Atomic force microscope-assisted scanning tunneling

0 downloads 0 Views 442KB Size Report
21 Aug 2014 - (FEMTO-DLPCA-200, Berlin, Germany) with a gain of. 109 V/A. An external lock-in amplifier (Signal Recovery. 5210; AMETEK, TN, USA) was ...
Microscopy Advance Access published August 21, 2014

Microscopy, 2014, 1–5 doi: 10.1093/jmicro/dfu028

Technical Report

Atomic force microscope-assisted scanning tunneling spectroscopy under ambient conditions Amin Vakhshouri1,*, Katsushi Hashimoto1,2, and Yoshiro Hirayama1,2,3 1

*To whom correspondence should be addressed. E-mail: [email protected] Received 8 February 2014; Accepted 17 July 2014

Abstract We have developed a method of atomic force microscopy (AFM)-assisted scanning tunneling spectroscopy (STS) under ambient conditions. An AFM function is used for rapid access to a selected position prior to performing STS. The AFM feedback is further used to suppress vertical thermal drift of the tip–sample distance during spectroscopy, enabling flexible and stable spectroscopy measurements at room temperature. Key words: combined atomic force/tunneling microscopy, room temperature tunneling spectroscopy, thermal drift suppression

Introduction Scanning tunneling microscopy (STM) [1] is a powerful method for probing electronic structures at atomic resolution. The probed structures provide information on the surface topography and, more significantly, the local density of states (LDOS), which can be further measured as a function of the energy [2]. The latter method, i.e. scanning tunneling spectroscopy (STS), has been utilized for revealing specific confined band structures in local atomic structures, such as carbon nanotubes and artificial atomic chains, deposited on atomically flat conducting substrates [3]. STM is limited to the conducting substrates. In contrast, atomic force microscopy (AFM) [4] can be done on any surface, including insulators. Moreover, AFM is capable of sensing long-range (several nanometers) tip–sample interaction forces, i.e., van der waals forces. Yet, the tip–sample distance in STM is limited to the decay length of tunneling

current, which is of the order of a few angstroms. As a result, by performing AFM in non-contact mode at a large tip–sample distance of the nanometer scale [5], which is one order of magnitude larger than for STM [6], one can image a non-atomically flat surface without tip crash faster than when using STM. Thus, by combining AFM and STM, one can perform rapid positioning by AFM prior to STS measurements. This method is applicable, in particular, to STS studies of electrically wired conductive objects, which are inhomogeneously distributed across surfaces with uneven topography on insulating substrates. Good examples are devices made from layered material, such as graphene and MoS2, deposited on an insulating substrate (typically SiO2 film) [7,8] by micromechanical exfoliation [9] and electrically connected by post-deposited ohmic electrodes. Because the typical lateral size of a target single layer is a few micrometers [10], it is difficult to accurately position the tip onto

© The Author 2014. Published by Oxford University Press on behalf of The Japanese Society of Microscopy. All rights reserved. For permissions, please e-mail: [email protected]

1

Downloaded from http://jmicro.oxfordjournals.org/ by guest on August 24, 2014

Department of Physics, Graduate School of Science, Tohoku University, Sendai, Miyagi 980-8578, Japan, 2JST, ERATO Nuclear Spin Electronics Project, Sendai 980-8578, Japan, and 3WPI-AIMR, Tohoku University, Sendai 980-8577, Japan

Microscopy, 2014, Vol. 0, No. 0

2

Experimental Figure 1 shows a schematic model of our microscope. We used Asylum Research AFM (MFP-3D, CA, USA) with a standard chip holder. In order to utilize it for STM, we used a metallic Pt/Cr-coated cantilever and connected it to an external operational-amplifier current-to-voltage convertor (FEMTO-DLPCA-200, Berlin, Germany) with a gain of 109 V/A. An external lock-in amplifier (Signal Recovery 5210; AMETEK, TN, USA) was used for the STS measurement. To switch the z-FBL input between deflection (q) and the tunneling current signal (the upper right block in Fig. 1), we modified the source code of the software (Asylum Research-MFP3D12084). The AFM images were taken in amplitude-modulation (AM) attractive mode, confirmed by a phase larger than 90° during imaging. To avoid a tip crash, we carefully chose a ratio of the free air (A0) and set point (As) oscillation amplitudes As/A0 (0.9) close to

Fig. 1. Schematic representation of combined AFM/STM. External operational amplifier (ext. op amp) and lock-in amplifier added for detection of It and STS signal. In the case of AFM (STM), the z-feedback loop input is the deflection (tunneling current) signal.

1. Several cantilevers were tested for STM measurements. Cantilevers with spring constant of K ∼50 N/m (Budget Sensors Tap300-G; TED PELLA, CA, USA) were found to be stable enough for STM measurement. The spring constant of the cantilevers used in this work were determined by the thermal noise method [11] after each experiment. A lower-K cantilever caused a tip crash that occurred right after the tip started moving. This might be due to mechanical instability of the cantilever [6]. All the measurements were done on highly ordered pyrolytic graphite (HOPG) (SPI Supplies, PA, USA) under ambient conditions.

Results and discussion Figure 2a shows an AFM image taken in the relatively large area of 10 × 10 µm2, which includes a maximum surface corrugation of 11 nm (bright strand passing from upper to lower). This image was recorded at the scanning speed of 15 µm/s for a total time of 5.6 min. If STM was used to scan the same area, the scanning speed (total scanning time) should have been reduced (increased) at least three orders of magnitude to not get a tip crash. After scanning the large area, we selected a smaller area of 0.5 × 0.5 µm2 (square in Fig. 2a) and performed STM imaging (Fig. 2b). Next, we examined the VTD, which may affect the standard spectroscopy measurement performed with disabling FBL. Figure 3a shows time dependence of z-piezo displacement measured by a sensor (linear variable differential transformer (LVDT) sensor, which accurately measures the z-piezo displacement [12]) with enabling AFM FBL. The positive slope of the graph indicates that the FBL retracted the tip from the sample surface as a result of FBL compensation for the reduced tip–sample distance (Zg) due to the thermal drift. The drift rate estimated using linear fitting was v = 1.1 Å/s. These results imply that the time required for a single spectroscopy is limited to below 1 s; otherwise,

Downloaded from http://jmicro.oxfordjournals.org/ by guest on August 24, 2014

a layer located in between thicker layers by using only an optical microscope. AFM offers further advantages when used with STS measurements, especially at room temperature. Each LDOS spectroscopy is done by recording the differential conductivity dIt/dVs with sweeping the sample voltage Vs, which determines the sample energy, where It is the tunneling current. During spectroscopy, STM feedback loop (FBL) is disabled to maintain a constant tip–sample distance that otherwise would change with sweeping Vs [2]. Without the FBL, the vertical thermal drift (VTD) is not compensated for and therefore affects the intrinsic LDOS curve. This problem limits the total Vs sweeping time determined by the sweeping range, the number of the data acquisition and the averaging time for each acquisition, which, respectively, determine the energy range, energy resolution and signal-to-noise ratio (S/N) of the data. Using AFM FBL, one can compensate for the thermal drift during spectroscopy and therefore increase the total Vs sweeping time. This enables us to perform room temperature (RT) measurements of LDOS over a wide energy range at a high energy resolution and high S/N. In this article, using our combined AFM–STM, we show alternating AFM and STM measurements using a metallic cantilever under ambient conditions. We demonstrate AFM-assisted rapid access to a selected area prior to STM measurements. Moreover, we present tunneling spectroscopy with activated AFM FBL that suppresses the effect of VTD on the It with the help of repulsive force due to a thin layer of insulating contamination formed on the tip and/or sample. Our methodology best shows its merits when used for STS on the layered material devises under ambient conditions where the formation of contamination is unavoidable.

Microscopy, 2014, Vol. 0, No. 0

Fig. 3. (a) Measurement of VTD. This measurement was done in contact AFM mode using a cantilever with K = 3.7 N/m. The VTD rate of υ = 1.1 Å/s was estimated by fitting a line to the graph. (b) Simultaneous measurement of It and F as a function of Zp. In this measurement, grounded tip approached to the biased sample (Vs = 0.5 V) until It = 500 pA was detected and then tip retracted. This procedure repeated on the same position and It and F averaged over three repeated measurements. The origin point of Zp = 0 nm is arbitrary chosen to be at the point where It starts to flow. The inset shows log scale of It (Zp) curve indicating exponential behavior of the current. (c and d) Switching STM FBL into AFM FBL to compensate for VTD. The stability of q and It before and after FBL switching indicates stable Zp while compensating for thermal drift. The set point of tunneling current and sample bias are Is = 50 pA and Vs = 0.5 V. The spring constant is K = 44.76 N/m for (b) and is K = 52 N/m for (c) and (d).

Downloaded from http://jmicro.oxfordjournals.org/ by guest on August 24, 2014

Fig. 2. Alternating AFM and STM imaging on selected area of an HOPG. (a) Fast and large area scanning by AFM. (b) STM image of the selected area marked by the white square in the center of (a). The solid line is the profile along the same line as the dashed line. The cantilever spring constant is K = 52 N/m. Sample voltage and tunneling current set point are Vs = 1.0 V and Is = 50 pA.

3

Microscopy, 2014, Vol. 0, No. 0

4

Fig. 4. Tunneling spectroscopy with long averaging time and the effect of electrostatic force on cantilever deflection. The curve shows the long measurement of dIt/dVs. The total measurement time was 30.1 s. The AFM FBL was activated during spectroscopy to compensate for the VTD. High S/N of dIt/dVs signal indicates a stable Zg. The inset illustrates bias voltage dependence of the deflection. The resulting maximum change in tunneling gap is ≈8.6 pm at Vs = 0.5 V, which leads to 2% change of It. The set point of It and Vs are Is = 20 pA and Vs = 0.5 V. The spring constant is K = 44.76 N/m.

spectroscopy while flexibly tuning the energy range, energy resolution and S/N. It is known that the applied Vs induces the electrostatic force. Sweeping the Vs indeed varies the deflection (Δq) (inset of Fig. 4), which has a minimum near Vs = −0.1 V due to a mismatch of tip and surface potential [21]. Within the range used for the tunneling spectroscopy −0.5 V < Vs < 0.5 V, the maximum Δq is found to be 16 pm. To estimate the change in the tip–sample gap (ΔZg), we used a previously proposed model [22]. According to this model, ΔZg = ΔZp − Δq − ΔZs. Here, ΔZs is the deformation of the sample which is negligible because the sample is relatively rigid [22] compared with the cantilever and the gap (insulating layer) in our system. By roughly estimating the force gradient of ∼−29 N/m between points 1 and 2 (Fig. 3b), where ΔZg is nonzero, and dividing it by cantilever force constant (K = 44.76 N/m), we obtain Δq = 0.65ΔZp. Substitution of the latter in the first equation yields ΔZg = 0.54Δq. Therefore, a change in tip–sample gap ΔZg ≈ 8.6 pm may cause current change of ΔIt ≈ 0.4 pA. This value is only 2% of the tunneling current set point Is = 20 pA, implying that the effect of electrostatic force on dIt/dVs curves is negligibly small.

Conclusion We succeeded in performing alternating AFM and STM measurements on an HOPG surface using a Pt/Cr-coated cantilever under ambient conditions. This usage of

Downloaded from http://jmicro.oxfordjournals.org/ by guest on August 24, 2014

the thermal drift can cause a substantial artifact in the measured LDOS curve and even a tip crash. This limitation is not really a problem at low temperature, because the thermal drift is negligible. However, at RT, as we showed, a much larger VTD limits the measurement time and hence limits the energy range, the energy resolution and the S/N of the LDOS data. To overcome this common difficulty of the RT measurement, we used the AFM FBL to keep the Zg constant during the tunneling spectroscopy. The AFM FBL source that we used is a static deflection signal instead of an AM signal excited by cantilever vibration, which may affect It. Figure 3b shows typical It and force (F) curves as a function of z_piezo displacement (Zp) measured simultaneously without feedback. The force between points 2 and 3 is in repulsive regime ( positive q) with a gradient of ΔF/ΔZp = −44 N/m, which we utilized for the AFM FBL during STS measurements. The It–Zp curve shows an exponential behavior (inset of Fig. 3b), i.e. signature of the tunneling current. The repulsive forces at which we performed STM and STS measurements indicate that the tip and the sample were mechanically coupled. The same tendency was also observed in STM measurements performed under ambient conditions [13–17] especially prior to electrical contact (i.e. in the tunneling current regime) [18]. This behavior interpreted as tunneling through an insulating layer caused by unavoidable contamination on a tip and/or sample surface [19]. The AFM-assisted tunneling spectroscopy was performed as follows. First, the tip was stabilized at a fixed Zp in STM mode. Second, the deflection signal was acquired to determine the deflection value at the corresponding Zp. Finally, before starting tunneling spectroscopy, STM FBL was switched to AFM FBL, which used the determined deflection value as the FBL reference to keep the Zg constant. To investigate the stability of mechanical vibration and It before/after switching, we measured the It (Fig. 3c) and q (Fig. 3d) versus time. We found that switching the FBL does not change the deflection and hence the tunneling current remains the same. Moreover, after switching the FBL, the tunneling current was maintained for >70 s, indicating that AFM FBL compensated for the VTD. Finally, we conducted AFM-assisted single tunneling spectroscopy at Zp = −3.1 nm and It = 20 pA (marked by the vertical dashed line in Fig. 3b), using a lock-in amplifier at a total Vs sweeping time T = 30.1 s determined by the number of the data acquisition, n = 101, and the averaging time for each acquisition, τ = 300 ms. The obtained dIt/dVs curve (Fig. 4) shows the typical ‘V-shaped’ characteristic of HOPG in agreement with a previous result [20]. Moreover, the observed smooth curve implies a fairly high S/N and no tip crash in the long spectroscopy time. This indicates that AFM-assisted tunneling spectroscopy enables us to perform

Microscopy, 2014, Vol. 0, No. 0

combined microscopy enables us to perform AFM-assisted tunneling spectroscopy. In this method, we select a spot/ area of interest on wide and non-atomically flat samples by AFM and compensate for VTD by using AFM feedback for stable tunneling spectroscopy measurements at RT. The AFM-assisted STS is appropriate for the study of local electronic structures on layered material devices, e.g. for determining layer-dependent band gaps of MoS2.

References

11. Hutter J L, Bechhoefer J (1993) Calibration of atomic-force microscope tips. Rev. Sci. Instrum 64: 1868–1873. 12. Proksch R, Cleveland J, Bocek D (2006) Linear variable differential transformers for high precision position measurements. Google patents: US 20060186876A1. 13. Mate C M, Erlandsson R, McClelland G M, Chiang S (1989) Direct measurement of forces during scanning tunneling microscope imaging of graphite. Surf Sci 208: 473–486. 14. Sugawara Y, Ishizaka T, Morita S (1990) Origin of anomalous corrugation height of STM images of graphite. Jpn J. Appl. Phys. 29: 1533–1538. 15. Sugawara Y, Ishizaka T, Morita S (1991) Scanning force tunneling microscopy of a graphite surface in air. J. Vac. Sci. Technol. 9: 1092–1095. 16. Meepagala S C, Real F, Reyes C B (1991) Tip–sample interaction forces in scanning tunneling microscopy—effects of contaminants. J. Vac. Sci. Technol. 9: 1340–1342. 17. Suganuma Y, Dhirani A (2003) A hybrid scanning tunnelingatomic force microscope operable in air. Rev. Sci. Instrum. 74: 4373–4377. 18. de Pablo P J, Martinez M T, Colchero J, Gomez-Herrero J, Maser W K, Benito A M, Munoz E, Baro A M (2000) Mechanical and electrical properties of nanosized contacts on singlewalled carbon nanotubes. Adv Mater 12: 573–576. 19. Mamin H J, Ganz E, Abraham D W, Thomson R E, Clarke J (1986) Contamination-mediated deformation of graphite by the scanning tunneling microscope. Phys Rev B 34: 9015–9018. 20. Niimi Y, Matsui T, Kambara H, Tagami K, Tsukada M, Fukuyama H (2006) Scanning tunneling microscopy and spectroscopy of the electronic local density of states of graphite surfaces near monoatomic step edges. Phys. Rev. B 73: 8. 21. Liscio A, Palermo V, Muellen K, Samori P (2008) Tip–sample interactions in Kelvin probe force microscopy: quantitative measurement of the local surface potential. J. Phys. Chem. C 112: 17368–17377. 22. Salmeron M, Ogletree D F, Ocal C, Wang H C, Neubauer G, Kolbe W, Meyers G (1991) Tip-surface forces during imaging by scanning tunneling microscopy. J. Vac. Sci. Technol. 9: 1347–1352.

Downloaded from http://jmicro.oxfordjournals.org/ by guest on August 24, 2014

1. Binnig G, Rohrer H (1982) Scanning tunneling microscopy. Helvetica Physica Acta 55: 726–735. 2. Stroscio J A, Kaiser W J (1993) Scanning Tunneling Microscopy (Academic Press). 3. Wildoer J W G, Venema L C, Rinzler A G, Smalley R E, Dekker C (1998) Electronic structure of atomically resolved carbon nanotubes. Nature 391: 59–62. 4. Binnig G, Quate C F, Gerber C (1986) Atomic force microscope. Phys. Rev. Lett. 56: 930–933. 5. Martin Y, Williams C, Wickramasinghe H (1987) Atomic force microscope force mapping and profiling on a sub 100-A scale. J. Appl. Phys. 61: 4723–4729. 6. Morita S, Wiesendanger R, Meyer E (2002) Noncontact Atomic Force Microscopy (Springer). 7. Novoselov K S, Geim A K, Morozov S V, Jiang D, Katsnelson M I, Grigorieva I V, Dubonos S V, Firsov A A (2005) Two-dimensional gas of massless Dirac fermions in graphene. Nature 438: 197–200. 8. Radisavljevic B, Radenovic A, Brivio J, Giacometti V, Kis A (2011) Single-layer MoS2 transistors. Nature Nanotechnology 6: 147–150. 9. Novoselov K S, Geim A K, Morozov S V, Jiang D, Zhang Y, Dubonos S V, Grigorieva I V, Firsov A A (2004) Electric field effect in atomically thin carbon films. Science 306: 666–669. 10. Mak K, Lee C, Hone J, Shan J, Heinz T (2010) Atomically thin MoS2: a new direct-gap semiconductor. Phys. Rev. Lett. 105: 136805.

5