Ir coated AFM tips for resistive switching measurements

Applied Surface Science 257 (2011) 7627–7632

Contents lists available at ScienceDirect

Applied Surface Science journal homepage: www.elsevier.com/locate/apsusc

The thermal stability of Pt/Ir coated AFM tips for resistive switching measurements M. Wojtyniak a,b,∗,1 , K. Szot a,b , R. Waser b,c a b c

Institute of Physics, University of Silesia, 40-007 Katowice, Poland Institut für Festkörperforschung, Forschungszentrum Jülich, 52425 Jülich, Germany Institut für Werkstoffe der Elektrotechnik, RWTH Aachen, 52056 Aachen, Germany

a r t i c l e

i n f o

Article history: Received 4 January 2011 Received in revised form 13 January 2011 Accepted 30 March 2011 Available online 9 April 2011 Keywords: Atomic force microscopy Tip temperature stability Pt coating Resistive switching

a b s t r a c t In this paper, we focus on the thermally treated atomic force microscope tips used in the investigation of the resistive switching phenomenon. Since the resistive switching phenomenon is often connected with the red-ox process, it is crucial to investigate the influence of oxidizing and reducing conditions at elevated temperatures on typical AFM tips. To fully characterize the influence of different conditions on the tip properties we used several techniques such as: X-ray photoelectron spectroscopy, scanning electron microscopy, transmission electron microscopy and local-conductivity atomic force microscopy. The chemical composition as well as the topography and morphology of the most popular Pt/Ir coated silicon tips were investigated. The influence of thermal treatment on the tip apex was also imaged and the changes in the electrical behavior of the tip coating were observed. Applied temperatures ranges were: 500–700 ◦ C for oxidizing conditions (air) and 300–700 ◦ C for reducing conditions (vacuum 10−6 Torr), the annealing time was set to 0.5 h. Results yielded the formation of Pt2 Si and SiO2 on the tip surface. The Pt tends to agglomerate into particles over time, depending on the temperature and conditions. The tip apex radius increases while the electrical conductivity decreases with the temperature. In conclusion, even the lowest applied temperature leads to changes in the tip properties, while these changes are much more pronounced under oxidizing conditions. © 2011 Elsevier B.V. All rights reserved.

1. Introduction Understanding the resistive switching phenomenon [1] demands deep insight into local conductivity on the nanoscale. This task can be accomplished only with the help of conducting tip and atomic force microscopy, which together constitute the local-conductivity AFM (LC-AFM) technique (Fig. 1). Since the investigation of the local conductivity of nanometer-size filaments or activation energy on the nano-scale requires measurements under both oxidizing and reducing conditions at elevated temperatures, changes in the properties of the typical conducting tips must be investigated. LC-AFM is not the only technique requiring conducting tips, more recently established AFM techniques also make use of such tips, for example: piezoresponse AFM (PMF) [2], Kelvin-probe AFM (SKPM) [3] or scanning capacitance microscopy

∗ Corresponding author. Present address: Institut für Festkörperforschung, Forschungszentrum Jülich, Wilhelm-Johnen-Strasse, 52428 Jülich, Germany. Tel.: +49 02461 61 2551; fax: +49 2461 61 2550. E-mail address: [email protected] (M. Wojtyniak). 1 Permanent address: Institute of Physics, University of Silesia, Uniwersytecka 4, 40-007 Katowice, Poland. Tel.: +48 32 359 1725; fax: +48 32 258 8431. 0169-4332/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.apsusc.2011.03.149

(SCM) [4]. Conducting tips are commercially available in a variety of forms ranging from solid metal to, most commonly, silicon with conductive coatings. Since solid metal tips are not very popular due to their low mechanical stability [5] we focus on the most popular – Pt/Ir coated silicon tips. Much work has been done on the mechanical characterization of such tips [6,7], but there is nearly no information about the behavior of the coating at elevated temperatures. Therefore, in this paper a set of tips were annealed at elevated temperatures under oxidizing and reducing conditions. The influence on the chemical composition, topography and electrical properties was investigated. In addition the tip apex radius before and after thermal treatment was measured. 2. Experimental details A set of silicon tips (PPP-CONTPt, Nanosensors), coated by means of sputter deposition with a 25 nm thick double layer of chromium and Pt/Ir with an initial 20 nm tip radius were used. Some of the tips were left without thermal treatment (reference) and the rest were annealed at various temperatures for 0.5 h: 500–700 ◦ C for oxidizing (air) and 300–700 ◦ C for reducing (vacuum 10−6 Torr) conditions. The heating and cooling rate was equal to several ◦ C/s

7628

M. Wojtyniak et al. / Applied Surface Science 257 (2011) 7627–7632

The equivalent resistivity per scanning area (R ) was calculated using current maps and statistic functions of the Gwyddion software. Then the size of the electric contact between the tip and the surface was estimated to be 10 nm × 10 nm and the resistivity per such a square was calculated followed by the calculation of R . The values from several areas in each sample were then averaged.

3. Results and discussion Fig. 1. Schematic presentation of LC-AFM technique.

on average. For X-ray photoelectron spectroscopy (XPS) measurements, we used a PHI 5600 spectrometer with monochromatized Al K␣, while the scanning electron microscopy (SEM) and transmission electron microscopy (TEM) images were obtained using an SEM Hitachi S4100 FEG and TEM Phillips CM 200, respectively. LCAFM measurements were performed using a JSPM-5200 scanning probe microscope. The areas of investigation were always on the same side of the probe as the cantilever and the tip. We therefore assumed that the coating was uniform at every point of the probe.

Fig. 2 shows the SEM images of Pt/Ir coated silicon cantilevers treated at various temperatures under oxidizing conditions. Fig. 2(a) corresponds to an untreated reference tip while Fig. 2(b)–(d) correspond to cantilevers annealed at 500, 600 and 700 ◦ C for 0.5 h under oxidizing (air) conditions. A comparison between the untreated (reference) tip and the tips annealed in air shows that the topography and morphology of Pt/Ir coating changes. While for the reference sample the coating is homogeneous and smooth, for higher temperatures hills becomes visible. Similar results (not presented) can be found for cantilevers annealed in vacuum, although the changes are much smaller. TEM measurements were performed in order to look closely into the

Fig. 2. SEM images of Pt/Ir tips: (a) without thermal treatment, (b) annealed at 500 ◦ C, (c) annealed at 600 ◦ C and (d) annealed at 700 ◦ C in air.

M. Wojtyniak et al. / Applied Surface Science 257 (2011) 7627–7632

7629

Table 1 Binding energies of Pt4f and Si2s lines for three samples – reference, annealed at 700 ◦ C under reducing conditions and annealed at 700 ◦ C under oxidizing conditions. Binding energy uncertainty ±0.03 eV. Sample

Pt4f (eV)

Si2s (eV)

Volume (%)

Reference Air 700 ◦ C

71.15 71.29

Vacuum 700 ◦ C

72.63

– 154.38 151.6 154.1

100 100 72.25 ± 10 27.75 ± 10

Table 2 The root mean square (RMS), average grain size (AGS), average grain height (AGH) and the R for all samples. The uncertainties are RMS = 0.1 nm, AGS = 1.2 nm, AGH = 1.3 nm. Sample

RMS (nm)

Reference

0.30

7.9

1.6

Air 500 C Air 600 ◦ C Air 700 ◦ C

1.24 2.38 4.92

19.6 28.4 56.0

4.4 9.0 17.5

Vacuum 300 ◦ C Vacuum 400 ◦ C Vacuum 500 ◦ C Vacuum 600 ◦ C Vacuum 700 ◦ C

0.39 0.51 0.67 0.79 0.89

9.8 10.9 14.8 16.3 18.4

2.2 2.5 3.4 3.8 4.3

◦

Fig. 3. TEM images of Pt/Ir tips: (a) without thermal treatment, (b) annealed at 700 ◦ C under reducing conditions and (c) annealed at 700 ◦ C under oxidizing conditions.

influence of heat treatment on the tip apex. In Fig. 3 the reference tip (a) and the tips annealed at 700 ◦ C in vacuum (b) and in air (c) are compared. The tip apex radius for the reference tip is around 20 nm, the same value as given by the manufacturer and reported from SEM measurements [8]. The polycrystalline character of the coating is visible as a grainy structure on the surface of the tip. Much more interesting is the image of the tip annealed at 700 ◦ C, which shows the bulk of single-crystalline silicon covered in the center by a thin layer of silicon oxide. Then on top of this layer several round dark particles can be found with various diameters ranging from several nanometers up to 30 nm. These particles are composed of Pt and a small fraction of Pt2 Si as will be shown in the XPS measurements. The Pt atoms tend to decrease the surface free energy. Thus

AGS (nm)

AGH (nm)

R () 6.61 ± 1.26 71.94 ± 5.35 452.25 ± 14.72 >10,000 58.12 ± 6.94 78.67 ± 6.35 84.58 ± 8.22 92.76 ± 7.44 108.55 ± 9.13

migration occurs on the surface and the islands are mainly formed due to the Ostwald ripening process. This behavior will have a huge impact on the electrical properties of the annealed tips, as will be shown from the LC-AFM measurements. The chemical composition of the tip surface was investigated using XPS presented in Fig. 4 and the peak binding energies are summarized in Table 1. In particular, the Pt4f and Si2s level photoemission spectra were analyzed. The Pt4f lines exhibits a spin–orbit ˆ doublet formed of asymmetric Doniach–Sunji c´ lines for all measured samples. This is a clear indication of a metallic state and a lack of Pt–O bonds [9]. It is known that the formation of platinum silicides starts from the relatively low temperature of 200 ◦ C [10] thus we expect the formation of Pt2 Si or PtSi. The binding energy values for Pt, Pt2 Si and PtSi are 71.3, 72.5 and 73.0 eV, respectively [11]. Thus for the reference tip the surface consists of platinum in a metallic state. For the tip annealed in air, the Pt4f peak is at 71.29 eV suggesting a small fraction of Pt2 Si besides Pt in the metallic state. For the tip annealed in vacuum, we find that the surface mostly consists of Pt2 Si and possibly a small fraction of PtSi. The data from Si2s peaks shows, as we would expect, that for the reference tip there is no silicon on the surface. The binding energy values for Si and SiO2 are 151.2 and 155.0 eV, respectively [12]. The result shows that the thermally treated tips are covered with SiO2 regardless of the annealing conditions applied. Hence for the tip annealed in vacuum the SiO2 component is only the fraction of surface silicon. The rest of the silicon comes from Pt2 Si, which corresponds well to the results from Pt4f spectra. Similar results were obtained by other authors [13–15]. The topography images of the same set of cantilevers are presented in Fig. 5(a)–(d). Averaged values of RMS and island sizes for several 1 ␮m × 1 ␮m scanning areas per sample are presented in Table 2. There is an increase in the coating roughness starting from the lowest annealing temperature for all conditions. In air at 500 ◦ C film stays fairly flat and homogeneous. As the annealing temperature was increased to 600 ◦ C the islands and holes became larger, thus platinum started to agglomerate forming partially connected patterns. After annealing at 700 ◦ C the sizes of island was further increased and they become more circular in shape. The RMS jumps to almost 5 nm, the AGS to 56 nm and the AGH to 17.5 nm, moreover the films break up and become discontinuous. The same effect was clearly visible in the TEM image.

7630

M. Wojtyniak et al. / Applied Surface Science 257 (2011) 7627–7632

Fig. 4. XPS spectra of Pt/Ir coating: (a, d) without thermal treatment, (b, e) annealed at 700 ◦ C under reducing conditions and (c, f) annealed at 700 ◦ C under oxidizing conditions.

Current maps of tips annealed in air presented in Fig. 6 mimic the behavior of the topography. For the reference sample (Fig. 6(a)) the polycrystalline character of the platinum coating is clearly visible, and many conducting spots with an average size of 8 nm can be found. With increasing annealing temperature the size of the conducting spots increases whereas the average current per scanning area decreases up to 700 ◦ C (Fig. 6(d)) where there is no current flow. The source of the rapid current decrease is twofold. First, as observed in the topography, platinum islands become discontinuous thus breaking the current paths. The connection between the ground electrode and the biased electrode (the scanning tip) is destroyed, thus there is no current flow. Second, the diffusion of Si and formation of SiO2 on the surface already shown in XPS measurements leads to a continuous decrease in conductivity. Island formation and growth is much slower for the vacuumannealed tips. The topography undergoes small changes and the

coating remains continuous up to the highest measured temperature. The topography images of the vacuum annealed cantilevers are presented in Fig. 7(a) and (c). The roughness and island size increase steadily with the annealing temperature up to 0.82 nm RMS, 18.4 nm AGS and 4.3 nm AGH for 700 ◦ C. These values are significantly lower than for the air-annealed cantilevers. This suggests that oxygen plays an important role in the formation of island structures and Pt surface mobility is thus enhanced. TEM measurements by other authors also confirm that annealing in oxygen and air leads to significantly larger Pt particles [16,17]. The local conductivity maps presented in Fig. 7(b) and (d) shows that the polycrystalline character of the reference sample is conserved. Large numbers of conducting grains can be found after annealing at both 300 ◦ C and 700 ◦ C in vacuum. The resistivity increases with the annealing temperature, as presented in Table 2. This is connected with the island formation and the variation

Fig. 5. Topography AFM images of Pt/Ir tips: (a) without thermal treatment, (b) annealed at 500 ◦ C, (c) annealed at 600 ◦ C and (d) annealed at 700 ◦ C in air.

M. Wojtyniak et al. / Applied Surface Science 257 (2011) 7627–7632

7631

Fig. 6. Local conductivity AFM images of Pt/Ir tips: (a) without thermal treatment, (b) annealed at 500 ◦ C, (c) annealed at 600 ◦ C and (d) annealed at 700 ◦ C in air.

Fig. 7. Topography and local conductivity AFM images of Pt/Ir tip annealed at (a)–(b) 300 ◦ C and (c)–(d) 700 ◦ C in vacuum.

in thickness and stoichiometry of the coating. Additionally, it is related to the formation of Pt2 Si, which has a higher resistivity value (30 ␮ cm [10]) compared with pure Pt (0.59 ␮ cm [18]). Moreover, the formation of surface SiO2 is increasing the resistivity of the tips. 4. Conclusion In summary, even the lowest applied temperature results in changes in the tip properties. The platinum coating tends to agglomerate and Pt2 Si and SiO2 is formed on the tip surface. The changes are much more pronounced under oxidizing conditions, hence we conclude that the maximum applicable temperature is around 500 ◦ C. Above that temperature the conductivity of the tip drops significantly. Also the roughness of the coating increases six-

fold, compared with the untreated tip leading to an increase of the tip apex. In the case of reducing atmospheres, heat treatment is not so crucial, so that it is fairly safe to use temperatures above 500 ◦ C. However, it must be kept in mind that in our experiments the annealing time was only set to 0.5 h. In the case of normal measurements, the time during which the tip stays at an elevated temperature may be much longer thus leading to greater changes in both roughness and conductivity.

Acknowledgments One of the authors (M.W.) is grateful for financial support from the Forschungszentrum Jülich. We are also very grateful to J. Friedrich for SEM and TEM measurements and discussions.

7632

M. Wojtyniak et al. / Applied Surface Science 257 (2011) 7627–7632

References [1] K. Szot, W. Speier, G. Bihlmayer, R. Waser, Nature Materials 5 (2006) 312–320. [2] A. Gruverman, O. Kolosov, J. Hatano, K. Takahashi, H. Tokumoto, Journal of Vacuum Science & Technology B: Microelectronics and Nanometer Structures 13 (3) (1995) 1095. ´ [3] M. Nonnenmacher, M.P. OBoyle, H.K. Wickramasinghe, Applied Physics Letters 58 (25) (1991) 2921–2923. [4] Y. Martin, D. Abraham, H. Wickramasinghe, Applied Physics Letters 52 (13) (2009) 1103–1105. [5] M. Palacio, B. Bhushan, Nanotechnology 19 (10) (2008) 105705. [6] B. Bhushan, K.J. Kwak, Nanotechnology 18 (34) (2007) 345504. [7] B. Bhushan, M. Palacio, K. Kwak, Acta Materialia 56 (16) (2008) 4233–4241. [8] G.M. Sacha, M. Cardellach, J.J. Segura, J. Moser, A. Bachtold, J. Fraxedas, A. Verdaguer, Nanotechnology 20 (28) (2009) 285704.

[9] S. Doniach, M. Sunjic, Journal of Physics C: Solid State Physics 3 (1970) 285–291. [10] A. Hiraki, Surface Science Reports 3 (7) (1983) 357–412. [11] P.J. Grunthaner, F.J. Grunthaner, A. Madhukar, Journal of Vacuum 20 (1982) 680. [12] L. Kövér, J. Tóth, L. Dávid, Vacuum 37 (1–2) (1987) 125–127. [13] R.M. Tiggelaar, R.G.P. Sanders, A.W. Groenland, J.G.E. Gardeniers, Sensors and Actuators A: Physical 152 (1) (2009) 39–47. [14] J. Yin, W. Cai, Y. Zheng, L. Zhao, Surface and Coatings Technology 198 (1–3) (2005) 329–334. [15] J. Lee, B. Kim, Materials Science and Engineering: A 449–451 (2007) 769–773. [16] W. Rothschild, H. Yao, H. Plummer Jr., Langmuir 2 (5) (1986) 588–593. [17] M. Chen, L. Schmidt, Journal of Catalysis 55 (3) (1978) 348–360. [18] D.B. Poker, C.E. Klabunde, Physical Review B 26 (12) (1982) 7012– 7014.