1. Introduction

Several effects are recognized to produce light using metal nanostructures in the field of nanoplasmonics. When applying a pulsed laser to a metal surface or nanostructure, light can be generated by nonlinear processes including second harmonic generation [1–3], third harmonic generation [4,5], two-photon photoluminescence [6–9] and three-photon photoluminescence [5,6]. In each of these effects, the photon energy of the emitted photons is greater than the incident energy. Notably, these nonlinear effects have extremely low conversion efficiency even for resonant metal structures including an additional nonlinear material. The conversion efficiency is typically less than a fraction of a percent [10–15].

Another well-known effect in nanoplasmonics is inelastic tunneling-induced light emission. Metal tunnel junctions under DC bias can produce light if the tunneling electron scatters inelastically [16–18]. The signature of this effect is that the photon energy produced has a cut-off equal to the bias for the electron across the tunnel junction. Recently, a high efficiency of 2% has been reported for this effect [19,20] and it was suggested that even higher efficiency values are possible with improved design [21].

It has also been noted that optical pulses can be used to create tunneling based ejection of electrons from metal surfaces [22–26]. The electric field within the femtosecond laser pulse lowers the barrier for tunneling based electron ejection. Many other works have explored the properties of electron ejection via tunneling out of a metal with an applied field and an ultrafast laser [27–36]. THz pulses have been used to produce ultrafast scanning tunneling microscopy [37], which followed from ultrafast scanning tunneling measurements [38]. Within these ultrafast works, the possibility of a tunneling electron inelastically scattering to produce a photon has not been explored.

Here we report bright upconverted light emission at a metal tunnel junction by the effect of inelastic tunneling: light-induced inelastic tunneling emission (LITE). We demonstrate that for sub-nanometer tunnel junctions excited by short infrared laser pulses, bright broad-spectrum visible light is emitted. The critical feature is that the emission shows the tunneling signature: the spectrum blue-shifts with increasing electric field of the incident pulse. The observed emission wavelength range is consistent with finite-difference time-domain (FDTD) simulations that show that the magnitude of the voltage across the junction is comparable to cut-off voltage (energy divided by electric charge) of the emitted photon energy for the peak pulse intensities used.

2. Results

The samples were gold nanoparticles (5 nm – 741949 Sigma-Aldrich) over an 30 nm thick ultra-flat gold (template stripped off silicon) with an amino-alkane-thiol self-assembled monolayer with varying carbon length (C2 – 30070 Sigma-Aldrich, C3 – 739294, Sigma-Aldrich, C6 – 733679 Sigma-Aldrich C8 – 745774 Sigma-Aldrich) [39,40]. (Samples were also fabricated with 20 nm and 60 nm, showing similar results, but these are not reported here – see Appendix for details of the fabrication and scanning electron microscope images). Dark-field scattering of the sample showed plasmon resonances. The transverse resonance was around 550 nm and the longitudinal resonances red-shifted with decreasing gap size produced by the self-assembled monolayer (around 680 nm for the C3 junction). The dark field measurement setup and results are shown in the Appendix. While the nanoparticles are stabilized in citrate, it is generally considered that the attraction to the amine group displaces the citrate and so this does not add thickness [12,39]. If the citrate were to stay on the nanoparticle, the thickness would be around 1 nm and the plasmon shift would be substantially less [41].

The femtosecond excitation with emission and the experimental setup for LITE is shown schematically in Figs. 1(a) and (b). 100 fs pulses at 1560 nm centre wavelength were incident on the sample at 58$^\circ$ angle. Our previous simulations [12] showed incident angles around $50^\circ$ resulted in the largest intensity, and for our experimental setup we found that the $58^\circ$ incident angle gave the best signal. A band-pass filter (Edmund Optics TECHSPEC Bandpass Filter 1575/50 nm 87871) was used to remove any spurious signal from the source’s pump laser or harmonic generation at the source.

 

Fig. 1. Light-induced Inelastic Tunneling Emission. (a) Schematic of metal nanoparticle over ultraflat gold with a self-assembled monolayer junction. An incident light pulse drives electron tunneling which emits a higher energy photon due to inelastic scattering. (b) Schematic of experimental setup for observing LITE emission: M - mirror, BF - band-pass filter, NDF - neutral dentsity filter, obj - 50$\times$ objective, FM - flip mirror, L - lens, CCD - CCD camera, spec - spectrometer. (c) Emission spectra for four different average powers, showing characteristic spectral shift of LITE. The sample was a 5 nm gold nanoparticle and a 0.69 nm junction. (d) Decay of LITE observed for a 0.51 nm junction at lower power excitation. The decay was much faster at high intensities (see Appendix).

Download Full Size | PDF

Figure 1(c) shows the recorded bright spectra of LITE recorded on a fiber-coupled photon-counting spectrometer (QE65000, Ocean Optics) with one second integration time. The spectra were corrected for transmission and detection efficiency by using a thermal source for calibration. The extremely bright emission from LITE can be observed by the naked eye. The total intensity over the spectrum from 400 nm to 950 nm is 33000 times brighter than third harmonic generation and 140000 times brighter than second harmonic generation (see Table 1). It is important to note that the signal was optimized in each case: maximizing the second harmonic signal and third harmonic signal each time (since the alignment is slightly different due to chromatic effects). We note that the instrumentation is not efficient for detection below 400 nm, and so the apparent cut-off at that wavelength is an artifact.

Table 1. Comparison of LITE with harmonic generation by integrated counts for the same acquisition time and maximized collection efficiency for each effect.

View Table

We also observed two photon photoluminescence (TPPL) for longer integration times, producing a broad spectrum. TPPL is easily distinguished from LITE because the TPPL spectra did not shift with varying excitation intensity. The characterization of the harmonic generation and TPPL is given in the Appendix (power and spectral dependence).

LITE was observed for two-carbon and three-carbon self-assembled monolayer junctions, with widths of 0.51 nm and 0.69 nm (see Appendix). The six-carbon and eight-carbon junctions did not produce LITE, which is consistent with these large junctions having negligible probability of tunneling, with widths of 0.94 nm and 1.16 nm. By contrast, the larger junctions still produced TPPL and harmonic generation (see Appendix).

LITE is short lived, which is suspected to be the result of tunneling induced breakdown of the junction. Pulsed laser field-induced breakdown in dielectrics has been studied extensively, and typically occurs at V/nm laser fields used here [42]. Even after LITE died out, the nonlinear effects of TPPL, SHG and THG persisted, although they were reduced.

Figure 1(d) shows the decay of the LITE for an incident average power of 45 mW with an exponential fit. The time-constant was twenty seconds. For higher intensities the decay was much faster; for example for 144 mW the decay was faster than the 100 ms integration of the spectrometer (see Appendix).

Other works have shown that the deformation of the shape of the metal nanoparticle can result in a blue shift, but this effect is permanent (melting), [7] which distinguishes it from the results presented here. Other works have also shown flickering of multi-photon fluorescence, but LITE differs from that flickering because it does not turn back on once it dies out [5]. The observed emission is only seen in junctions that are small enough to allow tunneling. Hot-electron effects can contribute to emission in the presence of tunneling; however, past analysis has suggested that this is a less probable emission process by two orders of magnitude [17].

The physical picture of LITE is that the AC field of the femtosecond pulse source produces enough voltage in the junction to induce electron tunneling. The tunneling electron has a finite probability of inelastic scattering, which produces photon emission from the tunneling event. The emitted photon has energy which is less than the applied voltage divided by the electron charge; the potential energy available to the tunneling electron. Thermal effects can give slightly higher energy photons. Figure 2(a) shows a schematic of this effect where the energy gap of the self-assembled monolayer presents a tunneling barrier and the energy bias across the barrier oscillates with the applied field from the femtosecond pulse.

 

Fig. 2. Numerical simulations of local field. (a) Schematic of LITE effect resulting from the AC field of a femtosecond laser oscillating the bias applied to the tunnel junction. The bias field is represented by a slope in the barrier potential. When the field induces tunneling, an upconverted photon can be emitted by inelastic scattering. (b) Finite-difference time-domain simulations provide the field strength as a function of the incident power in the experiment, which is translated to the cut-off wavelength for the LITE effect. The inset shows the local field intensity distribution in the junction region, normalized to the incident field intensity.

Download Full Size | PDF

To check the plausibility of this picture, we simulated the field amplitude at the 1560 nm pulse wavelength in the junction for the conditions of the experiment. We used FDTD simulations (FDTD - Lumerical v. 8.20.1731) with a mesh size of 0.05 nm, a total-field scattered field source, perfectly matched layer boundaries, and Johnson and Christy gold [43]. The inset to Fig. 2(b) shows the cross section of the field enhancement in the gap. We related the field amplitude to the cut-off wavelength for LITE:

Figure 2(b) shows simulated the cut-off wavelength as a function of the average power of the laser. The laser repetition rate was 80 MHz and peak intensity of the 33 mW laser was 1.6 $\times$ $10^{14}$ $\frac {W}{m^2}$. The cut-off is in the visible regime for the conditions of this experiment, showing a blue-shift as the peak intensity increases. Since the spectra were also influenced by the local plasmon resonances (see dark-field spectra in Appendix), we do not attempt to quantitatively fit the observed spectra although this may be attempted with techniques used elsewhere [21,44,45]. Figure 1(c) shows a small peak around 700 nm mainly buried in the emission. We believe is attributable to emission being influenced by the plasmon resonance. There is a similar effect seen for the C2 data just less than 800 nm (see Appendix).

We recognize potential for significant improvement to the LITE effect. First, we are not exploiting the plasmonic resonance in the present investigation: the incident photon energy is well-away from the plasmon resonance and so the field enhancement is only a factor of 6 with respect to the incident field (see Appendix). Better overall conversion efficiency is expected by fully exploiting resonances at the incident and emission wavelengths, particularly making use of the gap plasmon.

Second, we expect that the light emission is extremely short in duration, although we have not time-resolved the emission in this work. This is expected since the tunneling can operate at optical frequencies due to the small capacitance of the junction [46]. In particular, here we have a junction area of approximately 1-10 nm$^2$ with a gap of just under a nanometer; so the overall capacitance is below the aF range. In a well-designed case, the radiation resistance is less than a kOhm, so that the response time should be in the femtosecond regime. Therefore, this effect may be intriguing for ultra-fast nearfield excitation. In terms of the dynamic response, the present system does not have lead resistance since it is not coupled to a source by leads as with the usual inelastic tunneling; however, it is acting as an antenna with a radiative resistance which is similarly close to 100 Ohms [47]. Therefore, the RC time constant is expected to be short since the capacitance is in the aF range [48].

Third, LITE may be more efficient than conventional inelastic tunneling light emission because the light emission is limited in time to a cycle of the applied pulsed source. This means that the spectral bandwidth available is substantially narrower than DC-based emission and so less energy is expected to be lost to low energy photons that are typically not measured [49]. The inelastic tunneling emission has been viewed in the context of the power spectrum of the tunneling current, which allows emission for frequencies down to zero [50]. For the case of an oscillating tunneling current the tunneling direction switches and low frequency emission is suppressed. The oscillating field also means that the emission cut-off will sweep to longer wavelengths as the field is reduced, so this may negate the efficiency increase based on bandwidth limitations.

Finally, we note that the junctions here are extremely short-lived due to the self-assembled monolayer chosen. Others have investigated van der Waals hexagonal-BN as a stable junction for inelastic tunneling based emission [46,49]. That material is crystalline and has a large band-gap, so it is expected that LITE lifetimes will be extended substantially, as was observed in DC measurements. Despite this drop in intensity as the result of degradation of the junction, we note that the spectral shape is stable, as we have measured its evolution as the junction degrades and this is reported in the Appendix.

We recognize the possibility that multiphoton ionization is also occurring in this system as a way for electrons to transit across the barrier. However, if this were the dominant mechanism, then we would expect to see this effect even for barriers which are too large to allow for tunneling, since the electron would gain energy enough to go over the barrier. This is not the case in our present experiment – we only observed the flashes of emission for barriers formed by 2 and 3 carbon SAMs, and not for thicker barriers. While the present configuration is not one of ionization, we calculate the Keldysh parameter to be 0.8 for a peak field of 3V [51]. This suggests that multi-photon ionization is unlikely to be the dominant mechanism.

3. Conclusion

In summary, here we report bright upconverted emission from light induced inelastic tunneling between metals when they are incident with a pulsed laser source. Since we believe that this effect depends on tunneling based inelastic emission, it is limited by the efficiency of that effect, which is on the order of 2% per tunneling event and orders of magnitude more efficient than other nonlinear effects that operate in the near-field. By exploiting plasmonic resonances, the reduced tunneling bandwidth and more stable junction materials, it is expected that LITE will become a highly efficient ultrafast source of upconverted photons.

Appendix

Fabrication procedure

30 nm of Au was sputtered using a Mantis QUBE sputter deposition system on polished side of a silicon sample without any adhesion layer to allow for stripping. Silicon was used because of its crystalline properties that allows ultraflat gold after stripping. For stripping, UV cured epoxy was sandwiched between the gold on silicon and a glass slide. After exposing the epoxy to UV light, it hardened attaching the gold to glass and thus revealing the ultraflat gold surface. Samples were immersed in 3 mM ethanol solution of different self assembled monolayers (SAM) for 18 h. This procedure was followed by rinsing with ethanol 3-4 times to remove unattached SAMs and drying with nitrogen. The final step was to deposit nanoparticles (5 nm – 741949 Sigma-Aldrich $5.5\times 10^{13}$ particles/mL, 20 nm – 741965 Sigma-Aldrich $6.54\times 10^{11}$ particles/mL, 60 nm – 742015 Sigma-Aldrich $1.9\times 10^{10}$ particles/mL) by drop coating 500 $\mu$l of NPs in solution on sample and leaving it for 30 min, followed by rinsing samples 3-4 times with deionized water to remove unattached NPs and drying with nitrogen.

Figure 3 shows scanninng electron microscope (SEM) images. The sample with 20 nm NPs is substantially more dense as compared to the sample with 60 nm NPs because its initial solution used for deposition is 10 times more concentrated.

 

Fig. 3. SEM images of samples with a) 20 nm NPs and b) 60 nm NPs.

Download Full Size | PDF

Dark field measurement

The dark field setup is shown in Fig. 4. The sample was positioned in a way to make a 65$^\circ$ angle between normal of the sample and incident light. This was necessary to ensure that only the scattered light was collected. Scratches and agglomerations were avoided by looking at the dark field image before taking a measurement. An Ocean Optics LS-1 Tungsten Halogen Light Source was used as white light source. A 20$\times$ 0.28 NA Mitutoyo objective was used for incident light and 10$\times$ 0.28 NA Mitutoyo objective was used to collect scattered light. A QE65000 Ocean Optics spectrometer was used for signal collection.

 

Fig. 4. Dark-field measurement setup. WLS - white light source, L - lens, 20x objective, 10x objective, FM - flip mirror, CCD camera. In b) Incident and reflected light are represented with blue line and scattered light is represented with dashed red line.

Download Full Size | PDF

Dark-field measurement results are presented in Fig. 5. All the spectra show a red shift as the SAM thickness is reduced, as expected. The transverse mode can also be seen as a shoulder around 550 nm.

 

Fig. 5. Dark-field scattering of samples with a) 5 nm NPs, b) 20 nm NPs and c) 60 nm NPs. d) Dark-field scattering CCD camera image.

Download Full Size | PDF

Third harmonic generation measurement

Third harmonic generation (THG) measurement results are presented in Fig. 6. A Calmar Carmel sub-100 femtosecond laser centred at 1560 nm with average power of 100 mW was used as a laser source and signal was collected using a QE65000 Ocean Optics spectrometer with a ten second integration time. A bandpass filter was used to remove harmonics present in the laser and a neutral density filter was used to regulate the average power. When the SAM thickness was decreased, larger THG intensity was observed with a maximum for C3. THG intensity dropped for C2 due to onset of tunneling. [12]

 

Fig. 6. a) Third harmonic generation intensities for different SAMs of samples with 5 nm NPs and b) 60 nm NPs. c) CCD camera image of third harmonic generation. d) Third harmonic generation spectra.

Download Full Size | PDF

Figure 7 shows the power dependence of third harmonic generation with the slope of 3 for different SAMs, as expected from a three photon process.

 

Fig. 7. Third harmonic generation power dependence of samples with 5 nm NPs. Slope is 3 for all, as expected from a three photon process.

Download Full Size | PDF

Two photon photoluminescence

Figure 8(a) shows two photon photoluminescence (TPPL) for different incident powers. No shift is observed. The wavelength axis is truncated at 950 nm because of the range of spectrometer. Figure 8(b) shows the power dependence of TPPL with the slope of 2, as expected from a two photon process.

 

Fig. 8. a) TPPL with different incident power from C3 with 5 nm Au NPs. b) power dependence of TPPL with slope equal 2, as expected from a two photon process.

Download Full Size | PDF

Light-induced inelastic tunneling emission

Figures 9 and 10 show the normalized to maximum count and normalized to instrument response of LITE of the C2 sample. A blue shift was observed with larger intensities of the incident laser. The spectra are truncated at 950 nm because of the range of spectrometer.

 

Fig. 9. LITE of C2 with 5 nm Au NPs normalized to maximum count.

Download Full Size | PDF

 

Fig. 10. LITE of C2 with 5 nm Au NPs.

Download Full Size | PDF

Figure 11 shows the changing of LITE peak for different powers. The peak wavelength blue shifts with increasing power. Figure 12 shows that for larger incident laser power, LITE is faster than 100 ms second. In an attempt to measure decay time, integration time was reduced to 50 ms and 10 ms without showing sufficient time resolution to resolve the decay for these conditions. Figure 13 shows the slow decay of LITE observed for a 0.51 nm junction at low power excitation (45 mW). Thirty-two spectrum data each contain 1 s integration time are depicted in the figure showing the slow decay process. The decay process of the same data (integration of the counts) is shown in Fig. 1(d) of the main text. The decay was much faster at high intensities.

 

Fig. 11. Wavelength of LITE peak for different powers of C2 with 5 nm Au NPs.

Download Full Size | PDF

 

Fig. 12. Brief LITE observed for C3 with 144 mW average power

Download Full Size | PDF

 

Fig. 13. Slow decay of LITE observed for a C2 sample at lower power excitation.

Download Full Size | PDF

Simulation

A finite difference time domain (FDTD - Lumerical v. 8.20.1731) simulation was used with the following parameters: 0.05 nm mesh size in the gap and 10 nm over nanoparticle, refractive index of the gap is 1.4, total field scattered field source (100 fs pulse source with centre wavelength at 1560 nm) at the 58 degree angle (to account for off-axis alignment). Perfectly Matched Layer boundaries were used. Although C3 and C2 have subnanometer thickness (0.69 nm and 0.51 nm) [39], tunneling was neglected in the FDTD simulation because other works show that for C3 field amplitude in the gap for quantum corrected and classical models are similar [12,52].

Comparing electric field at the source and in the gap on Fig. 14, an enhancement factor of 5.8 was determined, and it was used to calculate the actual experimental electric field in the gap using:

(2)$$E_\mathrm{exp} = F\sqrt{\frac{E_\mathrm{s}^2I_\mathrm{exp}}{I_\mathrm{s}}}$$

where $F$ is enhancement factor, $E_{\mathrm {exp}}$ is field amplitude of the experiment, $E_{\mathrm {s}}$ field amplitude of the simulation, $I_{\mathrm {exp}}$ source intensity of the experiment calculated using Eq. (3), $I_{\mathrm {s}}$ source intensity of the simulation.

(3)$$I = \frac{P_\mathrm{peak}}{A}$$

where $P_{\mathrm {peak}}$ is peak laser power and A is illuminated area. $P_{\mathrm {peak}}$ was calculated from:

(4)$$P_\mathrm{peak} = \frac{P_\mathrm{avg}}{R\tau}$$

where $P_{\mathrm {avg}}$ is average source power, $R$ is repetition rate, $\tau$ is pulse width. Repetition rate is 80 MHz, pulse width 100 fs.

 

Fig. 14. Time-domain representation of the field in the gap.

Download Full Size | PDF

Once electric field amplitude of the experiment is known, cut off wavelength is calculated using:

(5)$$\lambda_c = \frac{hc}{Edq}$$

where $\lambda _c$ is cut-off wavelength, $h$ is Planck’s constant, $c$ is speed of light, $E$ is field amplitude, $d$ is gap size, $q$ is electron charge.

Funding

Natural Sciences and Engineering Research Council of Canada (RGPIN-2017-03830).

Acknowledgments

The authors thank the reviewers for providing suggestions to improve the discussion section.

Disclosures

The authors declare no conflicts of interest.

References

1. C. Chen, A. R. B. de Castro, and Y. Shen, “Surface-enhanced second-harmonic generation,” Phys. Rev. Lett. 46(2), 145–148 (1981). [CrossRef]  

2. H. Simon, D. Mitchell, and J. Watson, “Optical second-harmonic generation with surface plasmons in silver films,” Phys. Rev. Lett. 33(26), 1531–1534 (1974). [CrossRef]  

3. M. Kauranen and A. V. Zayats, “Nonlinear plasmonics,” Nat. Photonics 6(11), 737–748 (2012). [CrossRef]  

4. T. Y. Tsang, “Surface-plasmon-enhanced third-harmonic generation in thin silver films,” Opt. Lett. 21(4), 245–247 (1996). [CrossRef]  

5. M. Lippitz, M. A. van Dijk, and M. Orrit, “Third-harmonic generation from single gold nanoparticles,” Nano Lett. 5(4), 799–802 (2005). [CrossRef]  

6. G. Boyd, Z. Yu, and Y. Shen, “Photoinduced luminescence from the noble metals and its enhancement on roughened surfaces,” Phys. Rev. B 33(12), 7923–7936 (1986). [CrossRef]  

7. A. Bouhelier, R. Bachelot, G. Lerondel, S. Kostcheev, P. Royer, and G. Wiederrecht, “Surface plasmon characteristics of tunable photoluminescence in single gold nanorods,” Phys. Rev. Lett. 95(26), 267405 (2005). [CrossRef]  

8. P. Zijlstra, J. W. Chon, and M. Gu, “Five-dimensional optical recording mediated by surface plasmons in gold nanorods,” Nature 459(7245), 410–413 (2009). [CrossRef]  

9. M. R. Beversluis, A. Bouhelier, and L. Novotny, “Continuum generation from single gold nanostructures through near-field mediated intraband transitions,” Phys. Rev. B 68(11), 115433 (2003). [CrossRef]  

10. H. Aouani, M. Rahmani, M. Navarro-Cía, and S. A. Maier, “Third-harmonic-upconversion enhancement from a single semiconductor nanoparticle coupled to a plasmonic antenna,” Nat. Nanotechnol. 9(4), 290–294 (2014). [CrossRef]  

11. M. Airola, Y. Liu, and S. Blair, “Second-harmonic generation from an array of sub-wavelength metal apertures,” J. Opt. 7(2), S118–S123 (2005). [CrossRef]  

12. G. Hajisalem, M. S. Nezami, and R. Gordon, “Probing the quantum tunneling limit of plasmonic enhancement by third harmonic generation,” Nano Lett. 14(11), 6651–6654 (2014). [CrossRef]  

13. P. N. Melentiev, A. E. Afanasiev, A. A. Kuzin, A. S. Baturin, and V. I. Balykin, “Giant optical nonlinearity of a single plasmonic nanostructure,” Opt. Express 21(12), 13896–13905 (2013). [CrossRef]  

14. J. Van Nieuwstadt, M. Sandtke, R. Harmsen, F. B. Segerink, J. Prangsma, S. Enoch, and L. Kuipers, “Strong modification of the nonlinear optical response of metallic subwavelength hole arrays,” Phys. Rev. Lett. 97(14), 146102 (2006). [CrossRef]  

15. M. S. Nezami, D. Yoo, G. Hajisalem, S.-H. Oh, and R. Gordon, “Gap plasmon enhanced metasurface third-harmonic generation in transmission geometry,” ACS Photonics 3(8), 1461–1467 (2016). [CrossRef]  

16. J. Lambe and S. McCarthy, “Light emission from inelastic electron tunneling,” Phys. Rev. Lett. 37(14), 923–925 (1976). [CrossRef]  

17. B. Persson and A. Baratoff, “Theory of photon emission in electron tunneling to metallic particles,” Phys. Rev. Lett. 68(21), 3224–3227 (1992). [CrossRef]  

18. P. Johansson, R. Monreal, and P. Apell, “Theory for light emission from a scanning tunneling microscope,” Phys. Rev. B 42(14), 9210–9213 (1990). [CrossRef]  

19. W. Du, T. Wang, H.-S. Chu, and C. A. Nijhuis, “Highly efficient on-chip direct electronic–plasmonic transducers,” Nat. Photonics 11(10), 623–627 (2017). [CrossRef]  

20. H. Qian, S.-W. Hsu, K. Gurunatha, C. T. Riley, J. Zhao, D. Lu, A. R. Tao, and Z. Liu, “Efficient light generation from enhanced inelastic electron tunnelling,” Nat. Photonics 12(8), 485–488 (2018). [CrossRef]  

21. M. Parzefall and L. Novotny, “Light at the end of the tunnel,” ACS Photonics 5(11), 4195–4202 (2018). [CrossRef]  

22. P. Hommelhoff, Y. Sortais, A. Aghajani-Talesh, and M. A. Kasevich, “Field emission tip as a nanometer source of free electron femtosecond pulses,” Phys. Rev. Lett. 96(7), 077401 (2006). [CrossRef]  

23. P. Hommelhoff, C. Kealhofer, and M. A. Kasevich, “Ultrafast electron pulses from a tungsten tip triggered by low-power femtosecond laser pulses,” Phys. Rev. Lett. 97(24), 247402 (2006). [CrossRef]  

24. G. Herink, D. Solli, M. Gulde, and C. Ropers, “Field-driven photoemission from nanostructures quenches the quiver motion,” Nature 483(7388), 190–193 (2012). [CrossRef]  

25. P. Dombi, A. Hörl, P. Ràcz, I. Márton, A. Trürgler, J. R. Krenn, and U. Hohenester, “Ultrafast strong-field photoemission from plasmonic nanoparticles,” Nano Lett. 13(2), 674–678 (2013). [CrossRef]  

26. M. Krüger, M. Schenk, M. Förster, and P. Hommelhoff, “Attosecond physics in photoemission from a metal nanotip,” J. Phys. B 45(7), 074006 (2012). [CrossRef]  

27. H. Yanagisawa, M. Hengsberger, D. Leuenberger, M. Klöckner, C. Hafner, T. Greber, and J. Osterwalder, “Energy distribution curves of ultrafast laser-induced field emission and their implications for electron dynamics,” Phys. Rev. Lett. 107(8), 087601 (2011). [CrossRef]  

28. H. Yanagisawa, C. Hafner, P. Doná, M. Klöckner, D. Leuenberger, T. Greber, M. Hengsberger, and J. Osterwalder, “Optical control of field-emission sites by femtosecond laser pulses,” Phys. Rev. Lett. 103(25), 257603 (2009). [CrossRef]  

29. C. Ropers, D. Solli, C. Schulz, C. Lienau, and T. Elsaesser, “Localized multiphoton emission of femtosecond electron pulses from metal nanotips,” Phys. Rev. Lett. 98(4), 043907 (2007). [CrossRef]  

30. R. Bormann, M. Gulde, A. Weismann, S. Yalunin, and C. Ropers, “Tip-enhanced strong-field photoemission,” Phys. Rev. Lett. 105(14), 147601 (2010). [CrossRef]  

31. M. Schenk, M. Krüger, and P. Hommelhoff, “Strong-field above-threshold photoemission from sharp metal tips,” Phys. Rev. Lett. 105(25), 257601 (2010). [CrossRef]  

32. S. Irvine, A. Dechant, and A. Elezzabi, “Generation of 0.4-kev femtosecond electron pulses using impulsively excited surface plasmons,” Phys. Rev. Lett. 93(18), 184801 (2004). [CrossRef]  

33. B. Piglosiewicz, S. Schmidt, D. J. Park, J. Vogelsang, P. Groß, C. Manzoni, P. Farinello, G. Cerullo, and C. Lienau, “Carrier-envelope phase effects on the strong-field photoemission of electrons from metallic nanostructures,” Nat. Photonics 8(1), 37–42 (2014). [CrossRef]  

34. D. J. Park, B. Piglosiewicz, S. Schmidt, H. Kollmann, M. Mascheck, and C. Lienau, “Strong field acceleration and steering of ultrafast electron pulses from a sharp metallic nanotip,” Phys. Rev. Lett. 109(24), 244803 (2012). [CrossRef]  

35. J. Vogelsang, J. Robin, B. J. Nagy, P. Dombi, D. Rosenkranz, M. Schiek, P. Groß, and C. Lienau, “Ultrafast electron emission from a sharp metal nanotaper driven by adiabatic nanofocusing of surface plasmons,” Nano Lett. 15(7), 4685–4691 (2015). [CrossRef]  

36. S. Irvine and A. Elezzabi, “Surface-plasmon-based electron acceleration,” Phys. Rev. A 73(1), 013815 (2006). [CrossRef]  

37. T. L. Cocker, V. Jelic, M. Gupta, S. J. Molesky, J. A. Burgess, G. De Los Reyes, L. V. Titova, Y. Y. Tsui, M. R. Freeman, and F. A. Hegmann, “An ultrafast terahertz scanning tunnelling microscope,” Nat. Photonics 7(8), 620–625 (2013). [CrossRef]  

38. S. Weiss, D. Botkin, D. Ogletree, M. Salmeron, and D. Chemla, “The ultrafast response of a scanning tunneling microscope,” Phys. Status Solidi 188(1), 343–359 (1995). [CrossRef]  

39. C. Ciracì, R. Hill, J. Mock, Y. Urzhumov, A. Fernández-Domínguez, S. Maier, J. Pendry, A. Chilkoti, and D. Smith, “Probing the ultimate limits of plasmonic enhancement,” Science 337(6098), 1072–1074 (2012). [CrossRef]  

40. R. T. Hill, J. J. Mock, A. Hucknall, S. D. Wolter, N. M. Jokerst, D. R. Smith, and A. Chilkoti, “Plasmon ruler with angstrom length resolution,” ACS Nano 6(10), 9237–9246 (2012). [CrossRef]  

41. J.-W. Park and J. S. Shumaker-Parry, “Structural study of citrate layers on gold nanoparticles: role of intermolecular interactions in stabilizing nanoparticles,” J. Am. Chem. Soc. 136(5), 1907–1921 (2014). [CrossRef]  

42. N. Bloembergen, “Laser-induced electric breakdown in solids,” IEEE J. Quantum Electron. 10(3), 375–386 (1974). [CrossRef]  

43. P. B. Johnson and R.-W. Christy, “Optical constants of the noble metals,” Phys. Rev. B 6(12), 4370–4379 (1972). [CrossRef]  

44. F. Wang and Y. R. Shen, “General properties of local plasmons in metal nanostructures,” Phys. Rev. Lett. 97(20), 206806 (2006). [CrossRef]  

45. L. Novotny and B. Hecht, Principles of nano-optics (Cambridge university press, 2012).

46. M. Parzefall, P. Bharadwaj, A. Jain, T. Taniguchi, K. Watanabe, and L. Novotny, “Antenna-coupled photon emission from hexagonal boron nitride tunnel junctions,” Nat. Nanotechnol. 10(12), 1058–1063 (2015). [CrossRef]  

47. A. Alu and N. Engheta, “Input impedance, nanocircuit loading, and radiation tuning of optical nanoantennas,” Phys. Rev. Lett. 101(4), 043901 (2008). [CrossRef]  

48. D. Xiang, J. Wu, and R. Gordon, “Coulomb blockade plasmonic switch,” Nano Lett. 17(4), 2584–2588 (2017). [CrossRef]  

49. M. Parzefall, Á. Szabó, T. Taniguchi, K. Watanabe, M. Luisier, and L. Novotny, “Light from van der waals quantum tunneling devices,” Nat. Commun. 10(1), 292 (2019). [CrossRef]  

50. D. Hone, B. Mühlschlegel, and D. Scalapino, “Theory of light emission from small particle tunnel junctions,” Appl. Phys. Lett. 33(2), 203–204 (1978). [CrossRef]  

51. A. Zheltikov, “Keldysh parameter, photoionization adiabaticity, and the tunneling time,” Phys. Rev. A 94(4), 043412 (2016). [CrossRef]  

52. A. Khademi, T. Dewolf, and R. Gordon, “Quantum plasmonic epsilon near zero: field enhancement and cloaking,” Opt. Express 26(12), 15656–15664 (2018). [CrossRef]