291

# Ultrafast Optical Pump-Probe Scanning Probe Microscopy/Spectroscopy

Authored by: Hidemi Shigekawa , Shoji Yoshida

# 21 Century Nanoscience – A Handbook

Print publication date:  April  2020
Online publication date:  April  2020

Print ISBN: 9780815384731
eBook ISBN: 9780429340420

10.1201/9780429340420-3

#### Abstract

With size reduction, the differences in the electronic properties of materials and current devices, for example, those caused by the structural nonuniformity in each element, have an ever-increasing effect on macroscopic functions. For further advances in nanoscale science and technology, the development of a method for exploring the transient dynamics of local quantum functions in organized small structures is essential. Here, we overview laser-combined scanning tunneling microscopy/spectroscopy (STM/STS) based on the optical pump-probe method, which enables the real-space imaging of nanoscale quantum dynamics with atomic resolution even in the femtosecond range.

#### 3.1  Introduction

The understanding and control of quantum dynamics, such as the transition and transport in nanoscale structures, are the key factors for continuing the advances in nanoscale science and technology. However, with size reduction, the differences in the electronic properties of materials and current devices, for example, those caused by the structural nonuniformity in each element, have an ever-increasing effect on macroscopic functions. For example, atomic-scale defects have markedly changed the entire situation: defects, which were once considered as a problem to be avoided, are now actively designed and controlled to realize desired functions. The fluctuation in the distribution of dopant materials governs the characteristic properties of macroscopic functions. Therefore, for further advances in science and technology, the development of a method for exploring the transient dynamics of local quantum functions in organized small structures is essential.

The spatial resolution of scanning tunneling microscopy (STM) is excellent. We can analyze local structures and electronic properties such as the local density of states with atomic resolution. Since the invention of STM, the direct imaging of atomic-scale structures has been lifting the veil from various long-standing problems and extending the frontiers of science and technology (Binning et al. 1982; Bhushan 2010; Wisendanger 1994). In STM, a sharp tip is placed above the target material, and information immediately below the probe tip is obtained through measurement of the tunnel current, spin, force, and so forth (Figure 3.1). Since the tunnel current logarithmically depends on the tipsample distance, a 0.1 nm change in the distance produces a one-order change in the current. Therefore, if the STM tip is scanned over the sample surface while the tunnel current is kept constant using piezoelectric elements, three-dimensional imaging of the sample surface can be realized with atomic resolution.

Figure 3.1   Schematic illustration of STM.

In basic STM, the bias voltage applied between the STM tip and the sample is adjusted and the corresponding change in tunneling current is measured, giving information about the local electronic structures with atomic resolution. With the modulation of additional parameters, such as temperature, magnetic field, and tip-sample distance depending on the purpose, further information can be obtained. However, since the temporal resolution of STM is limited, in general, to less than 100 kHz owing to the circuit bandwidth (Mamin et al. 1994; Wintterlin et al. 1997; Kemiktarak et al. 2007), the ultrafast dynamics in materials has been beyond its field of vision. In contrast, the advances in ultrashort-pulse laser technology have opened the door to the world of ultrafast phenomena. A prominent method is optical pump-probe (OPP) measurement, which has enabled ultrafast dynamics to be observed in the femtosecond range. However, the spatial resolution of such optical methods is generally limited by the wavelength, which may be averaged over the light spot area. Therefore, since the invention of STM in 1982, one of the most challenging goals has been to combine STM with ultrashortpulse laser technology to simultaneously realize ultimate spatial and temporal resolutions. This issue has a long history, and many attempts have been made to achieve this goal using various approaches (Terada et al. 2010a; Shigekawa et al. 2010a).

In the following section, we focus on the development of laser-combined STM based on the OPP method.

#### 3.2.1  Optical Pump-Probe Method and Optical Pump-Probe STM

In the OPP method, a sample is illuminated by a train of pulse pairs with a certain delay time. First, pulses are used as a pump to excite the sample, and second, pulses are used as a probe to observe the relaxation of the excited state (Figure 3.2a). When carriers excited by the pump pulse remain in the excited states, the absorption of the probe pulse is suppressed, that is, absorption bleaching occurs. Thus, if the reflectivity of the second pulse, for example, is measured as a function of delay time, we can obtain information on the relaxation of the excited state induced by the pump pulse through the change in the reflectivity of the probe pulse (Figure 3.2b). In this case, the time resolution is only limited by the pulse width, namely, to the femtosecond range.

Figure 3.2   (a) Schematic illustration of OPP technique. (b) Normalized change in reflectivity ΔR/R as a function of delay time td .

In the new microscopy technique of OPP-STM, the sample surface below the STM tip is excited by a paired-pulse train with a certain delay time td , similar to that in the OPP method, but the signal is the tunnel current I, instead of change in the reflectivity of the probe pulse, as a function of delay time (Figure 3.3a). Namely, we probe the change in the number of carriers excited by the second pulse due to absorption bleaching through the change in the tunnel current as a function of the delay time. The optical pulses give rise to current pulses in the raw tunneling current I *, which reflects the excitation and relaxation of the sample. If these current pulses decay rapidly compared with the timescale of the STM preamplifier bandwidth, they are temporally averaged in the preamplifier and cannot be detected directly in the signal I. Even in such a case, the relaxation dynamics can be probed through the td dependence of I (Shah 1999).

In OPP-STM measurement, we usually employ pump and probe pulses with equal intensity. When td is sufficiently long, the paired optical pulses with the same intensity independently induce two current pulses with the same height (I *) as shown by A in Figure 3.3b. In contrast, when td is short and the second pulse illuminates the sample in the excited state caused by the first pulse, the second current pulse may have a different height dependence on td (B and C in Figure 3.3b). Since the change in the total current I * induced by the first and second pulses causes a change in the average tunneling current I, as shown in Figure 3.3c, as a function of the delay time, we can elucidate the dynamics of the excited states through the change in tunneling current. In this case, we can obtain the temporal resolution of the OPP method together with the spatial resolution of STM.

#### 3.2.2  How to Measure a Weak Signal

Since signals are weak, we need to use the lock-in detection method in OPP-STM. The excitation is oscillated at a certain frequency and the corresponding change in the signal is measured. In general, laser intensity is modulated in the OPP method, which, however, causes thermal expansion of the STM tip and sample (Grafström 2002). Since a 0.1 nm change in the tip-sample distance produces a one-order change in the tunneling current, it is difficult to detect a weak signal under this condition. A promising option is to modulate the delay time between the pump and probe pulses instead of the intensity. Using this modulation technique, the laser intensity is not changed and a modulation frequency independent of the noise frequency originating from thermal expansion can be chosen (Takeuchi et al 2004b, 2006; Terada et al. 2010a, b).

To realize a microscopy technique that enables us to visualize carrier dynamics in a nanometer-scale potential landscape, we developed a method for rectangular modulation of the delay time by using a pulse-picking technique. As shown in Figure 3.4, pulse trains are generated by two synchronized Ti:sapphire lasers at a 90 MHz repetition rate (11 ns intervals) with a pulse width of 150 fs in this case. The relative timing of the two pulse trains is controlled by a synchronizing circuit, which provides a delay time that ranges from 0 to 11 ns with a time jitter. Each train is guided to a pulse picker that can selectively transmit 1 pulse per 90 pulses, resulting in the reduced repetition rate of 1 MHz. The pulse selection enables the production of an additional longer delay time that can be adjusted in multiples of 11 ns. Consequently, td can be adjusted continuously from zero to a large value as needed. This control of the delay allows nanometer-scale structures to be probed with a wide range of relaxation lifetimes.

Figure 3.3   (a) Schematic illustration of OPP-STM. (b) Measurement mechanism. (c) OPP-STM spectrum corresponding to the mechanism in (b).

Figure 3.4   Schematic illustration of OPP-STM system. Rectangular modulation of the delay time is shown in the lower right of the figure.

In the pulse-picking method, td can be modified by large and discrete amounts by changing the timing of selecting pulses that are transmitted by pulse pickers, making the method suitable for modulating td in OPP-STM. For the modulation of td between $t d 1$ and $t d 2$ in a rectangular waveform, the lock-in detection of the tunneling current gives the value $Δ I ( t d 1 , t d 2 ) = I ( t d 1 ) − ( t d 2 )$ as shown in the lower right of Figure 3.4. As $t d 2$ is set to a large value compared with the relaxation time of the probed dynamics, $Δ I ( t d 1 , t d 2 )$ can be approximated as $Δ I ( t d 1 ) = I ( t d 2 ) − I ( ∞ )$ , where I(∞) is the tunneling current for a delay time that is sufficiently long for the excited state to be relaxed. Therefore, $Δ I ( t d 1 )$ is accurately obtained through the lock-in detection of I by sweeping $t d 1$ . In addition, since the modulation can be performed at a high frequency (1 kHz in our case), the measurement is not significantly affected by low-frequency fluctuations in the laser intensity and tunneling current. Accordingly, this method reduces the measurement time and, hence, enables the spatial mapping of time-resolved (TR) signals (Terada et al. 2011; Yoshida et al. 2012; Yamashita et al. 2005).

#### 3.3  Probing of Carrier Dynamics by OPP-STM

In this section, some examples of the OPP-STM measurement of semiconductors are shown. In STM for a semiconductor, a nanoscale metal-insulator-semiconductor (MIS) junction is formed by the STM tip, tunneling gap, and sample. In the case of a reverse-bias voltage between the STM tip and sample, bias voltage leakage causes tip-induced band bending (TIBB). Under optical illumination, the redistribution of photocarriers reduces the electric field, thereby reducing the band bending, i.e., surface photovoltage (SPV)(McEllistrem et al. 1993; Takeuchi et al. 2004a, b; Yoshida et al. 2007) is generated.

Figure 3.5 shows schematic illustrations to explain the mechanism. TIBB appears under a reverse-bias voltage (a). Under photoillumination, the redistribution of the photocar-riers reduces the band bending, i.e., SPV is generated (b). Then, the excited state subsequently relaxes to the original state through two processes (c). One is the decay of photo-carriers on the bulk side via recombination, drift, and diffusion, known as bulk-side decay. The other is the decay of minority carriers trapped at the surface via recombination and thermionic emission, called surface-side decay (Yokota et al. 2013).

When the second pulse arrives during the bulk-side decay, the density of carriers induced by the second pulse decreases owing to absorption bleaching. In such a case, the SPV, and thus the total tunneling current, changes as a function of the delay time. On the other hand, when the second pulse arrives during the surface-side decay, fewer photocarriers are trapped at the surface owing to the existence of SPV induced by the first pulse, resulting in a change in the dependence of the total tunneling current on the delay time. Therefore, by measuring the tunneling current as a function of delay time, we can obtain information about the carrier dynamics in both processes.

Figure 3.5   Band structures of a semiconductor under OPP-STM measurement (a) before, (b) during, and (c) after illumination.

In the next section, the decay processes of excited carriers probed by bulk-side decay are first discussed. Then, local carrier dynamics is discussed using the results obtained by the observation of surface-side decay.

#### 3.3.1  Carrier Dynamics in a GaAs/AlGaAs/LT-GaAs Heterostructure

Figure 3.6 shows the result obtained for a heterostructure consisting of GaAs, AlGaAs, and low-temperature-grown (LT-GaAs). Figure 3.6a shows a schematic illustration of the experimental setup. Figure 3.6b and c show an STM image and OPP-STM spectra obtained at each region, respectively (Terada et al. 2011). The lifetimes of photocarriers in GaAs and LT-GaAs obtained by the OPP method are 6 ns and 1.5 ps, respectively. Since the excitation energy is lower than the gap energy of AlGaAs, no signal is acquired there.

Figure 3.6   (a) Schematic illustration of OPP-STM measurement setup. (b) STM image of the sample of a GaAs/AlGaAs/ LT-GaAs heterostructure sample (LT-GaAs: low-temperature-grown GaAs). (c) OPP-STM spectra obtained above GaAs, AlGaAs, and LT-GaAs.

In OPP-STM, measurement can be carried out wherever required because the probe method is STM. The decay constants obtained by OPP-STM are 4.8 ns for GaAs and 4.0 ps for LT-GaAs, which are in good agreement with those obtained by the OPP method. Of course, there is a slight difference because OPP-STM picks up local information.

Using this microscopy technique, the imaging of carrier dynamics information becomes possible. Figure 3.7a shows the photocarrier dynamics obtained along a line crossing an AlGaAs/GaAs interface. Here, instead of measuring a full spectrum by changing the delay time, the STM tip was scanned with the delay time fixed at 300, 600 fs, and so forth. Namely, the lines in Figure 3.7a show the carrier density at each delay time after photoexcitation. The carrier density decreases the delay time. If a two-dimensional (2D) scan is performed over the surface, as schematically shown in Figure 3.7b, 2D map of the time-dependent signal can be obtained, as shown in Figure 3.7c. A decay constant map can be obtained from the full series of 2D maps by fitting the change in density at each point.

#### 3.3.2  Modulation of Carrier Dynamics in GaAs Pin Structure

Information on the carrier dynamics modulated by a local potential is important for understanding nanoscale physics and its application to the development of current devices with nanoscale structures. Here, an example of the carrier dynamics obtained for a GaAs pin structure is discussed (Yoshida et al. 2012). Figure 3.8a and b, respectively, show schematic illustrations of the experimental setup and the inner potential of the pin structure. In the p- and n-type regions, since the band is flat, recombination is the main process for the decay of photocarriers. In contrast, since there is a slope in the inner potential in the i-region, there must be some effect of the inner potential on the carrier dynamics in the region.

Figure 3.7   (a) One-dimensional OPP-STM signals obtained along a line crossing the /LT-GaAs interface of the sample shown in Figure 3.6b. (b) Schematic illustration of two-dimensional scan over a sample. (c) Two-dimensional OPP-STM images obtained at four delay times (0, 31 ps, 11 ns, 200 ns).

Figure 3.8   (a) Schematic illustration of the experimental setup of OPP-STM over a GaAs pin structure. (b) Schematic illustration of the inner potential of a GaAs pin structure. (c) Four OPP-STM images obtained above the GaAs pin sample at td = 0, 2, 4, and 11 ns. (E) Map of decay constant obtained from a series of two-dimensional OPP-STM images. (F) Cross section along the white line in (E).

Figure 3.8c (A-D) shows four images taken from a series of 2D maps of TR signals. Since a reverse-bias voltage between the STM tip and sample is necessary to measure the TR signal, only the left half of each figure is considered here. To observe the dynamics in the right half, we must change the sign of the bias voltage.

The carrier density decreases with the time after photoex-citation. The decay-constant map is obtained from the full series of 2D maps by fitting the change in carrier density at each point. Figures E and F in Figure 3.8c show the decay-constant map and the cross section along the white line in e, respectively. As expected, the decay constant decreases in the i-region. Namely, the carrier density decreases via drift and diffusion rather than via recombination in this region.

Another noteworthy point is that there is a fluctuation in the decay-constant map, suggesting the effect of local structures, such as atomic-scale defects, on the carrier dynamics. The examination of such phenomena by OPP-STM will be discussed in Section 3.3.3 and 3.3.4.

#### 3.3.3  Atomic-Level Analysis

When some metals are deposited on a semiconductor surface, gap states are formed, which also modulate carrier dynamics. In STM on a semiconductor under a reverse-bias, TIBB occurs owing to the leakage of the applied bias voltage (McEllistrem et al. 1993; Takeuchi et al. 2004a, b; Yoshida et al. 2007), as explained in Section 3.2 (Figure 3.5). When a sample surface is illuminated using this condition, holes are trapped at the surface to reduce the band bending, thus generating SPV. If there is a gap state, holes trapped at the surface recombine with the electrons tunneling from the STM tip at the gap state as shown in Figure 3.9a. There are two limitations in this process, i.e., the injection of tunneling current from the STM tip and the capture rate of holes at the gap states. For a sufficient amount of tunneling current, the capture rate becomes the limiting process, which can be adjusted by changing the tip-sample distance. OPP-STM was applied to directly measure the hole capture rate at the atomic scale.

Figure 3.9c and d show STM images of manganese and iron atoms deposited on a GaAs(110) surface, respectively. Ga atoms are replaced by them as shown in Figure 3.9b. TR-STM measurements were carried out by placing the STM tip above these structures. Figure 3.9e and f show the spectra obtained. The decay constants were 1.6 and 14.3 ns for Mn/GaAs and Fe/GaAs, respectively. Although their structures are similar, the hole capture rate at the Mn site is one order faster, which is caused by the difference in their energy levels. Single atomic-level analysis is thus possible, which is expected to play an important role in analyzing the effects of dopants and atomic-level defects in semiconductor technologies.

Understanding and control of the quantum dynamics in nanoparticles and their interfaces with surrounding materials are important and play essential roles in various fields, such as semiconductor devices, catalysis, and energy transfer. In the OPP-STM measurement described above, the size dependence of carrier dynamics can be analyzed. The decay process should depend on the gap-state density, namely, the nanoparticle size. As expected, the decay constant was observed to increase with the decreasing nanoparticle size (Terada et al. 2010b; Yoshida et al. 2013b).

#### 3.3.4  Effect of Atomic Step on Carrier Dynamics

Understanding and control of the effects of atomic steps and dislocations on carrier dynamics are important issues in material physics and the development of current devices, such as power devices using SiC and GaN (Bergman et al. 2001; Zhang et al. 2003). Here, OPP-STM is carried out on a GaAs surface step, and its effect on the carrier dynamics is observed from the surface-side decay, which is sensitive to the local carrier dynamics.

Figure 3.9   (a) Schematic illustration of the band structure during OPP-STM measurement above a semiconductor with a gap state. (b) Schematic illustration of (Mn/Fe)/GaAs structure. A Ga atom is replaced by a Mn/Fe atom. (c and d) STM images of GaAs surfaces with Mn and Fe atoms deposited on them, respectively. (e and f) OPP-STM spectra obtained above a Mn atom and an Fe atom in (c) and (d), respectively.

Figure 3.10a shows an STM image of the GaAs surface with an atomic step, and Figure 3.10b shows the spectra obtained at an area without the step and above the step. To obtain accurate spectra, the laser intensity was adjusted to enhance the effect of the step (Terada et al. 2010a, b). The large amplitude of the signal at the atomic step is due to the large SPV originating from the Coulomb potential caused by the negative charges at the defect (Yoshida et al. 2008). The decay constants away from and at the atomic step were obtained to be 118 and 81 ns, respectively, clearly showing the effect of the atomic-step defect.

An atomic step forms gap states owing to the existence of dangling bonds, which act as traps to enhance the recombination of carriers. In OPP-STM, electrons tunneling from the STM tip combine with holes trapped at the gap state, as in the case of (Mn, Fe)/GaAs discussed in Section 3.3.3. Under a condition without TIBB [ ], holes recombine with electrons in the valence band via gap states, which is considered to occur at dislocations by a similar mechanism.

#### 3.4  Probing Spin Dynamics by OPP-STM

Using circularly polarized light for pump and probe optical pulses, spin dynamics, which has been studied by, for example, spin-polarized STM (SP-STM) (Loth et al. 2010; Brede et al. 2012), can be observed, as has been carried out by the OPP method with optical orientation techniques (Takeuchi et al. 1990; Mirlin 1984). The mechanism is similar to that of absorption bleaching. Spins are optically oriented by circularly polarized light and their dynamics are probed by STM. Here, as an example, right-handed circularly polarized light is used for the pump and probe pulses and down spins are excited (Figure 3.11a), which are randomized with time. When down spins excited by the first pulse remain in the excited states, the excitation of down spins by the second pulse as a function of the delay time is suppressed, by absorption bleaching (Figure 3.11b). In this case, the number of carriers excited by the second pulse increases with increasing delay time. The change in the number of excited carriers with the delay time produces a TR signal reflecting the spin dynamics (Yoshida et al. 2014).

First, we show the results of measurements above quantum wells (QWs) in Figure 3.12a, where the randomization of spins oriented in QWs was observed (Figure 3.12b). The sample was grown on a GaAs(100) surface from left to right, and two QWs with widths of 6 and 8 nm were formed. A 200 nm GaAs layer was placed as a spacer to isolate the two QWs. The sample was cleaved to prepare a clean (110) surface. Since OPP-STM is an STM, we can deduce where the location of the QWs are. Therefore, after observing the surface, the STM tip was placed above the QWs and measurements were carried out. The spectra shown in Figure 3.12c and d show the decay of spin orientation observed for the 6 and 8 nm QWs, respectively. The spin lifetime in the GaAs substrate was about 12 ps. In the QWs, the spin lifetime increased with increasing QW width and was as 68 ps for the 6 nm QW and 112 ps for the 8 nm QW.

Figure 3.10   (a) STM image of a GaAs surface with a surface step. (b) OPP-STM spectra obtained above the step (∎) and in a terrace far from the step (⋳).

Figure 3.11   (a) Schematic illustration of OPP-STM setup with pump-probe pulses of circularly polarized light. (b) Mechanism of absorption bleaching for spin excitation.

Figure 3.12   (a) Experimental setup of OPP-STM for GaAs QWs. (b) Schematic illustration of relaxation of spins oriented in a QW. (c and d) OPP-STM spectra obtained above QW of 6 and 8 nm width, respectively.

When a magnetic field is applied, spin precession occurs around the axis of the magnetic field. Thus, the number of carriers excited by the second pulse depends on the rotation angle owing to absorption bleaching. Namely, the signal intensity decreases with oscillation at the Larmor frequency as shown in Figure 3.13. From the relationship between the observed frequency and the applied magnetic field, ωL = BB/h, the local g factor can be evaluated. In addition, the decay constant of the spin orientation indicates the environmental conditions of electrons, for example, whether the electrons are free or trapped by defects.

The combination of this TR technique of examining spin dynamics with SP-STM (Loth et al. 2010; Brede et al. 2012) is an interesting target.

Figure 3.13   OPP-STM signal of spin precession obtained for a GaAs sample under several magnetic fields. Schematic of spin precession producing the oscillation of signal is shown in the lower part.

#### 3.5.1  Optical Pulses

In this section, we discuss OPP-STMs that can be realized using new laser technologies. In ordinary laser pulses, some cycles are included as shown in Figure 3.14a, whose phase is called the carrier envelope phase (CEP) (Jones et al. 2000). The CEP is random and fluctuates in pulses, which is the reason why the pulse width limits the time resolution. Recently, new laser technologies have become applicable, where the CEP is the same and locked in all pulses, as shown in Figure 3.14b. Furthermore, the CEP can be controlled as shown in Figure 3.14c. In Figure 3.14b and c, CEPs of zero and π are shown, respectively. On the basis of such control of the CEP, a new microscopy technique, THz-STM, has been developed, which is explained in 3.5.2.

#### 3.5.2  Terahertz-STM (THz-STM)

One of the new microscopy techniques that has been extensively studied is THz-STM (Cocker et al. 2013, 2016; Yoshioka et al. 2016). Figure 3.15a shows a schematic illustration of THz-STM. THz pulses are generated by irradiating a crystal with a titanium-sapphire laser. As shown in Figure 3.15b, the THz pulse obtained has an electric field of almost a single cycle, and the wavelength depends on the pulse width of the excitation laser. Here, an example of a THz pulse with a width of about 1 ps generated by a 130 fs pulse laser is shown.

Figure 3.14   (a) Ordinary ultrashort laser pulses that include several waves in each pulse. The phase of such a wave in a pulse is called the carrier envelope phase (CEP). The CEP is random and fluctuates in each pulse. (b and c) Newly developed ultrashort pulses in which the CEP is fixed (zero and π for (b) and (c), respectively) and the same in all pulses.

Figure 3.15   (a) Schematic illustration of THz-STM. (b) Example of a THz electric field. The inset shows its spectrum in the frequency space. OAP: off-axis parabolic mirror.

In the case of a THz pulse, the CEP is locked automatically and can be controlled as shown in Figure 3.16a. The electric field has opposite directions for CEP values of zero and π, for example, which can be used to apply a bias voltage between the STM tip and the sample. In addition, since the bias voltage is applied as a very short pulse, a high bias voltage, which generally causes damage, can be applied in this case. Figure 3.16b shows an image of a graphite obtained by THz-STM. Atomic resolution was achieved.

In the OPP—STM described in the previous sections, TR signals were measured on the basis of the mechanism of absorption bleaching. In contrast, in THz-STM, a bias voltage can be applied at any delay time, thus behaves similarly to a stroboscope, allowing snapshot to be taken. There are several ways of carrying out TR measurements. One way is to combine infrared (IR) laser pulses with THz pulses, as shown in Figure 3.16c, in which an IR pulse is used as a pump pulse to excite the sample and a THz pulse is used as a probe to observe the dynamics induced by the IR pulse. Furthermore, when CEP-controlled pulses with a single electric field are used for pump and probe pulses, sub-cycle STM measurement can be achieved, in which the shorter pulse is used as a probe pulse. Namely, dynamics controlled by the electric field in a CEP-controlled pump pulse can be examined in detail by a CEP-controlled probe pulse with sub-cycle time resolution.

Figure 3.16   (a) THz electric field for CEP = zero, π/2 and π. The schematic illustrations on the right side show the directions of the electric fields. (b) OPP-STM image obtained for graphite, in which a THz electric field was used as a bias voltage. (c) Schematic of OPP-THz-STM.

#### 3.5.3  Analysis of Laser Pulse Shape

The electric field below an STM tip is modulated from the incident field, as shown in Figure 3.17a, and also revealed by the antenna theory (Wang et al. 2004). Thus, to obtain accurate results using THz-STM, we need a method of evaluating the laser beam shape after modulation during STM measurement. In general, the original THz pulse shape is measured by an electro-optic (EO) sampling method, which is a technique for analyzing the shape of an electric field (Figure 3.17b). When a THz pulse and a shorter IR pulse enter a crystal, the polarization plane of the IR pulse is rotated by an amount depending on the intensity of the simultaneously existing THz electric field. Thus, a THz waveform can be probed if the rotation angle is measured while varying the delay time.

Next we discuss how to observe a THz pulse modulated at the tip apex. The combination of an IR pulse and a THz pulse is used for the measurement, as shown in Figure 3.17c. The tip apex is irradiated by the shorter pulse (517 nm, here) while the barrier height is reduced by the THz pulse, as shown in Figure 3.17d, which produces photoemission upon the shorter-pulse irradiation. Since the photocurrent depends on the intensity of the THz pulse, the THz pulse can be measured by varying the delay time (Wimmer et al. 2014; Herink et al. 2014).

Figure 3.17   Schematics showing (a) EO sampling, (b) modulation of THz electric field by STM tip to Near field indicate incident and modulated THz pulses, respectively (c) measurement setup to probe THz electric field using photoemission, and (d) measurement mechanism in which hot electrons are produced by an IR pulse while the barrier height is modulated by a THz pulse.

Figure 3.18   (a) Image of the STM tip used in an experiment to measure the laser pulse shape. (b) Map of photocurrent intensity. (c) Map of THz pulse intensity at td = 0. (d) Incident electric field (top) and THz electric fields obtained at positions A, B, and C in (c).

Figure 3.18 shows an example obtained by the measurement (Yoshida et al. 2019). Figure 3.18a shows an image of the tungsten tip used for this measurement. Measurements were carried out by changing the IR spot position on the STM tip. Figure 3.18b shows a map of the photocurrent. The measurements were carried out at each pixel, and the values were mapped in the color scale. THz waveforms were measured by varying the delay time at each pixel. Figure 3.18c shows a map of the signal intensity at zero delay, and Figure 3.18d shows the THz waveform obtained at the three positions (A, B, and C in Figure 3.18c), which have different shapes from that of the original THz pulse shown in the top of Figure 3.18d. At the tip apex, the signal intensity is strongest according to the enhancement by the tip, while the photocurrent is weak because the size of the tip is smaller than the spot size of the IR pulse. This illustrates the importance of checking pulse shape under the measurement conditions. This technique is applicable to the evaluation of near field in other experiments.

#### 3.6  Other Techniques

The combination of the OPP technique with multiprobe STM (Hasegawa 2007) is an attractive application. Using this microscopy, we can probe local dynamics under well-controlled nanoscale operating conditions, enabling photo-induced ultrafast dynamics in organized small structures to be understood in more detail. For example, anisotropy in conductivity can be analyzed by changing the directions of the two probes. Furthermore, the depth profile of dynamics may be analyzed using multiple probes. When circularly polarized light is used, spin dynamics can be probed, as described in Section 3.4. The technique in combination with SP-STM is expected to be used to examine spin dynamics in more detail.

To increase the generality of this microscopy technique and its applicability to various types of materials and structures, additional techniques are desired. For example, as an application of TR-STM based on a method similar to that used for the analysis of a molecular electronic structure, two-photon absorption measurement was used together with TR-STM (Wu & Ho 2010). For further advances in OPP-STM, the development of direct techniques for detecting photocur-rent in transient dynamics on the nanoscale is expected to play an important role in determining the optical characteristics of materials and devices. OPP-STM measurements of the transient photocurrent dynamics in the layered n-type semiconductor n-WSe2 have been carried out (Yoshida et al. 2013a, b). Since WSe2 has an indirect bulk band gap, the recombination lifetime of photoexcited carriers is significantly long (~10 μs) compared with the diffusion process. Therefore, under a forward bias voltage, i.e., a negative sample bias voltage, excited electrons are considered to directly tunnel to the STM tip as photocurrent, which has been clearly observed.

When the energies of pump and probe pulses are chosen appropriately, excited states may be included in the analysis similar to in the OPP method. The development of these techniques is considered to increase the generality of this microscopy technique, making it more practical and applicable to various phenomena. The essential mechanism of this microscopy technique involves the nonlinear interference between the excitations in the transient tunneling current generated by the two laser pulses; therefore, the introduction of new ideas is desirable to achieve further advances. By performing a stochastic analysis, the transition rate of a molecular conformation was clearly observed (Li et al. 2017).

Near-field scanning optical microscopy (NSOM) is used not only for spectroscopy but also for manufacturing nanoscale structures (Ohtsu 1998, 2003, 2008), and STM combined with synchrotron radiation has been opening the door to the characterization of atomic species by exciting core-level electrons (Okuda et al. 2009; Saito et al. 2006; Cummings et al. 2012).

#### 3.7  Summary

Spectroscopy techniques for nanoscale analysis developed by the combination of laser technologies with STM have been overviewed. For the combination of STM with ultrashortpulse laser technologies, the spatial resolution of STM and the temporal resolution of an ultrashort-pulse laser in the femtosecond range are simultaneously achieved by the delay-time modulation method developed using a pulse-picking technique. On the basis of the examples shown in this chapter, the development of other new techniques is expected to achieve further advances in this field.

#### References

Bergman, J. P. , Lendenmann, H. , Nilsson, P. A. . et al. 2001. Crystal defects as source of anomalous forward voltage increase of 4H-SiC diodes. Mater. Sci. Forum 353–356: 299–302.
Bhushan, B. 2010. Scanning Probe Microscopy in Nanoscience and Nanotechnology. Berlin, Heidelberg, Germany: Springer.
Binning, G. , Rohrer, H. , Gerber, Ch . et al. 1982. Surface studies by scanning tunneling microscopy. Phys. Rev. Lett. 49: 57–61 (1982).
Brede, J. , Chilian, B. , Khajetoorians, A. A. . et al. 2012. Atomic-scale spintronics. In Handbook of Spintronics, ed. D. Awschalom , J. Nitta , Y. Xu , 757–784. Berlin, Heidelberg, Germany: Canopus Academic Publishing and Springer.
Cocker, T. L. , Jelic , V, Gupta, M. et al. 2013. An ultrafast terahertz scanning tunnelling microscope. Nat. Photonics. 7: 620–625.
Cocker, T. L. , Peller, D. , Yu, P. et al. 2016. Tracking the ultrafast motion of a single molecule by femtosecond orbital imaging. Nature. 539: 263–267.
Cummings, M. L. , Chien, T. Y. , Preissner, C. et al. 2012. Combining scanning tunneling microscopy and synchrotron radiation for high-resolution imaging and spectroscopy with chemical, electronic, and magnetic contrast. Ultramicroscopy 112: 22–31.
Grafström, S. 2002. Photoassisted scanning tunneling microscopy. J. Appl. Phys. 91: 1717–1753.
Hasegawa, S. 2007. Multi-probe scanning tunneling microscopy. In Scanning Probe Microscopy, ed. S. Kalinin , A. Gruverman , 480–505. New York: Springer.
Herink G. , Wimmer, L. , and Ropers . C. 2014. Field emission at terahertz frequencies: AC-tunneling and ultrafast carrier dynamics. New. J. Phys. 16: 123005.
Jones, D. J. , Diddams, S. A. , Ranka, J. K. . et al. 2000. Carrier-envelope phase control of femtosecond mode-locked lasers and direct optical frequency synthesis. Science 288: 635–639.
Kemiktarak, U. , Ndukum, T. , Schwab, K. C. et al. 2007. Radio-frequency scanning tunnelling microscopy. Nature 450: 85–88.
Li, S. , Chen, S. , and Li, J. 2017. Joint space-time coherent vibration driven conformational transitions in a single molecule. Phys. Rev. Lett. 119: 176002.
Loth, S. , Etzkorn, M. , Lutz, C. P. . et al. 2010. Measurement of fast electron spin relaxation times with atomic resolution. Science 329: 1628–1630.
Mamin, H. J. , Birk, H. , Wimmer, P. et al. 1994. High-speed scanning tunneling microscopy: Principles and applications. J. Appl. Phys. 75: 161–168.
McEllistrem, M. , Haase, G. , Chen, D. et al. 1993. Electrostatic sample-tip interactions in the scanning tunneling microscope. Phys. Rev. Lett. 70: 2471–2474.
Mirlin, D. N. 1984. Optical alignment of electron momenta in GaAs-type semiconductors. In Optical Orientation, ed. F. Meier , B. P. Zakharchenya , 133–172. Elsevier Science Publishers, North-Holland, Amsterdam.
Ohtsu, M. 1998. Near-Field Nano/Atom Optics and Spec-troscopy. Berlin, Heidelberg, Germany: Springer-Verlag.
Ohtsu, M. 2003. Progress in Nano-Electro-Optics II. Berlin, Heidelberg, Germany: Springer-Verlag.
Ohtsu, M. 2008. Progress in Nano-Electro-Optics VI. Berlin, Heidelberg, Germany: Springer-Verlag.
Okuda, T. , Eguchi, T. , Akiyama, K. et al. 2009. Nanoscale chemical imaging by scanning tunneling microscopy assisted by synchrotron radiation. Phys. Rev. Lett. 102: 105503.
Saito, A. , Maruyama, J. , Manabe, K. et al. 2006. Development of a scanning tunneling microscope for in situ experiments with a synchrotron radiation hard-X-ray microbeam. J. Shynchrotron Rad. 13: 220.
Shah, J. 1999. Ultrafast Spectroscopy of Semiconductors and Semiconductor Nanostructures. Berlin, Heidelberg, Germany: Springer.
Shigekawa, H. , Takeuchi, O. , Terada, Y. et al. 2010. Series: Handbook of Nanophysics, ed. K. Sattler , Boca Raton, FL: Taylor & Francis. vol. 6, Principles and Methods.
Takeuchi, A. , Muto, S. , Inata, T. et al. 1990. Direct observation of picosecond spin relaxation of excitons in GaAs/AlGaAs quantum wells using spin-dependent optical nonlinearity. Appl. Phys. Lett. 56: 2213–2215.
Takeuchi, O. , Yoshida, S. , and Shigekawa, H. 2004a. Light-modulated scanning tunneling spectroscopy for nanoscale imaging of surface photovoltage. Appl. Phys. Lett. 84: 3645–3647.
Takeuchi, O. , Aoyama, O. , Oshima, R. et al. 2004b. Probing subpicosecond dynamics using pulsed laser combined scanning tunneling microscopy. Appl. Phys. Lett. 85: 3268–3270.
Takeuchi, O. , Aoyama, M. , Kondo, H. et al. 2006. Nonlinear dependences in pulse-pair-excited scanning tunneling microscopy. Jpn. J. Appl. Phys. 45: 1926–1930.
Terada, Y. , Yoshida, S. , Takeuchi, O. et al. 2010a. Laser-combined STM for probing ultrafast transient dynamics. J. Phys. Cond. Mat. 22: 264008–264015.
Terada, Y. , Yoshida, S. , Takeuchi, O. et al. 2010b. Realspace imaging of transient carrier dynamics by nanoscale pump-probe microscopy. Nat. Photonics 4: 869–874.
Terada, Y. , Yoshida, S. , Takeuchi, O. et al. 2011. Laser-combined scanning tunneling microscopy on the carrier dynamics in low-temperature-grown GaAs/AlGaAs/GaAs. Adv. Opt. Tech. 2011: 510186–510196.
Wang, K. , Mittleman, D. M. , van der Valk, N. C. J. . et al. 2004. Antenna effects in terahertz apertureless near-field optical microscopy. Appl. Phys. Lett. 85: 2715–2717.
Wimmer, L. , Herink, G. , Solli, D. R. . et al. 2014. Terahertz control of nanotip photoemission. Nat. Photonics 10: 432–436.
Wintterlin, J. , Trost, J. , Renisch, S. et al. 1997. Realtime STM observations of atomic equilibrium fluctuations in an adsorbate system: O/Ru(0001). Surf. Sci. 394: 159–169.
Wisendanger, R. 1994. Scanning Probe Microscopy and Spectroscopy. Cambridge: Cambridge University Press.
Wu, S. W. and Ho, W. 2010. Two-photon-induced hotelectron transfer to a single molecule in a scanning tunneling microscope. Phys. Rev. B 82: 085444–085451.
Yamashita, M. , Shigekawa, H. , and Morita, R. 2005. Mono-Cycle Photonics and Optical Scanning Tunneling Microscopy-Route to Femtosecond Angstrom Technology. Berlin, Heidelberg, Germany: Springer.
Yokota, M. , Yoshida, S. , and Mera, Y. 2013. Bases for time-resolved probing of transient carrier dynamics by optical pump-probe scanning tunneling microscopy. Nanoscale 5: 9170–9175.
Yoshida, S. , Kanitani, Y. , Oshima, R. et al. 2007. Microscopic basis for the mechanism of carrier dynamics in an operating p—n junction examined by using light-modulated scanning tunneling spectroscopy. Phys. Rev. Lett. 98: 026802–026805.
Yoshida, S. , Kanitani, Y. , Takeuchi, O. et al. 2008. Probing nanoscale potential modulation by defect-induced gap states on GaAs(110) with light-modulated scanning tunneling spectroscopy. Appl. Phys. Lett. 92: 102105–102107.
Yoshida, S. , Terada, Y. , Oshima, R. et al. 2012. Nanoscale probing of transient carrier dynamics modulated in a GaAs-PIN junction by laser-combined scanning tunneling microscopy. Nanoscale 4: 757–761.
Yoshida, S. , Terada, Y. , Yokota, M. et al. 2013a. Direct probing of transient photocurrent dynamics in p-WSe2 by time-resolved scanning tunneling microscopy. App. Phys. Exp. 6: 016601–016604.
Yoshida, S. , Yokota, M. , Takeuchi, O. et al. 2013b. Singleatomic-level probe of transient carrier dynamics by laser-combined scanning tunneling microscopy. App. Phys. Exp. 6: 032401.
Yoshida, S. , Aizawa, Y. , Wang, Z. et al. 2014. Probing ultrafast spin dynamics with optical pump-probe scanning tunnelling microscopy. Nat. Nanotechnol. 9: 588–593.
Yoshida, S. , Hirori, H. , Tachizaki, T. et al. 2019. Sub-cycle transient scanning tunneling spectroscopy with visualization of enhanced terahertz near-field. ACS Photonics. 6: 1356–1364.
Yoshioka, K. , Katayama, I. , Minami, Y. et al. 2016. Realspace coherent manipulation of electrons in a single tunnel junction by single-cycle terahertz electric fields. Nat. Photonics. 10: 762–765.
Zhang, A. P. , Rowland, L. B. , Kaminsky, E. B. . et al. 2003. Correlation of device performance and defects in AlGaN/GaN high-electron mobility transistors. Eastman J. Electron. Mater. 32: 388–394.

## Use of cookies on this website

We are using cookies to provide statistics that help us give you the best experience of our site. You can find out more in our Privacy Policy. By continuing to use the site you are agreeing to our use of cookies.