Tip-enhanced Electron Emission Microscopy Coupled with the Femtosecond Laser Pulse

Abstract

The ultrashort electron pulse, laser-emitted from the metal tip apex has been characterized and used as a probing source for a new electron microscope to visualize the morphology of the gold-mesh in the nanometric resolution. As the gap between the tungsten tip and Au-surface is approached within a few nm, the large electromagnetic field enhancement for the incident P-polarized laser pulse with respect to the tip-sample axis is strongly observed. Here, we demonstrate that the time-resolved tip-enhanced electron emission microscope (TEEM) can be implemented on the laboratory table top to give the two-dimensional image, opening lots of challenges and opportunities in the near future.

Introduction

Nano science and technology1 have been blossomed with the advent of various electron microscopes with which one can nowadays get the spatial resolution of ∼1Å.2,3 Materials of specially-designed nanostructures are vastly produced these days, and their optical, physical, or functional properties in the sub-nanometer spatial resolution are strongly demanded. Compared to the optical microscopes of which the spatial resolution is often diffraction-limited, the electron microscope provides a much better opportunity to investigate the material at the molecular level

Here, we report an electron microscope with which, in principle, one can visualize the nano-scale morphology of the sample with femtosecond time resolution. The timeresolved electron microscope has been realized by the Zewail group4,5 through the combination of the conventional transmission electron microscope (TEM) with the femtosecond laser pulse. In the setup, they utilize the photoelectric effect induced by the femtosecond laser pulse to generate the ultrashort electron pulse, which is then further accelerated and focused to give the highly-resolved TEM image as a function of the reaction time. This 4D microscopy has been enormously successful, providing unprecedented information about the ultrafast structural changes occurring on the surface. The technique, however, is costly, and the wider application is rather limited

A quite different approach for attaining the same goal can be implemented on the laboratory table-top. In this approach, the tungsten tip is irradiated by the femtosecond laser pulse. Due to the lightening rod effect,6-11 the laserinduced field enhancement is peaked at the apex of the tip, resulting in the emission of electron in the tip-pointing direction. Many research groups have recently investigated the detailed mechanism of photoemission from the nanosized metallic tip quite intensively, and the photo-induced electron emission from the metal tip seems to be well characterized.12-17 When the tip is very near to the sample being probed, the dispersion of the emitted electron will be confined in nanometer scaled space whereas the time duration of the electron pulse remains within a few tens of femtosecond. If one uses this ultrashort and yet spatially confined electron pulse, it may be possible to investigate the spatiotemporal property of surface material, Figure 1. The first demonstration of the electron microscope based on this novel concept actually had been reported by Ropers et al.17 to give the local electromagnetic field distribution as the illuminated tungsten tip is linearly scanned over the nanometrically grooved gold structure, showing the great promise of this technique as a time-resolved electron microscope. Since this pioneering work was reported, however, the development of this tool as a practical electron microscope has rather slowly progressed

Here, we report the image of a gold mesh, providing the first realistic TEEM image which reflects the two-dimensional morphology of the surface structure. The effect of the bias voltage, laser intensity, tip-sample distance, and laser polarization have been systematically investigated as an endeavour toward the realization and wide application of this new technique as a table-top time-resolved electron microscopic tool, eventually for the study of the molecular level mechanism of photo-catalysis, solar cell, or photochemical energy conversion.

 

Experimental

Materials. Tungsten wires (300 um-thick, 99.95%, Nilaco Corporation) Potassium hydroxide (85%, JUNSEI), Gold mesh (MG-47, Precision Eforming), Carbon Tape (Okenshoji), Tuning fork (32.768 kHz, Microquartz Electronics).

Electrochemical Etching. A tungsten tip was manu-factured by an electrochemical etching technique using a tungsten rod with 300 μm-thickness in 3 N KOH solution. A tungsten rod was etched by two steps. In the first step, we applied 10 V of constant voltage to process the electrochemical reaction. When the cuurent was decreased to about 0.09 Ampere, the first step is finished. In the second the voltage was 3 V until the electrochemical etching was completed. Cut-off time was determined by a homemade program using LabView software (National Instruments, LabView2009). Produced tungsten tips are dipped into 48% HF and rinsed in de-ionized water to remove the layer of electrolyte. In our electrochemical etching setup, tungsten tips are reproduced with a curvature of tens of nanometer.

Shear Force Microscopy. An electrochemically etched tungsten tip was mounted on the side of a tuning fork using instant glue and sine wave produced by a function generator (DS345, Stanford Research Systems) is supplied to a tuning fork and a pre-amplifier. At the resonance frequency, the tuning fork is mechanically excited. Then, the amplitude at the resonance frequency was amplified by home-made electronic circuit and detected by lock-in amplifier (SR810, Stanford Research Systems).

Tip-enhanced Electron Emission Microscopy (TEEM). Figure 1 shows the TEEM experimental setup schematically. The solid-state diode-pumped Nd:YVO4 laser (Verdi V-5, Coherent, 532 nm single-frequency output) was used to pump the cavity dumped Ti:sapphire oscillator (KM Lab.). Bragg cell was used for cavity dumping by acousto-optic modulation from a RF driver (NEOS). We tuned laser pulses so that the output power was around 24 mW at a central wavelength of 810 nm and 15 fs pulse width at 800 kHz repetition rate. A prism pair compensates the group delay dispersion (GDD) of the output pulse. This pulse is separated by two. One acts as pump pulse and the other is served as probe pulse. These pulses are focused on the tungsten tip using 75 mm-focal length planoconvex lens which can travel 25.4 mm-forward and backward using x-axis picomotor stage. And electron pulse is generated by a multi-photon ionization process. Because of the phase interference of two laser pulses, we have observed the strong temporal modulation in the electron signal of which the Fourier-transformed frequency corresponds to the central wavelength of 810 nm. This pulse interacts with surface. After then, these electrons are detected by a home-made microchannel plate detector and integrated by a photon-counter (SR400, Stanford Research Systems). All data acquisition was obtained by home-made program using LabView software.

Figure 1.Experimental setup of the tip-enhanced electron emission microscope. The femtosecond laser pulses with the repetition rate of 800 kHz are focused on the tip for the emission of the electron. The electron signal is temporally modulated by the interferometric autocorrelation of two time-delayed coherent femtosecond laser pulses, giving the corresponding pulse width of ∼25 fs, as shown in the inet. (PLCX: plano convex lens, WP: λ/2 waveplate, BS: beamsplitter).

 

Results and Discussion

Figure 2.Gold mesh images taken by (a) scanning electron microscope and (b) shear force microscope (SFM). The tipenhanced electron emission microscope (TEEM) images taken with the P-polarized femtosecond laser pulse at the tungsten tip bias voltages of (c) 0, (d) −50, (e), −100, (f) −200, (g) −300, and (h) −500 V are shown.

The resulting two dimensional TEEM images, taken at the same position after shear force microscopy, is shown in Figure 2(c)-(h). At zero bias voltage, the electron signal is enhanced at the hole whereas it gives the much less electron signal when the tip is at the gold surface. This experimental observation may indicate that the electron emitted along the tip-pointing direction is reflected when the tip is positioned at holes as the base-plate for the gold mesh is made of quartz, whereas it is absorbed when the tip is at the gold surface. The TEEM image taken at the zero bias voltage, however, does not reflect the realistic shape of the gold-mesh. The size of the hole is much smaller than that of the real one, and the electron signal from the upper-half part of the image is much stronger than that from the lower-half part. In the lower-half part, only the left hole gives the weak electron signal. This asymmetry should come from the tilt of the sample with respect to the plane perpendicular to the tip-pointing direction. Interestingly, the TEEM image becomes inverted when the small negative bias voltage is applied to the tungsten tip, Figure 2(d)-(h). Since the sample is electrically floated, the application of this bias voltage to the tip gives the poorly defined electric field. Nonetheless, the effect of the negative bias voltage is found to be dramatic. The electron signal is much enhanced when the negatively biased tip is positioned at the gold surface, whereas it remains more or less constant when the tungsten tip is positioned at the hole of the gold mesh. The TEEM images taken at the bias voltages between ‒50 and ‒200 V are very similar to the SFM image, indicating that the TEEM technique can be employed as a practical high-resolution electron microscope. As the bias voltage is further increased, it is found that the TEEM images become blurred.

It should be noted that no electron is detected when the femtosecond laser pulse is irradiated solely on the goldsurface. Secondly, the gap between the tip and sample should be very tiny to observe the significant enhancement of the electron signal, as previously reported in the study by Ropers et al..17 Thirdly, no electron is detected without the laser irradiation. Therefore, the laser pulse, the tungsten tip, and gold sample should be at right positions in order to get the electron signal enhancement. Another requirement is the polarization of the laser pulse. The P-polarization is essential for obtaining the TEEM image. Both the lightning rod (LR) and surface plasmon polariton (SPP)18-24 effects seem to be responsible for these phenomena. The LR effect comes from the tip apex. The laser irradiation on the tungsten tip induces the free-electron oscillation along the tip surface and its amplitude becomes maximized at the apex, so that the electron emission gets efficient only at the tip apex. The application of the bias voltage to the tungsten tip then lowers the tunneling barrier of the ionization from tungsten tip. Another factor mainly contributing to the TEEM image of the goldmesh should come from the SPP effect.25 In this case, the electron on the gold surface oscillates with the irradiated laser pulse, and this oscillation induces the local electromagnetic field enhancement26-28 at the small gap between the tip and sample.

Figure 3.FDTD calculations of the electric field distribution for (a) a single tungsten tip and (b) a tungsten tip on the gold surface. The distance between the tungsten tip and the gold surface is set to be 2 nm. The polarization E and wave vector k of the propagating light are illustrated as arrows. One can notice that the electric field intensity has been increased by ∼30 fold as the Tungsten tip approaches to the gold surface

Calculations are carried out using the finite-difference in time-domain (FDTD)29-32 technique for the current experimental condition. The curvature radius of the tungsten tip and its cone angle are assumed to be 50 nm and 35°, respectively. The gap between the tungsten tip and the gold surface is fixed at 2 nm. In the simulation, the Yee cell size is set to be 2 × 2 × 2 nm3 in all regions. The broadband P-polarized light with the central wavelength of 810 nm is shone in the simulation at the incidence angle of 10° with respect to the plane of the gold substrate. Calculated electric field distributions generated around the tungsten tip with and without the gold substrate nearby are shown in Figure 3(a) and (b), respectively. The electric field intensity is found to be approximately 30 times enhanced as the tungsten tip is in close proximity to the gold substrate, confirming the essential role of SPP in the electron signal enhancement in the present work. The application of the negative bias voltage to the tungsten tip automatically induces the strong local electric field enhancement at the gap between the tungsten tip and sample, giving the synergistic effect to the SPP excitation. Therefore, the electron signal is strongly enhanced when the tip is close to the gold surface whereas it is only weakly detected when the tip is distant from the gold surface at the hole of the gold mesh, giving the realistic TEEM image. The TEEM images in Figure 2 show somewhat inhomogeneous intensity distribution especially along the χ-coordinate.

This inhomogeneity is most likely due to the tilt of the sample, resulting in the change of the tip-sample distance along the z-direction, which is consistent with the TEEM image taken at the zero bias voltage (vide supra). This suggests that the TEEM image could be much more sensitive than the SFM image along the z-direction, giving the plausibility of TEEM as a tool for the three-dimensional mapping of the surface morphology. For testing the possibility of the time-resolved image recording, the femtosecond laser pulse is divided into two parts using a beam-splitter, and the resulting pump and probe laser pulses are optically delayed to give the surface image as a function of the delay time.

Actually, because of the phase interference of two laser pulses, we have observed the strong temporal modulation in the electron signal of which the Fourier-transformed frequency corresponds to the central wavelength of 810 nm. This reflects that the actual pulse width of the emitted electron is also in the femtosecond time scale, giving the bright future of TEEM as a time-resolved nano-probing tool. This also verifies again that the TEEM image reported in this work is produced by the ultrashort electron pulse generated by the femtosecond laser pulse irradiation. It should be noted that our first two-dimensional TEEM image is rather preliminary in terms of quantitative measures, and there is much room for this TEEM technique to be much improved and refined as a microscopic tool in terms of both spatial and temporal resolutions.

 

Conclusion

In conclusion, we have demonstrated that the tip-enhanced electron emission microscope can be implemented on the laboratory table top to give the two-dimensional image of the surface structure. The ultrashort time duration of the electron pulse from the tip, which is being used as a nanoprobe in TEEM, should be ideal for the real time-resolved nano electron microscope. The lightning rod effect induced by the tip apex with the aid of the surface plasmon polariton effect maximized at the tiny tip-sample gap gives the clear picture of the surface structure of the gold-mesh. The image should depend on the surface plasmon resonant property of the material, and thus TEEM may also be quite useful to study the structure of alloy materials with different dielectric constants. Moreover, since the time duration of the electron pulse is in the femtosecond time scale, it should be plausible to see the nanometer confined structural change of the surface including the morphology, electric property, or vibrational structure in real time.