lu, J. P. Veazey, M. Huang, P. Irvin, S. Ryu, H. Lee, C.-B. Eom, C. S. Hellberg, and J. Levy, “Electron pairing without superconductivity,”Nature521(7551), 196–199 (2015).

24F. Bi, D. F. Bogorin, C. Cen, C. W. Bark, J.-W. Park, C.-B. Eom, and J.

Levy, “Water-cycle mechanism for writing and erasing nanostructures at the LaAlO3/SrTiO3interface,”Appl. Phys. Lett.97(17), 173110 (2010).

25Kolibrisensor, “Advanced quartz sensor technology for NC-AFM,”

(SPECS Surface Nano Analysis GmbH, 2015).

26C. Cancellieri, N. Reyren, S. Gariglio, A. D. Caviglia, A. Fete, and J.-M.

Triscone, “Influence of the growth conditions on the LaAIO3/SrTiO3inter- face electronic properties,”Europhys. Lett.91(1), 17004 (2010).

27E. Lesne, N. Reyren, D. Doennig, R. Mattana, H. Jaffres, V. Cros, F.

Petroff, F. Choueikani, P. Ohresser, R. Pentcheva, A. Barthelemy, and M.

Bibes, “Suppression of the critical thickness threshold for conductivity at the LaAlO3/SrTiO3interface,”Nat. Commun.5, 4291 (2014).

28Y. Xie, Y. Hikita, C. Bell, and H. Y. Hwang, “Control of electronic con- duction at an oxide heterointerface using surface polar adsorbates,”Nat.

Commun.2, 494 (2011).

29R. J. Hunter,Foundations of Colloid Science, 2nd ed. (Oxford University Press, 2001).

30H.-J. Butt and M. Kappl, “Normal capillary force,”Adv. Colloid Interfac.

146(1–2), 48–60 (2009).

31C. Blaser and P. Paruch, “Subcritical switching dynamics and humidity effects in nanoscale studies of domain growth in ferroelectric thin films,”

New J. Phys.17(1), 013002 (2015).

32P. Maksymovych, S. Jesse, P. Yu, R. Ramesh, A. P. Baddorf, and S. V.

Kalinin, “Polarization control of electron tunneling into ferroelectric surfaces,”Science324(5933), 1421–1425 (2009).

33After the contacts deposition process, the total resistance between the electrodes is10 MX.

34As reported before, the tip-sample distance is 1.5 nm, the oscillation amplitude is 0.3 nm and the tip radius is 110 nm.

35For this analysis, the writing is performed using aMFP-3D InfinityAFM fromAsylum Research-Oxford Instrumentsequipped with a standard canti- lever. The tip bias during the cut was set atVb¼ 0.5 V.