betaFeedback

WorksPhysics and Astronomy

Detection of Surface Waves During Femtosecond Filamentation

Abstract

Ultrashort pulsed lasers (USPL) can produce thin columns of plasma in air via femtosecond filamentation, and these plasmas have been found to generate broadband TeraHertz (THz) and Radio Frequency (RF) radiation. A recent theory argues that the currents driven at the boundary of the plasma excite a Surface Plasmon Polariton (SPP) surface wave (in particular a Sommerfeld-Goubau wave given the cylindrical symmetry), which proceeds to detach from the end of the plasma to become the RF pulse. We have performed near-field measurements of these plasmas with a D-dot probe, and find an excellent agreement with this theory. The radial field dependence is precisely fit by a Hankel function, with an outer length scale in agreement with plasma conductivity and radius, and a measured longitudinal drift in frequency maxima closely matches both SPP simulations and analytic expectations.

Introduction

Surface waves can be excited across a spectrum of environments, including solitonic shallow water gravity waves [1, 2, 3], Rayleigh and Love surface acoustic waves (from microfluidic to planetary scale) [4, 5, 6], Dyakonov, Bloch, and Tamm electromagnetic surface waves along special dielectrics [7, 8, 9, 10, 11], and surface plasmon polaritons [12, 13] on the boundaries of conducting substrates with free electron number density nen_e. Surface plasmon polaritons are composed of an electromagnetic wave coupled to a charge density wave on the boundary of the metal, semiconductor or plasma region, and can exist under a cutoff frequency ωs\omega_s, which for a flat plane boundary equals

ωs=ωpl2,\omega_s = \frac{\omega_{\mathrm{pl}}}{\sqrt{2}},

where ωpl\omega_{\mathrm{pl}} is the plasma frequency.

SPPs are typically considered along the boundaries of metals and semiconductors in the optical and infrared bands, where they can be driven by Otto or Kretschmann coupling [14, 15], but they can be excited along the boundary of a plasma as well. An Ultra Short Pulse Laser (USPL) can generate volumes of plasma in a unique way: given sufficient power Kerr self-focusing will boost the intensity of a standard 800nm, 50fs, Ti:Sapphire pulse propagating in air to roughly 5×10175 \times10^{17} W/m2^2, at which point strong field ionization [16] creates a plasma of density roughly ne∼1022n_e \sim10^{22} m−3^{-3} and radius rpl∼100r_{\mathrm{pl}} \sim100 μm. USPL filamentation thus generates thin columns of plasma that grow longer at the speed of light [17, 18, 19], which have a wide array of potential applications [20, 21, 22, 23, 24].

In addition to creating the plasma, the femtosecond pulse also endows the plasma with initial current densities. A J×BJ \times B Lorentz force drives a direct longitudinal current density in the second half of the laser pulse (JzJ_z for a pulse propagating in z^\hat{z}), which is closely linked to broadband single color filament THz [25, 26, 27, 28] (with an associated low frequency RF tail). Zhou et al [29] noted that transverse current densities (JxJ_x for a linear x^\hat{x} polarized laser pulse, that evolves into JrJ_r with collisions) are also generated by the laser pulse, and they argue that the resulting plasma wake trailing behind the laser pulse will also excite RF radiation.

Stimulated by far field measurements of broadband RF from filamentation [30, 31, 32, 33, 34], several years ago our group developed a quantitative model of RF generation due to the transverse plasma currents [35]. Particle In Cell (PIC) simulations [36, 37] confirmed that a plasma wake is established on the outer boundary of the plasma; the magnitude of the associated JrJ_r current is limited by electron neutral collisions at higher pressures, and by the build up of an electrostatic well at lower pressures (e.g. 1 Torr).

Larger scale continuum Drude simulations [38, 39] then lead to a surprise: the presence of the plasma wake current JrJ_r on the outer boundary of the plasma acts as an antenna and invariably excites a SPP (see also [40]). Furthermore the RF frequencies of this SPP are far below the cutoff at ωs∼1\omega_s \sim1 THz, and they thus travel with a group velocity close to cc. The plasma wake currents, translating and behind the laser pulse, thus co-propagate coherently with and steadily amplify the SPP surface wave at the leading edge of the plasma column. Simulations also show that the surface wave detaches efficiently from the end of the plasma column (end fire radiation [41, 42]) to become the forward directed pulse of RF.

The magnitude and spatial profile of the RF produced by this plasma wake surface wave model was found to accurately match far field broadband horn laboratory measurements [35]. In this work we revisit the theory with near-field D-dot measurements in the vicinity of the filamentation plasma. We find strong evidence of the Sommerfeld-Goubau SPPs predicted by the theory, with an excellent fit to a Hankel function radial profile and the outer length scale. We have also detected a longitudinal frequency drift in the SPPs that closely matches new axisymmetric Drude simulations, and new theoretical derivations for the phase velocity and resistive attenuation length scales. The measurement and characterization of these surface waves thus provides a new window into the physics of filamentation plasmas and for chirped pulse laser science in general.

Theory

The theory developed in [35] was developed through exploratory simulations at a variety of length scales, and then simplified into analytic approximations. Small scale PIC simulations that resolve the 800 nm wavelength and make use of the ionization rate WW given in [16] reproduce the electron velocity distributions given in [29] for linearly and circularly polarized laser pulses (and the large net transverse current for two color systems [43]). In the simplest PIC simulations the electrons are only accelerated by the remainder of the laser pulse, which yields the highest residual velocities for electrons born with instantaneous ∣E∣=0|E| = 0, but one can also include interactions with the parent ion which leads to additional heating (see also [44]).

We then perform larger PIC simulations of transverse slices of this plasma column where the initial electron velocities are set by the previous simulation. With 50 mJ of energy, the 50 fs pulses are in the multifilamentary regime [45], and we approximate the bundle of filament plasmas as one larger plasma column with radius rpl=500 μmr_{\mathrm{pl}} = 500\ \mu\mathrm{m}. At atmospheric pressure the tail of the distribution is at about Ktail=4K_{\mathrm{tail}} = 4 eV, and with an electron-neutral collision frequency ν\nu of 5 THz these diffuse out to roughly rΔ=30 μmr_{\Delta} = 30\ \mu\mathrm{m} over 100 ps, with an effective late time diffusion velocity of veff=7×104v_{\mathrm{eff}} = 7 \times10^{4} m/s. An analytic approximation for the radial current density JrJ_r is given by: Jr=ϵ0veffKtail/rΔ2J_r = \epsilon_0 v_{\mathrm{eff}} K_{\mathrm{tail}} / r_{\Delta}^{2}, which is about 3000 A/m2^{2} at 10 GHz.

The presence of this broadband radial current pulse JrJ_r on the outer boundary of the plasma column was found to excite a broadband SPP in axisymmetric Drude simulations [35]; Figure 1 shows an updated simulation. This surface wave is well approximated by the transverse magnetic Sommerfeld-Goubau solution [46, 47, 48, 49, 50, 51]. At a particular angular frequency ω\omega the radial component of the electric field ErE_r outside of the plasma is given by:

Log plots of the radial electric field for an axisymmetric Drude electromagnetic simulation at three points in time

Figure 1. Log plots of ErE_r for an axisymmetric Drude electromagnetic simulation at 3 points in time, with the SW near z=10z = 10 cm and z=18z = 18 cm at the top and middle, and free RF radiation shortly after detachment from the 25 cm plasma column on the bottom. The plasma column has a central number density of ne=1023 m−3n_e = 10^{23}\ \mathrm{m}^{-3}, and has been given a sin⁡2\sin^2 envelope in the z^\hat{z} direction (inspired by [52]), which causes the SW to both decay and smoothly detach over the final 5 cm with minimal upstream reflection (in contrast to the substantial reflection seen in previous uniform z^\hat{z} density simulations [35]).

Er(r,z,t)=−πrplE02routerei(ωt−hz)H1(1)(γairr),(1)E_r(r,z,t) = -\frac{\pi r_{\mathrm{pl}} E_0}{2r_{\mathrm{outer}}}e^{i(\omega t-hz)}H_1^{(1)}(\gamma_{\mathrm{air}}r), \tag*{(1)}

where E0E_0 is the amplitude of the wave at the surface of the plasma at rplr_{\mathrm{pl}} (about 10410^{4} V/m at atmospheric pressure), the external γair\gamma_{\mathrm{air}} is given by γair2=k02ϵair−h2\gamma_{\mathrm{air}}^{2} = k_0^{2}\epsilon_{\mathrm{air}} - h^{2}, with ϵair=ϵair/ϵ0\epsilon_{\mathrm{air}} = \epsilon_{\mathrm{air}}/\epsilon_0, free space wave number k0=ω/ck_0 = \omega/c, complex wavenumber hh which encodes both the SPP wavelength and attenuation length scale (Ldecay=1/Im⁡(h)L_{\mathrm{decay}} = 1/\operatorname{Im}(h)), and H1(1)H_1^{(1)} is a Hankel function of the first kind with order 1, with outer length scale router=1/∣γair∣r_{\mathrm{outer}} = 1/|\gamma_{\mathrm{air}}|. For radii r^\hat{r} between rplr_{\mathrm{pl}} and routerr_{\mathrm{outer}} the Hankel function is approximately H1(1)≃−2router/(πr)H_1^{(1)} \simeq-2r_{\mathrm{outer}}/(\pi r), and for radii larger than routerr_{\mathrm{outer}} it transitions to falling off exponentially: we fit the experimental D-dot data at a variety of frequency bands and radii to this predicted profile.

The routerr_{\mathrm{outer}} length scale depends on hh, which is found by matching the internal EzE_z and HϕH_\phi fields (having J0J_0 and $J_1

Bessel function dependences respectively) to the external fields (described by H0(1)H_{0}^{(1)} and H1(1)H_{1}^{(1)} Hankel functions) at the plasma column outer radius rplr_{\mathrm{pl}}:

γairH0(1)(γairrpl)εairH1(1)(γairrpl)=γplJ0(γplrpl)εplJ1(γplrpl),(2)\frac{\gamma_{\mathrm{air}}H_{0}^{(1)}(\gamma_{\mathrm{air}}r_{\mathrm{pl}})}{\varepsilon_{\mathrm{air}}H_{1}^{(1)}(\gamma_{\mathrm{air}}r_{\mathrm{pl}})} = \frac{\gamma_{\mathrm{pl}}J_{0}(\gamma_{\mathrm{pl}}r_{\mathrm{pl}})}{\varepsilon_{\mathrm{pl}}J_{1}(\gamma_{\mathrm{pl}}r_{\mathrm{pl}})}, \tag*{(2)}

with γpl2=k02εpl−h2\gamma_{\mathrm{pl}}^{2}=k_{0}^{2}\varepsilon_{\mathrm{pl}}-h^{2} and relative plasma permittivity εpl\varepsilon_{\mathrm{pl}} [49] (and setting εair=1\varepsilon_{\mathrm{air}}=1 from here on).

Sommerfeld originally solved (2) by using asymptotic approximations for the Bessel and Hankel functions and then developing an iterative method [46] to find hh, and recent work has shown that those approximations can also yield a direct solution via the Lambert WW function [53, 54]. We used this direct Lambert WW method to estimate hh in [35], but to get more accurate values for this work we have dropped the approximations and switched to an iterative method that uses the full Bessel and Hankel functions (see [49]). In particular, we include both the low and high frequency terms in the plasma permittivity:

εpl(ω)=1+ωpl2iων−ω2,(3)\varepsilon_{\mathrm{pl}}(\omega)=1+\frac{\omega_{\mathrm{pl}}^{2}}{i\omega\nu-\omega^{2}}, \tag*{(3)}

since ωε0\omega\varepsilon_{0} is not too much smaller than σDC=ε0ωpl2/ν\sigma_{\mathrm{DC}}=\varepsilon_{0}\omega_{\mathrm{pl}}^{2}/\nu. At lower frequencies the skin depth 2/σDCμω\sqrt{2/\sigma_{\mathrm{DC}}\mu\omega} can be comparable to the plasma radius rplr_{\mathrm{pl}}, and so we also do not use the J0/J1=iJ_{0}/J_{1}=i approximation. We thus rearrange the full (2) equation to give the N+1N+1 iteration for γair\gamma_{\mathrm{air}}:

γairN+1=γplNH1(1)(γairNrpl)J0(γplNrpl)εplH0(1)(γairNrpl)J1(γplNrpl),(4)\gamma_{\mathrm{air}}^{N+1}=\frac{\gamma_{\mathrm{pl}}^{N}H_{1}^{(1)}(\gamma_{\mathrm{air}}^{N}r_{\mathrm{pl}})J_{0}(\gamma_{\mathrm{pl}}^{N}r_{\mathrm{pl}})}{\varepsilon_{\mathrm{pl}}H_{0}^{(1)}(\gamma_{\mathrm{air}}^{N}r_{\mathrm{pl}})J_{1}(\gamma_{\mathrm{pl}}^{N}r_{\mathrm{pl}})}, \tag*{(4)}

with γplN=(γairN)2+(εair(ω)−1)k02\gamma_{\mathrm{pl}}^{N}=\sqrt{(\gamma_{\mathrm{air}}^{N})^{2}+(\varepsilon_{\mathrm{air}}(\omega)-1)k_{0}^{2}}, an initial N=1N=1 guess of γairN=1=0.5k0\gamma_{\mathrm{air}}^{N=1}=0.5k_{0}, and h=k02−γair2h=\sqrt{k_{0}^{2}-\gamma_{\mathrm{air}}^{2}} once the iteration has converged (typically 10-30 steps - see [54] for further discussion on convergence criteria). For a plasma with an estimated values rpl=500 μmr_{\mathrm{pl}}=500\ \mu\mathrm{m}, ν∼5 THz\nu\sim5\ \mathrm{THz}, ne∼3×1022 m−3n_{e}\sim3\times10^{22}\ \mathrm{m}^{-3} (giving σDC∼170 S/m\sigma_{\mathrm{DC}}\sim170\ \mathrm{S/m}), and a surface wave frequency of 10 GHz, (4) yields h≃214+21ih\simeq214+21i, which gives router∼1.0 cmr_{\mathrm{outer}}\sim1.0\ \mathrm{cm}, and a longitudinal 1/e1/e attenuation length scale of Ldecay∼4.7 cmL_{\mathrm{decay}}\sim4.7\ \mathrm{cm}. We note that with longer laser pulses the electron density nen_{e} can grow much higher, especially as collisional heating and ionization become important [55], which leads to higher conductivity, a larger plasma radius and better surface wave propagation and growth.

We base our theory on the Sommerfeld-Goubau solution (1), but additional dynamics occur in the surface wave simulations (see Figure 1) that are not easily captured by analytic approximations. The analytic solutions assume a Heaviside step function profile in the conductor electron density, which is a good approximation for optical SPPs on metal substrates, but less so for the boundaries of plasmas. However, in simulation when we switch from a step function to a radial electron density gradient at the outer boundary of the plasma we get qualitatively similar SPPs, with the amplitude E0E_{0} of these gradient-boundary waves somewhat suppressed, and the spatial wave length moderately increased.

We also observe that the plasma wake radial current directly excites some free radiation in addition to the surface wave: this lower amplitude free radiation can be seen to slowly pull ahead of the surface wave over long simulations (and can be seen in Figure 1 given the log scale). We suspect that the D-dot probe is also detecting this prompt free radiation in the experiments.

The original simulations performed in [35] used plasma columns with constant density in the longitudinal z^\hat{z} direction, but given the experimental measurements in [52] we have switched to a more realistic tapered ne(z)=ne,max⁡sin⁡2(πz/Lpl)n_{e}(z)=n_{e,\max}\sin^{2}(\pi z/L_{\mathrm{pl}}) profile, with low plasma densities at the beginning and end of the column. The use of this tapered profile results in modulation of the SPP amplitude that closely mirrors the near field experimental data. It also helps to explain the absence of higher frequency far field RF in experiments such as [31] (given the presence of high amplitude, high frequency content in JrJ_{r}), as higher frequency SPPs are more strongly attenuated in low density regions (see Figure 2). We also find a frequency dependent drift in the simulated SPP peak amplitudes that closely matches the longitudinal D-dot measurements. We note that the dispersion of SPPs in a cylindrical geometry is more complex than on a plane, with the group velocity of low frequency waves dropping rapidly instead of converging to cc [50, 51], as can be seen in Figure 2. On a filamentation plasma where Kerr focusing and plasma defocusing can compete to form islands of high density plasma this can cause the lowest frequency SPPs to become trapped on the high density islands.

Plots of the phase velocity, decay length, and outer length scale as a function of SPP frequency

Figure 2. Plots of the phase velocity, decay length, and outer length scale as a function of SPP frequency as calculated by Eq. (4), for a plasma column with radius rpl=0.5r_{\mathrm{pl}} = 0.5 mm, collision frequency ν=5\nu= 5 THz, and density nen_e ranging from 102210^{22} to 1023 m−310^{23}\ \mathrm{m}^{-3}. The phase velocity slowly climbs toward cc as the frequency decreases (as for planar SPPs), and then plunges when the skin depth becomes comparable to rplr_{\mathrm{pl}}. LdecayL_{\mathrm{decay}} and routerr_{\mathrm{outer}} both increase for lower frequencies, with LdecayL_{\mathrm{decay}} falling rapidly with decreasing nen_e, and routerr_{\mathrm{outer}} more slowly.

Experimentation

To investigate the near-field region of a femtosecond plasma, an 800 nm wavelength 50 mJ laser was used with a 50 fs pulse duration and a 10 Hz repetition rate. Plasma formed near the geometric focal point of a 3 m focusing optic (leading to somewhat higher nen_{e}). Figure 3 shows the visible fluorescence of the plasma column which spans approximately 18 cm across the top of an optical bench.

Photograph of the laser plasma column and D-dot probe

Figure 3. The 800nm laser plasma column was generated with the aid of lens assisted filamentation to conserve laboratory space. The D-dot probe is located on an r^\hat{r} direction translation stage and is swept in the z^\hat{z} direction using an optical rail.

A Prodyn FD-5C2 D-dot probe (bandwidth of 2-40 GHz) was used to measure the electric field generated by the plasma column. From preliminary experiments, the strongest signal was received by the D-dot probe when placing the probe tip normal to the direction of plasma propagation (see Figure 3). The probe was mounted to a 2 inch micrometer stage (movable in the r^\hat{r} direction) in the same plane as the plasma column. The stage was attached to two consecutive 33 inch optical rails mounted along the direction of propagation. This allowed for the probe to be positioned both along the z^\hat{z} direction of the plasma and at various radius locations in the r^\hat{r} direction. By mounting the probe in such a configuration, the plasma column was able to be examined in increments of 1/16 inch in the z^\hat{z} direction and 0.001 mm in the r^\hat{r} direction. The laser was carefully aligned through two irises to ensure the D-dot probe would remain at a fixed radius from the plasma column while moving along the zz axis. An index card flush with the D-dot was periodically ablated by the plasma column at various z^\hat{z} distances to confirm the distance between the probe tip and the plasma column remained constant. The experimental setup is depicted by the rendering in Figure 4.

Rendering of the experimental setup

Figure 4. The above rendering shows the components utilized during experimentation. The Prodyn FD-5C2 D-dot probe is mounted atop a 2 inch micrometer stage which moves along the r^\hat{r} axis. Additionally, the stage is mounted to a 72.5 inch optical rail; this allows for movement along the z^\hat{z} axis. The plasma column propagates along the positive z^\hat{z} axis as shown above. K-type coaxial cable was used to connect to the oscilloscope, and microwave foam was placed parallel to the plasma to reduce signal reflections.

Data was collected by selecting a stationary radius rr and longitudinally scanning in the z^\hat{z} direction with the D-dot probe; rr values of 0.5, 0.75, 1.0, 1.5, and 2.0 cm were

probed at 28 locations along the z^\hat{z} axis ranging from 248.9 cm to 320.0 cm from the focusing optic. Steps along the z^\hat{z} axis were taken in 2.54 cm increments. Additionally, radial data was collected at the fixed longitude location of highest signal (z=294.6z = 294.6 cm) for 19 evenly spaced radii between 0.25 cm and 5.0 cm from the plasma surface (note that we later shift our zz coordinate system to begin with of plasma formation, so that z=295z = 295 cm becomes z=18z = 18 cm). For both experiments laser output energy was measured at the beginning and end of each test. The signal received by the D-dot probe was digitally captured using a Tektronix DP077002SX 33 GHz oscilloscope which limited the upper frequency limit of the measurements. The D-dot probe was connected to the oscilloscope using a K-type (40 GHz) cable.

Results

After collecting the raw data in the Tektronics digital oscilloscope, a Fast Fourier Transform (FFT) was applied to the time dependent signals to obtain the frequency content within the 2-33 GHz microwave regime. The noise floor from the laboratory environment was also recorded, and the resulting noise FFT subtracted from the signal FFTs. We note that developing an accurate, frequency dependent calibration of the D-dot probe (see e.g. [31]) is non-trivial, and we leave it for future work. As the D-dot amplitudes are uncalibrated, the measured values of the electric fields at different frequencies should only be considered in a relative context.

Figure 5 shows the FFTs taken at 4 radial separations ranging from r=0.5r = 0.5 cm to r=1.5r = 1.5 cm, taken near the peak at z=18z = 18 cm, as well as the noise signal recorded when the laser pulse is blocked at the focusing lens and no plasma forms (note that each curve is the average of 50 individual shots). Although the data is uncalibrated the ultra-broadband nature of the near field waves is clear, (as can also be seen in the original time series data), and the overall profile resembles the far field RF recorded in [31].

FFTs of measured electron field strength and an example time-domain signal

Figure 5. (a) FFTs of measured electron field strength measured at z=18z = 18 cm (near the peak) for radial separations ranging from 0.5 cm to 1.5 cm, along with the measured noise spectrum taken when no filamentation is occurring. The near field D-dot probe has not been calibrated for this experiment, but the broadband signal resembles the far field data taken in [31]. In all frequency bins the amplitude of the signal above the noise floor falls off rapidly with increasing distance. (b) An example time domain signal recorded by the oscilloscope at a separation of 0.5 cm.

We next average the radial FFT data over large frequency bins so that we can compare with theory. In general the radial fall off is very well fit by the expected Hankel function dependence in (1), and is much better than naive polynomial fits. Figure 6 shows the single best fit that occurs at 14 GHz (which is also about the frequency of largest signal - see Figure 5), with an R2R^2 value of 0.9993. All of the frequency bins are best fit by the Hankel function dependence, with typical R2R^2 values of about 0.99. The radial dependence of simulated SWs (see Figure 1) is also very well fit by a Hankel function, with R2R^2 values better than 0.9999.

Radial signal decay and outer length scale as a function of frequency

Figure 6. (a) Log scale plot of the 14 GHz bin data (collected at z=18z = 18 cm) for radii between r=0.25r = 0.25 cm and 3.75 cm. The radial data in all frequency bins is well fit by the expected Hankel function dependence (1), with typical R2R^2 values of 0.99, and the 14 GHz bin yields the best fit, with R2=0.9993R^2 = 0.9993. Visually the Hankel fit is very good out to a distance of about 2 cm, much better than the naive 1/r1/r, 1/r21/r^2, and 1/r31/r^3 profiles, and then it slightly underpredicts the measured signal at larger radii. We suspect that this is due to the presence of free radiation, in addition to the bound SPP wave, which the simulations also generically predict to be excited (see Figure 1). (b) Comparison of the routerr_{\mathrm{outer}} values from Hankel function fits as a function of frequency, compared to the analytic expectation from (4) with an electron density of ne=3×1022 m−3n_e = 3 \times10^{22}\ \mathrm{m}^{-3}. The overall magnitude is quite close, although the slope of the experimental data is flatter than the theoretical expectation.

We note that the Hankel function fit underpredicts the measured data starting at about 2.5 cm in Figure 6, and this trend is seen in the other frequency bins. In simulation some free radiation is also excited by the radial JrJ_r currents, and the amplitude of this radiation will fall off more slowly than the Hankel function at large radius (r>routerr > r_{\mathrm{outer}}). We thus suspect that we are also detecting this free radiation, although the overall amplitude of the measured signal is small at this larger radius.

Given the Hankel function fits we can also extract the fit values for the outer length scale routerr_{\mathrm{outer}} and compare with theory. We find that the experimental routerr_{\mathrm{outer}} values match well with an electron density of ne=3×1022 m−3n_e = 3 \times10^{22}\ \mathrm{m}^{-3} (see Figure 2), with an experimental fit router=0.75r_{\mathrm{outer}} = 0.75 cm at 15 GHz being close to the theory value given by (4). However the slope of the experimental values as a function of frequency is lower than the theoretical curve, as can be seen in Figure 6.

We next consider the longitudinal variation in the measured D-dot signal. Figure 7 shows the amplitude as a function of zz for 5 frequency bins ranging from 5 GHz to 25 GHz, as measured at r=2.5r = 2.5 mm in both the experimental and Drude simulation data (see Figure 1). As a D-dot calibration is not available the different frequencies have all been normalized to 1 for comparison purposes. We

Normalized traces of measured electric field strength in five different frequency bins for experimental data and simulation

Figure 7. Normalized traces of measured electric field strength in 5 different frequency bins as a function of longitudinal position zz, for both the experimental data and the corresponding simulation shown in Figure 1. In both cases the higher frequency components of the wave reach their maximums at earlier z^\hat{z} positions, and the lowest frequencies reach their maximums last. The experimental data also becomes successively more noisy at lower frequencies, which we suspect is due to a combination of phase velocity drop off at low frequencies, and focusing and defocusing effects modulating the plasma column density.

note that there is a consistent drift of the peaks of the different frequencies to larger zz as the frequency decreases in both the experimental data and the simulations.

This longitudinal drift towards lower frequencies is consistent with the theoretical expectations shown in Figure 2: from higher frequencies down to about 5 GHz the phase velocity of the SPPs is approaching cc, and thus the lower frequency surface waves have better phase coherence with the JrJ_r current source. The lower frequency waves also decay more slowly, helping them to build to a later peak. At even lower frequencies below 5 GHz the phase velocity rapidly drops (depending sensitively on the electron density), and the waves no longer propagate efficiently with JrJ_r, and the amplitude of the SPPs falls off quickly. We note that the experimental curves become much noisier below 5 GHz, which we suspect is due to density fluctuations along the plasma column due to focusing and de-focusing effects, and which the lower frequencies are more sensitive to. At the larger zz values towards the end of the plasma column the electron density nen_e drops off, which causes the waves to both decay more quickly and detach from the plasma column, leading to a collapse of the signal.

Inspired by exploratory simulations we argued [35] that SPPs are both generated during femtosecond filamentation, and grow in strength due to coherent copropagation near the speed of light with the plasma wake currents that excite them along the leading edge of the plasma column. In this paper we present new near field data that is in excellent agreement with this theory. We find that the radial profile of the D-dot measurements is in very good agreement with the expected Hankel function dependence from the Sommerfeld-Goubau solution, and the characteristic outer length scale is well matched by a plasma with density ne=3×1022 m−3n_e = 3 \times10^{22}\ \mathrm{m}^{-3}. In turn the longitudinal frequency drift data seen in the data is in good agreement with both theoretical expectations for the phase velocity and decay length (4), and with new simulations (Figure 1) that include variations in plasma density [52]. The near field measurement of these surface waves thus provides a powerful new scientific tool for chirped pulse filamentation research.

Acknowledgments

The authors thank the Air Force Office of Scientific Research (AFOSR) for support via Laboratory Task No. FA9550-24RDCOR002. This work was supported in part by high-performance computer time and resources from the DoD High Performance Computing Modernization Program. Approved for public release; distribution is unlimited. Public Affairs release approval No. AFRL-2024-6758.

References

  1. [1]D Kordeweg and G de Vries, “On the change of form of long waves advancing in a rectangular channel, and a new type of long stationary wave,” Phil. Mag 39, 422–443 (1895).
  2. [2]Norman J Zabusky and Martin D Kruskal, “Interaction of “solitons” in a collisionless plasma and the recurrence of initial states,” Physical review letters 15, 240 (1965).DOI
  3. [3]Mark J Ablowitz and Harvey Segur, “On the evolution of packets of water waves,” Journal of Fluid Mechanics 92, 691–715 (1979).DOI
  4. [4]Lord Rayleigh, “On waves propagated along the plane surface of an elastic solid,” Proceedings of the London mathematical Society 1, 4–11 (1885).DOI
  5. [5]Augustus Edward Hough Love, Some Problems of Geodynamics: Being an Essay to which the Adams Prize in the University of Cambridge was Adjudged in 1911 (University Press, 1911).
  6. [6]Per Delsing, Andrew N Cleland, Martin JA Schuetz, Johannes Knörzer, Géza Giedke, J Ignacio Cirac, Kartik Srinivasan, Marco Wu, Krishna Coimbatore Balaram, Christopher Bäuerle, et al., “The 2019 surface acoustic waves roadmap,” Journal of Physics D: Applied Physics 52, 353001 (2019).
  7. [7]Michel I D’yakonov, “New type of electromagnetic wave propagating at an interface,” Sov. Phys. JETP 67, 714–716 (1988).
  8. [8]Pochi Yeh, Amnon Yariv, and Chi-Shain Hong, “Electromagnetic propagation in periodic stratified media. i. general theory,” JOSA 67, 423–438 (1977).DOI
  9. [9]John A Polo Jr and Akhlesh Lakhtakia, “Surface electromagnetic waves: a review,” Laser & Photonics Reviews 5, 234–246 (2011).
  10. [10]Osamu Takayama, AA Bogdanov, and Andrei V Lavrinenko, “Photonic surface waves on metamaterial interfaces,” Journal of Physics: Condensed Matter 29, 463001 (2017).DOI
  11. [11]Saman Jahani and Zubin Jacob, “All-dielectric metamaterials,” Nature nanotechnology 11, 23–36 (2016).DOI
  12. [12]Rufus H Ritchie, “Plasma losses by fast electrons in thin films,” Physical review 106, 874 (1957).DOI
  13. [13]JM Pitarke, VM Silkin, EV Chulkov, and PM Echenique, “Theory of surface plasmons and surface-plasmon polaritons,” Reports on progress in physics 70, 1 (2006).
  14. [14]Andreas Otto, “Excitation of nonradiative surface plasma waves in silver by the method of frustrated total reflection,” Zeitschrift für Physik A Hadrons and nuclei 216, 398–410 (1968).DOI
  15. [15]Erwin Kretschmann and Heinz Raether, “Radiative decay of non radiative surface plasmons excited by light,” Zeitschrift für Naturforschung A 23, 2135–2136 (1968).DOI
  16. [16]SV Popruzhenko, VD Mur, VS Popov, and D Bauer, “Strong field ionization rate for arbitrary laser frequencies,” Physical review letters 101, 193003 (2008).DOI
  17. [17]A Braun, G Korn, X Liu, D Du, J Squier, and G Mourou, “Self-channeling of high-peak-power femtosecond laser pulses in air,” .
  18. [18]Francis Théberge, Weiwei Liu, Patrick Tr Simard, Andreas Becker, and See Leang Chin, “Plasma density inside a femtosecond laser filament in air: Strong dependence on external focusing,” Physical Review E 74, 036406 (2006).pubmed.ncbi.nlm.nih.gov/17025753
  19. [19]Arnaud Couairon and André Mysyrowicz, “Femtosecond filamentation in transparent media,” Physics reports 441, 47–189 (2007).DOI
  20. [20]Jérôme Kasparian and Jean-Pierre Wolf, “Physics and applications of atmospheric nonlinear optics and filamentation,” Optics express 16, 466–493 (2008).pubmed.ncbi.nlm.nih.gov/18521179
  21. [21]SL Chin, T J Wang, C Marceau, J Wu, JS Liu, O Kosareva, N Panov, YP Chen, J F Daigle, S Yuan, et al., “Advances in intense femtosecond laser filamentation in air,” Laser Physics 22, 1–53 (2012).DOI
  22. [22]Zhaxylk A Kudyshev, Martin C Richardson, and Natalia M Litchinitser, “Virtual hyperbolic metamaterials for manipulating radar signals in air,” Nature communications 4, 2557 (2013).DOI
  23. [23]Shermineh Rostami, Michael Chini, Khan Lim, John P Palastro, Magali Durand, Jean-Claude Diels, Ladan Arissian, Matthieu Baudelet, and Martin Richardson, “Dramatic enhancement of supercontinuum generation in elliptically-polarized laser filaments,” Scientific Reports 6, 20363 (2016).DOI
  24. [24]J Kasparian and J-P Wolf, “Ultrafast laser spectroscopy and control of atmospheric aerosols,” Physical Chemistry Chemical Physics 14, 9291–9300 (2012).DOI
  25. [25]Chung-Chieh Cheng, Ewan M Wright, and Jerome V Moloney, “Generation of electromagnetic pulses from plasma channels induced by femtosecond light strings,” Physical Review Letters 87, 213001 (2001).
  26. [26]Phillip Sprangle, JR Penano, Bahman Hafizi, and CA Kapetanakos, “Ultrashort laser pulses and electromagnetic pulse generation in air and on dielectric surfaces,” Physical Review E 69, 066415 (2004).
  27. [27]Ciro D’Amico, Aurélien Houard, Michel Franco, Bernard Prade, André Mysyrowicz, Arnaud Couairon, and VT Tikhonchuk, “Conical forward thz emission from femtosecond-laser-beam filamentation in air,” Physical review letters 98, 235002 (2007).
  28. [28]CD Amico, Aurélien Houard, Selcuk Akturk, Yi Liu, J Le Bloas, Michel Franco, Bernard Prade, Arnaud Couairon, VT Tikhonchuk, and André Mysyrowicz, “Forward thz radiation emission by femtosecond filamentation in gases: theory and experiment,” New Journal of Physics 10, 013015 (2008).DOI
  29. [29]Bing Zhou, Aurélien Houard, Yi Liu, Bernard Prade, André Mysyrowicz, Arnaud Couairon, Patrick Mora, Christopher Smeenk, Ladan Arissian, and Paul Corkum, “Measurement and control of plasma oscillations in femtosecond filaments,” Physical review letters 106, 255002 (2011).pubmed.ncbi.nlm.nih.gov/21770648
  30. [30]Benjamin Forestier, Aurélien Houard, Magali Durand, Yves-Bernard André, Bernard Prade, J-Y Dauvignac, Franck Perret, Ch Pichot, Michel Pellet, and André Mysyrowicz, “Radiofrequency conical emission from femtosecond filaments in air,” Applied Physics Letters 96, 141111 (2010).DOI
  31. [31]Alexander Englesbe, Jennifer Elle, Remington Reid, Adrian Lucero, Hugh Pohle, Matthew Domonkos, Serge Kalmykov, Karl Krushelnick, and Andreas Schmitt-Sody, “Gas pressure dependence of microwave pulses generated by laser-produced filament plasma,” Optics Letters 43, 4953–4956 (2018).DOI
  32. [32]A Janicek, E Thornton, T Garrett, A Englesbe, J Elle, and A Schmitt-Sody, “Length dependence on broadband microwave emission from laser-generated plasmas,” IEEE Transactions on Plasma Science 48, 1979–1983 (2020).DOI
  33. [33]AV Mitrofanov, DA Sidorov-Biryukov, MM Nazarov, AA Voronin, MV Rozhko, AB Fedotov, and AM Zheltikov, “Coherently enhanced microwave pulses from midinfrared-driven laser plasmas,” Optics Letters 46, 1081–1084 (2021).DOI
  34. [34]Alexander Englesbe, Jennifer Elle, Robert Schwartz, Travis Garrett, Daniel Woodbury, Dogeun Jang, Ki-Yong Kim, Howard Milchberg, Remington Reid, Adrian Lucero, et al., “Ultrabroadband microwave radiation from near-and mid-infrared laser-produced plasmas in air,” Physical Review A 104, 013107 (2021).DOI
  35. [35]Travis Garrett, Jennifer Elle, Michael White, Remington Reid, Alexander Englesbe, Ryan Phillips, Peter Mardahl, Erin Thornton, James Wymer, Anna Janicek, et al., “Generation of radio frequency radiation by femtosecond filaments,” Physical Review E 104, L063201 (2021).DOI
  36. [36]Charles K Birdsall and A Bruce Langdon, Plasma physics via computer simulation (CRC press, 2004).
  37. [37]Robert E Peterkin and John W Luginsland, “A virtual prototyping environment for directed-energy concepts,” Computing in Science & Engineering 4, 42–49 (2002).DOI
  38. [38]M Okoniewski, M Mrozowski, and MA Kutchy, “Simple treatment of multi-term dispersion in fdtd,” IEEE Microwave and Guided Wave Letters 7, 121–123 (1997).
  39. [39]Fernando L Teixeira, “Time-domain finite-difference and finite-element methods for maxwell equations in complex media,” IEEE Transactions on Antennas and Propagation 56, 2150–2166 (2008).DOI
  40. [40]Jiang Chen, Michaela Badioli, Pablo Alonso-González, Sukosin Thongrattanasiri, Florian Huth, Johann Ostmond, Marko Spasenović, Alba Centeno, Amaia Pesquera, Philippe Godignon, et al., “Optical nano-imaging of gate-tunable graphene plasmons,” Nature 487, 77–81 (2012).DOI
  41. [41]JB Andersen, “Radiation from surface-wave antennas,” Electronics Letters 3, 251–252 (1967).DOI
  42. [42]GI Stegeman, RF Wallis, and AA Maradudin, “Excitation of surface polaritons by end-fire coupling,” Optics letters 8, 386–388 (1983).pubmed.ncbi.nlm.nih.gov/19718123
  43. [43]Ki-Yong Kim, James H Glownia, Antoinette J Taylor, and George Rodriguez, “Terahertz emission from ultrafast ionizing air in symmetry-broken laser fields,” Optics express 15, 4577–4584 (2007).pubmed.ncbi.nlm.nih.gov/19532704
  44. [44]Paul B Corkum, “Plasma perspective on strong field multiphoton ionization,” Physical review letters 71, 1994 (1993).pubmed.ncbi.nlm.nih.gov/10054556
  45. [45]L Bergé, S Skupin, F Lederer, G Méjean, J Yu, J Kasparian, E Salmon, JP Wolf, M Rodriguez, L Wöste, et al., “Multiple filamentation of terawatt laser pulses in air,” Physical Review Letters 92, 225002 (2004).pubmed.ncbi.nlm.nih.gov/15245231
  46. [46]Arnold Sommerfeld, “Ueber die fortpflanzung elektrodynamischer wellen längs eines drahtes,” Annalen der Physik 303, 233–290 (1899).DOI
  47. [47]Georg Goubau, “Surface waves and their application to transmission lines,” Journal of Applied Physics 21, 1119–1128 (1950).DOI
  48. [48]Julius Adams Stratton, Electromagnetic theory, Vol. 33 (John Wiley & Sons, 2007).
  49. [49]Sophocles J Orfanidis, “Electromagnetic waves and antennas,” (2002).
  50. [50]CA Pfeiffer, EN Economou, and KL Ngai, “Surface polaritons in a circularly cylindrical interface: surface plasmons,” Physical review B 10, 3038 (1974).DOI
  51. [51]Kanglin Wang and Daniel M Mittleman, “Dispersion of surface plasmon polaritons on metal wires in the terahertz frequency range,” Physical Review Letters 96, 157401 (2006).pubmed.ncbi.nlm.nih.gov/16712193
  52. [52]Anna M Janicek, “Measuring energy deposition from a laser induced plasma in air which generates broadband microwave radiation,” (2023).
  53. [53]Denis Jaisson, “Simple formula for the wave number of the goubau line,” Electromagnetics 34, 85–91 (2014).DOI
  54. [54]J Ricardo G Mendonça, “Electromagnetic surface wave propagation in a metallic wire and the lambert w function,” American Journal of Physics 87, 476–484 (2019).DOI
  55. [55]Erin Thornton, Travis Garrett, and Jennifer Elle, “Boosting radio frequency radiation with collisional processes in picosecond laser filamentation,” Physics of Plasmas 31, 050702 (2024).DOI

Paper details

vlnc:W_db9qknwe
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents