betaFeedback

WorksPhysics and Astronomy

Terahertz light driven coherent excitation of a zone-folded Raman-active phonon mode in the Spin-Ladder System α′α'-NaV2_2O5_5

Abstract

We investigate the out-of-equilibrium lattice dynamics in the spin-ladder system α′α'-NaV2_2O5_5 using intense terahertz (THz) pump and near-infrared (NIR) probe spectroscopy. When quasi-single-cycle THz pulses interact with α′α'-NaV2_2O5_5 in its low-temperature, dimerized charge-ordered phase, they induce coherent oscillations in the time domain at the zone-folded Raman-active phonon frequency of 1.85 THz. By combining pump-probe measurements with lattice dynamics modeling based on equation-of-motion approach, we propose that these oscillations arise from a nonlinear coupling between Raman-active and infrared (IR)-active phonon modes, with the latter being resonantly excited by the THz pulses. In contrast, excitation with NIR femtosecond laser pulses does not produce measurable vibrational dynamics, highlighting the unique potential of THz-driven, nonlinear light-matter interactions for the coherent and selective control of structural dynamics in quantum materials.

Introduction

Studies of ultrafast phenomena in complex materials have been revolutionised by recent advances in coherent light sources that have enabled the generation of high-intensity, ultrashort light pulses at specified THz and mid-infrared (MIR) frequencies [1, 2, 3]. These sources have proven to be highly effective in the dynamic manipulation of electronic, magnetic, and structural properties on femtosecond and picosecond timescales [4, 5, 6]. Driving materials far from equilibrium with tailored THz and MIR fields, has revealed – both experimentally and theoretically – a rich landscape of metastable and nonlinear phenomena, transient quantum phases [7, 8, 9, 10, 11, 12], and novel pathways for the selective excitation of collective modes such as lattice vibrations [13, 14, 15, 16, 17], spin waves [18, 19, 20] and collective modes in superconductors [21, 22, 23].

THz and MIR pulses, are uniquely suited for driving low-energy collective excitations that are otherwise inaccessible via direct light-matter coupling. Through nonlinear coupling mechanisms, such as phonon-phonon, electron-phonon, and spin-phonon interactions, these pulses can be used to selectively manipulate individual degrees of freedom without introducing significant heating in the material [4, 5, 6]. This selective excitation permits precise control of strongly correlated materials, where the interplay among charge, spin, and lattice dynamics gives rise to macroscopic quantum phenomena, including multiferroicity, unconventional superconductivity, and charge or spin ordering [24].

Here, we investigate the ultrafast dynamics of Raman-active zone-folded phonon modes in the prototypical spin-ladder system α′\alpha'-NaV2_2O5_5, which is driven by nonlinear THz light-matter interaction. An intense THz pulse induces coherent oscillations at the frequency of a Raman-active zone-folded phonon mode (ΩR=1.85\Omega_{\mathrm{R}} = 1.85 THz). Temperature-dependent measurements show that these oscillations vanish during the structural phase transition, confirming their origin as zone-folded lattice vibrations associated with the low-temperature, charge-ordered dimerized phase. The quadratic dependence of the oscillation amplitude on the THz field strength also signals a nonlinear excitation mechanism. As the coherent dynamics are only evident when the spectral components of the THz pulse overlap with IR-active phonon frequencies, it can be concluded that the excitation mechanism is primarily governed by nonlinear lattice coupling between IR-active phonons and the Raman-active phonon. More specifically, interpretation of the results with phonon dynamics modeling based on coupled equations of motion suggests that the nonlinear excitation of the Raman-active zone-folded phonon mode is due to driving IR-active phonons with intense THz light as opposed to purely photonic nonlinear pathways such as frequency mixing of the THz light components. Importantly, femtosecond NIR laser excitation fails to induce any measurable coherent vibrational dynamics, highlighting the unique capability of THz-driven nonlinear light-matter interactions to selectively excite Raman-active phonons inaccessible through direct optical coupling.

The Experiment

The spin-ladder system α′\alpha'-NaV2_2O5_5 experiences an intriguing phase transition characterized by charge ordering, spin pairing, and lattice distortion at the critical temperature Tc∼34T_c \sim34 K [27]. Above TcT_c, the vanadium (V) ions form ladder structures in the aa-bb plane and have an average +4.5+4.5 charge. Below TcT_c, a charge ordering transition occurs, leading to an alternating arrangement of V4+\mathrm{V}^{4+} and V5+\mathrm{V}^{5+} ions along the bb axis, as depicted in Figure 1(a). This charge order phase leads neighboring V4+\mathrm{V}^{4+} ions to form spin dimers in a spin-singlet state, causing the opening of a spin gap in the magnetic spectrum [27, 28].

Schematic representation of the spin-ladder system, experimental setup, THz electric field and induced polarization changes, and Fourier spectra

Figure 1. (a) Schematic representation of the spin-ladder system formed by the Vanadium (V) and oxygen atoms (O). Below Tc=34 KT_c = 34\,\mathrm{K}, the system is in a charge-ordered phase, with in-equivalent V4+\mathrm{V}^{4+} and V5+\mathrm{V}^{5+}. (b) Experimental setup. An intense THz pulse (yellow) with the electric field polarized linearly along the spin-ladder (b-axis) of α′\alpha^\prime-NaV2_2O5_5 excites the crystal and induces a polarization change on a femtosecond NIR probe pulse (red). (c-d) THz electric field, measured by electro-optic sampling (EOS), and induced polarization changes of the probe as a function of the time delay tppt_{pp}. (e) Fourier amplitude spectrum of the THz pump waveform (yellow solid line) and real part of the optical conductivity σ1\sigma_1 along the b-axis of α′\alpha^\prime-NaV2_2O5_5 at 4.4 K, taken from [25]. (f) Fourier transform of the temporal derivative of ΔR/R\Delta R/R shown in panel (d). Black arrows indicate the frequencies of the Raman-active modes from Ref. [26].

Simultaneously with the formation of the charge-ordered phase, a lattice superstructure emerges due to lattice distortion accompanied by a spin dimerization [29, 30]. This new lattice configuration, in turn, leads to the appearance of spectrally narrow IR- and Raman-active phonon modes in the THz range due to Brillouin-zone folding [29, 30]. Raman spectroscopy has identified a low-frequency zone-folded phonon mode at ΩR=66 cm−1\Omega_R = 66\,\mathrm{cm}^{-1} (1.85 THz), emerging as the primary contribution to the Raman scattering cross-section in the aa-bb plane [26, 31, 32, 33]. In contrast, two additional zone-folded phonon modes at 107 cm−1\mathrm{cm}^{-1} (3.20 THz) and 134 cm−1\mathrm{cm}^{-1} (4.02 THz) exhibit a comparatively lower Raman scattering cross-section. Initially assigned to a magnetic excitation due to the comparable energy of the spin-gap [26, 33], further investigations have revealed the vibrational nature of the mode at ΩR\Omega_R through temperature and doping-dependent analyses employing Raman, IR, and ESR spectroscopies [25, 29, 30, 34]. However, it is important to acknowledge that the potential involvement of spin interactions in the vibrational dynamics of this mode cannot be categorically excluded a priori. Additionally, Raman studies, coupled with theoretical considerations, suggest that charge order fluctuations, modulating the spin super-exchange interactions, can result in a strong spin-phonon coupling below TcT_c [35]. This spin-phonon coupling has been suggested to significantly enhance vibrational anharmonicity [35], making α′\alpha^\prime-NaV2_2O5_5 a compelling system for investigating nonlinear THz-driven out-of-equilibrium dynamics and exploring potential nonlinear phononic mechanisms.

Infrared spectroscopy in the THz range, polarized along the bb axis, has revealed the emergence of IR-active, zone-folded phonon modes below TcT_c [25, 29]. These modes appear at ΩIR,1=101 cm−1\Omega_{\mathrm{IR},1} = 101\,\mathrm{cm}^{-1} (3.03 THz), ΩIR,2=127 cm−1\Omega_{\mathrm{IR},2} = 127\,\mathrm{cm}^{-1} (3.8 THz), and are associated with the formation of the lattice superstructure [25, 29]. Additionally, a phonon mode is observed at ΩIR,3=180 cm−1\Omega_{\mathrm{IR},3} = 180\,\mathrm{cm}^{-1} (5.4 THz) [25, 29]. Other phonons have a weaker IR absorption cross-section, including the low-frequency phonon cluster at around 35 cm−1\mathrm{cm}^{-1} (∼\sim 1 THz) [25]. All the zone-folded phonon modes rapidly disappeared above TcT_c.

For our experiment, high-quality single crystals of α′\alpha^\prime-NaV2_2O5_5 were grown by self-flux method [36]. The high crystal quality was confirmed by Laue diffraction. The samples were cut and polished on the aa-bb plane with dimensions of the order of 2 x 1.5 mm2\mathrm{mm}^2. To explore the potential for driving coherent structural dynamics in α′\alpha^\prime-NaV2_2O5_5, we performed ultrafast THz pump and NIR probe spectroscopy, illustrated in Figure 1(b). Intense quasi-single-cycle THz pulses were focused onto the sample mounted on a cold finger of a cryostat. The THz-induced response was then tracked by measuring the reflected intensity in a quasi-collinear configuration using ultrashort (∼\sim 55 fs) NIR laser probe pulses, at a photon energy of 1.55 eV, as a function of the pump-probe time delay tppt_{pp} [see Supplemental Material [37], which includes Refs. [38, 39, 40, 41, 42, 43, 44, 45, 46, 47]].

The THz-induced response was monitored by measuring the optical reflectivity modulation ΔR/R=(Ra−Rb)/R\Delta R/R = (R_a - R_b)/R. RaR_a and RbR_b denote the reflected probe intensities under THz pumping along the aa and bb axes, respectively, and R=Ra+RbR = R_a + R_b represents the reflected intensity without the THz pump pulse. The ΔR/R\Delta R/R response can effectively capture the coherent lattice dynamics of the Raman-active phonon mode at frequency ωR\omega_R, since its phonon displacement alters the polarization state of the NIR probe, which is supported by its activity in the cross-polarized (abab) Raman scattering channel [26, 29].

The probe pulses were circularly polarized to allow a more direct comparison with previous optical pump-probe studies [32, 48]. Additionally, we verified that the THz-induced dynamics remain unchanged by probing it with linear polarization at 45∘45^\circ with respect to the bb axis, see Sec. S1 of Supplemental Material [37].

The temporal waveform of the THz pump electric field E(t)E(t), see Figure 1(c), was measured at the sample position through electro-optic sampling utilizing a 200 μm200\,\mu\mathrm{m}-thick GaP crystal (see Sec. S1 of the Supplemental Material [37]). In Figure 1(d), the temporal evolution of the THz pump-induced ΔR/R(tpp)\Delta R/R(t_{\mathrm{pp}}) is depicted. The time origin (tpp=0t_{\mathrm{pp}}=0) is conventionally defined at the delay when the optical probe overlaps the peak of E(t)E(t). The observed THz-induced dynamics exhibit coherent oscillations superimposed on a gradually increasing background. It is important to emphasize that the envelope function of the oscillatory response extends well beyond the time delays when the amplitude of the THz field becomes negligible. These long-lasting oscillatory dynamics imply that the oscillations result from a coherent excitation of a vibrational mode. For a more comprehensive analysis of the THz-induced dynamics and their underlying frequency components, we employ the Fourier transform of both the THz pump waveform and the temporal derivative of the dynamics, represented as d(ΔR/R)/dtppd(\Delta R/R)/dt_{\mathrm{pp}}. The use of the temporal derivative is commonly adopted in ultrafast spectroscopy to mitigate the quasi-DC effect arising from slowly varying background signals in the pump-probe time traces [48]. Figure 1(e) shows the THz pump spectrum compared with the real part of the optical conductivity along the bb axis, which shows three main phonon excitations at ΩIR,1\Omega_{\mathrm{IR},1}, ΩIR,2\Omega_{\mathrm{IR},2}, and ΩIR,3\Omega_{\mathrm{IR},3}. As illustrated in Figure 1(f), the Fourier amplitude unveils a distinct peak at 1.851.85 THz, corresponding to the frequency of the zone-folded Raman-active phonon mode at ΩR\Omega_R. Furthermore, the presence of the weaker Raman-active phonon around 3.023.02 THz (107 cm−1107\,\mathrm{cm}^{-1}) is also detected [29].

To confirm that the observed oscillatory dynamics actually arise from the zone-folded phonons, we investigated the evolution of ΔR/R(tpp)\Delta R/R(t_{\mathrm{pp}}) dynamics across the temperature-driven phase transition. In Figure 2(a), the THz-induced ΔR/R(tpp)\Delta R/R(t_{\mathrm{pp}}) dynamics are presented for temperatures ranging from 55 K to 6060 K. Notably, as the temperature approaches TcT_c, coherent oscillations rapidly vanish while the contribution of the slowly varying background becomes dominant. In Figure 2(b), the Fourier transform amplitude of d(ΔR)/dtppd(\Delta R)/dt_{\mathrm{pp}} for two different temperatures (T=5 K<TcT=5\,\mathrm{K}<T_c and T=60 K>TcT=60\,\mathrm{K}>T_c) is illustrated, indicating the disappearance of the peak at ΩR\Omega_R

Temperature evolution and Fourier analysis of THz pump-induced dynamics

Figure 2. (a) Temperature evolution of the THz pump-induced ΔR/R\Delta R/R in α′\alpha^\prime-NaV2_2O5_5. No prominent signature of coherent oscillations is observed above Tc=34T_c=34 K. Vertical offsets have been added to the curves for better visualization. (b) Fourier transform of d(ΔR/R)/dtppd(\Delta R/R)/dt_{\mathrm{pp}} for T=5T=5 and 6060 K. Black dotted curves are Lorentzian fit, to evaluate the amplitude and the center mode frequency ΩR\Omega_R. (c-d) Normalized amplitude and center frequency ΩR\Omega_R as a function of temperature determined by Lorentzian fit.

for T>TcT>T_c. Moreover, the Fourier amplitude and the peak frequency as a function of temperature, obtained by fitting procedure using a Lorentzian function, are shown in Figure 2(c)-(Figure 2d).

To gain a comprehensive understanding of the driving mechanism of the observed phonon excitation, we conducted measurements to examine the relationship between the oscillation amplitude of the driven oscillatory dynamics and the electric field amplitude of the incident THz field, see Figure 3(a). The measured dynamics show a clear quadratic trend, providing compelling evidence for the nonlinear coupling of the mode to the driving electric field. Additionally, we exclude the possibility of a linear (dipolar) coupling of light to the observed coherent mode, since the mode is purely Raman active. As supporting evidence, Figure 3(b) demonstrates that the dynamics do not change upon polarity inversion of the THz field, providing a strong indication of no direct dipolar excitation.

Normalized oscillations amplitude and THz pump electric field waveform

Figure 3. (a), Normalized oscillations amplitude (Norm. Osc. Ampl.) vs normalized THz field strength. The amplitude scales quadratically with the pump electric field (the light blue solid line represents a quadratic fit). Error bars indicate the standard deviations of the mean determined from pump-probe scan measurements, (b) THz pump electric field waveform E(+)E^{(+)} (dashed line) and polarity inverted one E(−)E^{(-)} and corresponding THz induced ΔR/R(tpp)\Delta R/R(t_{\mathrm{pp}}) dynamics.

Next, we discuss the nonlinear excitation mechanism of the vibrational modes by THz driving that we observed. We conducted a series of THz pump - NIR probe experiments, employing THz spectral filters to tailor the spectrum of the THz driving field, see Figure 4(a)-(Figure 4d). Dotted lines indicate the frequencies of the zone-folded Raman and IR active phonons.

THz pump excitation spectra and corresponding experimental and numerical reflectivity dynamics

Figure 4. (a)–(d), THz pump excitation spectra. The dotted lines indicate the resonance frequency of the Raman phonon, ΩR\Omega_{\mathrm{R}}, and IR-active phonons, ΩIR,1\Omega_{\mathrm{IR},1}, ΩIR,2\Omega_{\mathrm{IR},2}, and ΩIR,3\Omega_{\mathrm{IR},3}. (e)–(f), Corresponding THz-induced ΔR/R\Delta R/R dynamics. α′\alpha'-NaV2_2O5_5 is kept at 5 K. The black dashed line in panel (h) is the exponential fit, see main text. (i)–(n), Numerical calculation using Eq. (4) based on nonlinear phononic model.

When the THz spectrum remains below 2.5 THz, our measurements reveal no significant oscillatory dynamics, see Figure 4(e)}, despite the overlap between the THz pump spectral frequencies with the Raman-active phonon mode frequency at $\Omega_{\mathrm{R}}$. This observation strongly suggests that either linear (dipolar) excitation or nonlinear excitation pathways purely involving photons are very unlikely. Extending the spectral contents up to 4.5 THz, see Fig. 4(b), the THz pump is able to produce the oscillatory vibrational dynamics as the pump frequency contents match those of the IR-active phonons at $\Omega_{\mathrm{IR},1}=3.03$ THz and at $\Omega_{\mathrm{IR},2}=3.8$ THz. Superimposed on the oscillatory dynamics, we also observe a concomitant rise of the slowly varying background. Further extending the THz pump spectral components up to 6 THz, see {xref}Fig. 4(c) reaching the frequency of the phonon at $\omega_{\mathrm{IR},3}=5.4$ THz, resulted in a substantial increase in the background. Finally, when the pump spectral contents are centered mainly around $\omega_{\mathrm{IR},3}$, only the slowly varying component in the dynamics is observable, see {xref}Fig. 4(d). The temporal dynamics are accurately reproduced by a single exponential function $(1-e^{-t_{\mathrm{pp}}/\tau})$, a characteristic often associated with thermal processes {citep}[32, 38]. The fitting of this function yields a time constant of $\tau=1.58$ ps. Remarkably, as one can see in {xref}Fig. 4`, the absence of oscillatory dynamics in cases where the pump spectral contents lie below 2.5 THz (Pump 1) and above 4 THz (Pump 4) strongly suggests that coherent oscillations of the Raman mode are driven through a nonlinear phononic process involving the excitation of the phonon modes at ΩIR,1\Omega_{\mathrm{IR},1}, ΩIR,2\Omega_{\mathrm{IR},2}, and ΩIR,3\Omega_{\mathrm{IR},3}.

Nonlinear Phononics as a Possible Excitation Mechanism

In the following, we model the lattice dynamics driven by nonlinear phononic excitation using an equation-of-motion approach. The indirect excitation of a coherent Raman mode, which arises from the resonant large-amplitude IR-active vibrations, can be described through the lattice potential to the lowest order beyond the harmonic approximation [13, 49]:

V=ΩR2QR22+∑iΩIR,i2QIR,i22−∑i,jcijRQIR,iQIR,jQR.(1)V = \frac{\Omega_{\mathrm{R}}^2 Q_{\mathrm{R}}^2}{2} + \sum_i \frac{\Omega_{\mathrm{IR},i}^2 Q_{\mathrm{IR},i}^2}{2} - \sum_{i,j} c_{ij\mathrm{R}} Q_{\mathrm{IR},i} Q_{\mathrm{IR},j} Q_{\mathrm{R}}. \tag*{(1)}

Here, QRQ_{\mathrm{R}} and ΩR\Omega_{\mathrm{R}} represent the phonon normal coordinate and the eigenfrequency of the Raman-active phonon mode, respectively, while the indices i,j=1,2,3i,j=1,2,3 denote those of the IR-active modes. The parameters cijRc_{ij\mathrm{R}} are the nonlinear coupling coefficients.

Given that the IR-active phonons are directly excited to a large amplitude, the phonon coordinate QIR,iQ_{\mathrm{IR},i} is significantly greater than QRQ_{\mathrm{R}}, i.e., QIR,i≫QRQ_{\mathrm{IR},i} \gg Q_{\mathrm{R}}. This allows us to neglect the coupling terms ciiRQIR,iQIR,iQR2c_{ii\mathrm{R}}Q_{\mathrm{IR},i}Q_{\mathrm{IR},i}Q_{\mathrm{R}}^2. Consequently, the relevant nonlinear anharmonic coupling to the lowest order is given by the term QIR,iQIR,jQRQ_{\mathrm{IR},i}Q_{\mathrm{IR},j}Q_{\mathrm{R}}.

The lattice energy UU is given by U=V−∑iZiQIR,iEU=V-\sum_i Z_i Q_{\mathrm{IR},i}E, which accounts for the interaction between the IR-active phonon modes and the electric field of the THz pump pulse E(t)E(t) [13, 50]. Here, ZiZ_i represents the phonon effective charge of the ii phonon mode.

Following the approach in Refs [13, 50], we derive the equations of motion for both the Raman and IR-active phonons from the lattice potential UU:

Q¨IR,i+2γIR,iQ˙IR,i+ΩIR,i2QIR,i−∑jcijRQIR,jQR=ZiE(2)\begin{aligned} \ddot{Q}_{\mathrm{IR},i}+2\gamma_{\mathrm{IR},i}\dot{Q}_{\mathrm{IR},i}+\Omega_{\mathrm{IR},i}^{2}Q_{\mathrm{IR},i} -\sum_j c_{ij\mathrm{R}}Q_{\mathrm{IR},j}Q_{\mathrm{R}} &= Z_i E \tag*{(2)} \end{aligned}
Q¨R+2γRQ˙R+ΩR2QR=∑i,jcijRQIR,iQIR,j(3)\ddot{Q}_{\mathrm{R}}+2\gamma_{\mathrm{R}}\dot{Q}_{\mathrm{R}}+\Omega_{\mathrm{R}}^{2}Q_{\mathrm{R}}=\sum_{i,j}c_{ij\mathrm{R}}Q_{\mathrm{IR},i}Q_{\mathrm{IR},j} \tag*{(3)}

The (3) describes the temporal evolution of the Raman phonon coordinates QRQ_{\mathrm{R}} as a harmonic oscillator driven by the term QIR,iQIR,jQ_{\mathrm{IR},i}Q_{\mathrm{IR},j}, which is quadratic in the IR-active phonon coordinates. In the presence of nonlinear coupling, this term becomes relevant when IR active phonon modes are driven to large amplitude by the intense and resonant light field. To model the dynamics of the Raman active mode QRQ_{\mathrm{R}}, we numerically solve Eq. (2a) and Eq. (2b), by employing the experimentally measured THz pump fields and the phonon parameters evaluated by infrared and Raman spectroscopy: ΩR=1.85\Omega_{\mathrm{R}} = 1.85 THz, γR=0.15\gamma_{\mathrm{R}} = 0.15 THz, ΩIR,1=3.03\Omega_{\mathrm{IR},1} = 3.03 THz, γIR,1=0.01\gamma_{\mathrm{IR},1} = 0.01 THz, ΩIR,2=3.8\Omega_{\mathrm{IR},2} = 3.8 THz, γIR,2=0.01\gamma_{\mathrm{IR},2} = 0.01 THz, ΩIR,3=5.4\Omega_{\mathrm{IR},3} = 5.4 THz, γIR,3=0.14\gamma_{\mathrm{IR},3} = 0.14 THz [25, 46]. The effective charge ZiZ_i can be estimated from the real part of the optical conductivity σ1\sigma_1 at the IR frequency ωIR,i\omega_{\mathrm{IR},i}, as detailed in Sec. S4 the Supplemental Material [37]. In the numerical calculation, σ1(ω1)\sigma_1(\omega_1), σ1(ω2)\sigma_1(\omega_2), σ1(ω3)\sigma_1(\omega_3) are taken from the optical conductivity in Figure 1, see Refs. [25, 46] for details.

We model the induced reflectivity changes as [42]:

ΔRR(tpp)=A⋅QR(tpp)+B⋅η(tpp),(4)\frac{\Delta R}{R}(t_{\mathrm{pp}}) = A \cdot Q_{\mathrm{R}}(t_{\mathrm{pp}}) + B \cdot\eta(t_{\mathrm{pp}}), \tag*{(4)}

where the first term is proportional to the phonon coordinate QRQ_{\mathrm{R}}, as discussed in Ref. [42], representing the excited nonlinear vibrational dynamics. The second term characterizes the slowly varying background, attributed to the thermal effect due to the deposition of optical energy. To model these thermal dynamics, it is reasonable to assume a single exponential form η(tpp)=(1−e−tpp/τTHz)\eta(t_{\mathrm{pp}}) = (1-e^{-t_{\mathrm{pp}}/\tau_{\mathrm{THz}}}), consistent with our experimental observations, as depicted in Figure 4(h). For the model, we set a time constant of τTHz=1.58\tau_{\mathrm{THz}} = 1.58 ps as determined by fitting the experimental data, see Figure 4(h).

Comparison between THz pump and NIR pump induced reflectivity changes

Figure 5. Comparison between THz pump and NIR pump induced ΔR/R(tpp)\Delta R/R(t_{\mathrm{pp}}) in α′\alpha'-NaV2_2O5_5. The sample is kept at 5 K. NIR pump fluence is ∼3\sim3 mJ cm−2^{-2}. The dotted black lines are single-exponential fit. Vertical offset has been added for better visualization.

Numerical results of Eq. (4) are reported in Figure 4(i)–(Figure 4n), while details about the AA and BB parameters are reported in Sec. S3 of the Supplemental Material [37]. Our calculations show a consistent agreement with observed dynamics and elucidate its dependence on the spectral content of the THz pump. Furthermore, in Sec. S2 of the Supplemental Material [37], we have extended our analysis by performing a comparative study of the numerical results of nonlinear phononics with other excitation mechanisms, such as nonlinear photonics and infrared resonant Raman scattering [13, 15, 41]. This analysis provides strong evidence that nonlinear phononics is the dominant driving mechanism underlying the observed dynamics. However, determining the selection rules for the nonlinear phononic excitation based on phonon symmetry compositions remains challenging due to unresolved ambiguities in the crystal symmetry of the low-temperature charge-order phase of α\alpha-NaV2O5\mathrm{NaV}_2\mathrm{O}_5. Competing structural models, including centrosymmetric C2/cC2/c, noncentrosymmetric Fmm2Fmm2, and monoclinic C23−A112C_{2}^{3}-A112 space groups, have been proposed based on infrared and Raman spectroscopy and X-ray diffraction studies [29, 44]. If the structure is centrosymmetric (C2/cC2/c) as discussed in Ref. [29, 44], the nonlinear coupling follows the composition rule Bu×Bu=AgB_u \times B_u=A_g, allowing IR phonons (BuB_u symmetry) to excited the zone-folded Raman-active phonon (AgA_g symmetry). Same conclusion is found for the monoclinic C23−A112C_{2}^{3}-A112 case. However, if the crystal symmetry is non-centrosymmetric (Fmm2Fmm2), as discussed in Refs. [29], conventional composition rules cannot explain the nonlinear phononics excitation. Detailed discussion for both cases is given in Sec. S5 of the Supplemental Material [37].

Further complexity arises from charge order, which can influence the local crystal symmetry. Infrared and X-ray studies have revealed coexisting ladder-plane configurations characterized by distinct diagonal charge-order patterns [25, 47], leading to local structural variations that deviate from the average symmetry. These local symmetry changes have been proposed as the origin of the anomalous symmetry rules observed for the zone-folded IR phonons (including those at ΩIR,1\Omega_{\mathrm{IR},1} and ΩIR,2\Omega_{\mathrm{IR},2}) in polarization-resolved infrared spectroscopy [25, 46]. Consequently, these variations may alter the effective selection rules governing nonlinear phononic coupling. Further discussion on these structural and symmetry-related effects can be found in Sec. S5 of the Supplemental Material [37]. Nonetheless, despite these ambiguities, the remarkable agreement between the nonlinear phononic model and the experimental data - especially in contrast to models based on alternative excitation mechanisms - strongly supports nonlinear phononics as the most likely excitation pathway for the observed phenomena (see Sec. S3 of the Supplemental Material for a full analysis [37]).

Next, we phenomenologically discuss the origin of the slowly varying background in the observed dynamics. The slow modulation of the ΔR/R\Delta R/R change at the probe photon energy (1.55 eV) under THz excitation is indicative of a temperature increase, consistent with previous steady-state polarization-resolved infrared spectroscopy investigations [51]. This temperature increase can be attributed to two possible mechanisms. First, through lattice excitation, the strongly driven IR-active phonons by THz pulse may generate acoustic phonons via phonon-phonon scattering. This scattering process gradually transfers energy from the IR-active phonons to the acoustic phonons, leading to a slow and steady increase in the lattice temperature. Among the IR-active phonons, the IR3 phonon is likely to play a particularly efficient role in this energy transfer. A second potential mechanism involves IR-active spin excitations. Infrared spectroscopy has shown that α′\alpha'-NaV2O5\mathrm{NaV}_2\mathrm{O}_5 exhibits a magnon continuum above 139 cm−1139\ \mathrm{cm}^{-1} (4.2 THz) [46]. It is possible that THz light directly excites these “charged” magnons, as they are dipole-active [52]. Consequently, through spin-phonon coupling, the energy from these excitations could then be transferred to the lattice, contributing to a rise in temperature.

The relaxation of these thermal processes leads to an overall increase in temperature, resulting in a slowly varying background in the pump-probe dynamics. This slow modulation of the reflectivity change reflects the gradual thermalization of the system following THz excitation, independent of the specific pathway, whether phonon-driven or spin-driven, that initiates the temperature increase, eventually leading to a picosecond melting of the charge order and lattice superstructure, as proposed in the photoexcitation of the low-temperature phase of α′\alpha'-NaV2O5\mathrm{NaV}_2\mathrm{O}_5 [32, 33].

Finally, we performed a near-infrared (NIR) pump - NIR probe experiment at 5 K, merely replacing the THz stimulus with the NIR pump at a photon energy of 1.55 eV. The dynamics induced by the NIR pump are characterized by a single exponential evolution with a time constant of τNIR=2.82\tau_{\mathrm{NIR}}=2.82 ps. Notably, this indicates that the photoinduced dynamics are slower compared to those achieved with THz excitation (τTHz=1.58\tau_{\mathrm{THz}}=1.58 ps). This suggests that direct absorption of the THz energy through lattice vibrations or spin excitations leads to a faster increase in lattice temperature, and eventually melting charge-ordered phase, compared to the photoexcitation electronic system. Strikingly, no significant oscillatory component is observed through optical photoexcitation. This result highlights the potential of nonlinear THz light-matter interaction in driving Raman-active phonon mode compared to purely optical pathways involving high-energy photons, such as impulsive stimulated Raman scattering [53].

Conclusion

In summary, our experimental results and analytical modeling strongly suggest that the nonlinear dynamics of zone-folded Raman-active phonons in α′\alpha'-NaV2O5\mathrm{NaV}_2\mathrm{O}_5 originate from the excitation of zone-folded IR-active phonons driven to large amplitudes by intense THz pulses. This nonlinear phononic excitation process might be enhanced by the presence of charge order and superstructure formation, as suggested by Raman spectroscopy, which reveals significant lattice anharmonicity in the low-temperature phase [35]. Further studies employing ab initio density functional theory calculations could quantitatively estimate these coupling coefficients in the charge-ordered phase of α′-NaV2O5\alpha^\prime\text{-NaV}_2\text{O}_5 and shed light on the excitation mechanism at the microscopic level. Moreover, while disentangling electronic and lattice-driven responses remains challenging in time-resolved optical spectroscopy, ultrafast X-ray diffraction experiments at free-electron lasers offer the potential to yield crucial insights into the magnitude of coherent lattice displacements and a deeper understanding of the underlying dynamics.

In conclusion, this study extends the exploration of nonlinear THz-driven lattice dynamics to the quantum spin-ladder compound α′-NaV2O5\alpha^\prime\text{-NaV}_2\text{O}_5. Future investigations could focus on the dynamical modulation of the Dzyaloshinskii-Moriya (DM) interaction by THz-driven transient lattice displacements. Steady-state spectroscopic studies have already implicated lattice-mediated DM interactions in generating finite Raman and infrared cross-sections for single- and two-magnon scattering processes [25]. By employing resonant THz-driven coherent phonons, future experiments could probe the ultrafast modulation of the infrared absorption cross-section under transient lattice displacements. Such efforts would provide novel insights into the dynamics of the DM interaction, offering experimental validation of spin-phonon coupling mechanisms that have thus far been explored primarily in steady-state regimes.

Overall, exploiting nonlinear THz light-matter interaction is emerging as an effective approach for controlling spin and lattice dynamics in strongly correlated systems. These materials hold great potential for applications in next-generation technologies, including high-speed spintronics and THz-driven real-time data processing, paving the way for groundbreaking advances in functional device applications.

Acknowledgments - The authors thank P. Barone for insightful discussions and gratefully acknowledge the support of the PSI-LNO GL group during the THz pump-probe experiments. FT-IR analyses were conducted at the DAΦNE-LIGHT Synchrotron Radiation Facility. The authors sincerely acknowledge the Synchrotron Radiation Service for their invaluable technical assistance with the IR spectroscopy experiments. system α′\alpha^\prime-NaV2_2O5_5,” Physical review letters 88, 127201 (2002).

[49] T. P. Martin and L. Genzel, “Ionic raman scattering and ionic frequency mixing,” physica status solidi (b) 61, 493–502 (1974).

[50] A. A. Melnikov, K. N. Boldyrev, Y. G. Selivanov, V. P. Martovitskii, S. V. Chekalin, and E. A. Ryabov, “Coherent phonons in a Bi2_2Se3_3 film generated by an intense single-cycle thz pulse,” Physical Review B 97, 214304 (2018).

[51] C. Presura, D. van der Marel, A. Damascelli, and R. K. Kremer, “Low-temperature ellipsometry of α′\alpha^\prime-NaV2_2O5_5,” Phys. Rev. B 61, 15762–15765 (2000).

[52] A. Damascelli, D. van der Marel, M. Grüninger, C. Presura, T. T. M. Palstra, J. Jegoudez, and A. Revcolevschi, “Direct two-magnon optical absorption in α′\alpha^\prime-NaV2_2O5_5:“charged” magnon,” Phys. Rev. Lett. 81, 918–921 (1998).

[53] R. Merlin, “Generating coherent thz phonons with light pulses,” Solid State Communications 102, 207–220 (1997), highlights in Condensed Matter Physics and Materials Science.

Supplemental Material

Experiment Setup

The experimental configuration of the THz pump and optical probe setup is sketched in Figure S1. The output of a 20 mJ, 55 fs, 800 nm Ti:Sapphire laser was used to drive an optical parametric amplifier (OPA), which provides ultrashort, multi-mJ pulses at longer wavelengths [38, 39]. The generation of single-cycle THz pulses was based on optical rectification in organic crystal OH1 (2-[3-(4-hydroxystyryl)-5, 5-dimethylcyclohex-2-enylidene] - from Rainbow Photonics Ltd.) pumped with the OPA signal beam at 1.5 μ\mum with an energy of 3.2 mJ.

Experimental setup for THz pump NIR probe reflection spectroscopy, THz beam profile, and temporal THz intensity waveform

Figure S1. (a) Experimental setup for THz pump NIR probe reflection spectroscopy. Optical Parametric Amplifier (OPA), parabolic mirrors (PM1-PM3), delay line (DL), wave plate (WP), Wollaston prism (W), and detectors (D1-D2). (b) THz beam profile at the sample position measured by the THz camera. Beam waists obtained by Gaussian fitting are respectively wx=118w_x = 118 µm and wy=122w_y = 122 µm (average waist w=120w = 120 µm). (c) Temporal THz intensity waveform obtained as the square of the electric field measured by electro-optic sampling, using a Gaussian-envelope fit to determine the pulse duration.

Following the THz generation crystal, three low-pass filters were used to isolate the THz pulses and block the residual OPA beam: two of them featured a 20 THz cut-off frequency, and the third had a 6 THz cut-off frequency. These three filters gave an optical extinction ratio exceeding 10510^5 at the OPA signal wavelength. Figure 1(c) of the main text illustrates the resulting THz pump waveform. Intense THz electric fields were reached by tight focusing of the THz beam using three parabolic mirrors [38]. To tailor the THz pump spectrum and obtain the four distinct pump pulses shown in Figure 4(a)-(Figure 4d) of the main text, specific spectral filters were employed. We used a 2 THz low-pass filter, a 4.2 THz low-pass filter, a 6 THz low-pass filter, and a 4.2 THz high-pass filter coupled with a 10 THz low-pass filter.

To measure NIR probe dynamics as a function of the pump strength, see Figure 3(a) in the main text, the THz electric field was attenuated by three wire-grid polarizers. The reversal of the THz pump field polarity, depicted in Figure 3(b) in the main text, was implemented by a 180° rotation of the OH1 crystal around the light propagation axis.

To estimate the electric-field strength of the pump pulses shown in Figure 1 of the main text, we apply the formula [40]

ETHz=z0Ep4ln⁡2ππw2τFWHM(5)E_{\mathrm{THz}} = \sqrt{\frac{z_0 E_p^4 \sqrt{\ln2}}{\pi\sqrt{\pi} w^2 \tau_{\mathrm{FWHM}}}} \tag*{(5)}

where z0z_0 is the vacuum impedance and we measured the THz energy per pulse, Ep=3E_p = 3 μ\muJ, using a calibrated THz energy meter (Gentec THZ12D3S-VP-D0); the beam waist, w=120w = 120 μ\mum, obtained by a Gaussian fit of the beam profile Figure S1, which was measured with a micro-bolometric THz camera (NEC IRV-T0830); the pulse duration, τFWHM=0.477\tau_{\mathrm{FWHM}} = 0.477 ps, obtained from the FWHM of the Gaussian envelope fitting the temporal intensity waveform of the THz pump Figure S1, which was taken as the square of the electric field measured at the sample position by electro-optic sampling in transmission in a 200 μ\mum-thick (110) GaP crystal with a 50 fs, 800 nm gating pulse obtained as a fraction of the Ti:Sapphire beam. Our estimated electric-field strength, ETHz=3.14E_{\mathrm{THz}} = 3.14 MVcm−1^{-1}.

For the THz pump-NIR probe measurements, the gating pulse was employed to capture the ultrafast pump-induced polarization modulation of the sample in reflectivity. The THz-induced polarization modulation was measured by splitting the probe beam into two orthogonal components using a Wollaston prism.

The THz electric field of the pump was polarized in the sample plane, perpendicular to the optical table, and along the bb axis of the α′\alpha^\prime-NaV2_2O5_5 crystal, with an uncertainty of ±5∘\pm5^\circ. The polarization of the incident probe beam sample was conventionally set to circular for a direct comparison with previous optical pump-probe experiments on α′\alpha^\prime-NaV2_2O5_5. We also verify that comparable pump-probe dynamics are obtained using a 45°-linearly polarized probe with respect to the pump polarization, see Figure S2. All measurements were conducted using a He cryostat, allowing us to achieve a minimum sample temperature of 5 K.

Comparison between circularly polarized NIR probe and 45° linearly-polarized NIR probe

Figure S2. THz pump NIR probe: comparison between circularly polarized NIR probe and 45° linearly-polarized NIR probe.

THZ-INDUCED COHERENT EXCITATION RAMAN-ACTIVE MODES IN α′\alpha^\prime-NaV2_2O5_5: NONLINEAR PHONONICS VS NONLINEAR PHOTONICS

When intense THz pulses excite the crystal lattice, distinct nonlinear pathways - phononic and photonic - can contribute to the generation of Raman-active coherent phonons [13, 15, 16]. In photonic pathways, the driving mechanism is governed by Raman scattering processes resonant with the sum- and difference-frequency components of the light field and commonly referred to as ”photonic” Raman scattering. In phononic processes, the excitation of Raman-active modes occurs through resonant coupling to IR vibrational modes driven by intense light and includes both ionic Raman scattering and infrared resonant Raman scattering. In the first case, the coupling between phonon modes is mediated by the anharmonic lattice potential, in the latter case by the nonlinear lattice polarizability [41].

The equation of motion for the Raman-active phonon coordinate QRQ_R can be expressed as follows:

Q¨R+2γRQ˙R+ΩR2QR=F,(6)\ddot{Q}_R + 2\gamma_R\dot{Q}_R + \Omega_R^2Q_R = \mathcal{F}, \tag*{(6)}

where F\mathcal{F} represents the driving force acting on the Raman-active phonon, originating from either photonic or phononic processes. Here, ΩR\Omega_R is the phonon eigenfrequency, and 2γR2\gamma_R represents the phonon linewidth. For the photonic Raman scattering (RS) mechanism, F\mathcal{F} is proportional to the square of the pump electric field E(t)E(t):

F=RE(t)2,\mathcal{F} = R E(t)^2,

being RR the phonon Raman tensor [13, 15]. In the ionic RS process, as discussed in the main text, the driving force is given by F=∑ijcijRQIR,iQIR,j\mathcal{F} = \sum_{ij} c_{ijR}Q_{IR,i}Q_{IR,j}, which arises from the ionic displacement of the IR-active phonons excited by the THz pulse, where cijRc_{ijR} represent the nonlinear coupling coefficients [13]. The index i,j=1,2,3i,j = 1,2,3 indicates the IR1, IR2 and IR3 phonon modes.

In infrared resonant Raman scattering (IRRS) [41], instead the driving force is F=∑ibiQIR,iE(t)\mathcal{F} = \sum_i b_iQ_{IR,i}E(t), where bib_i represents the change in the effective charge of the i-th IR-active mode relative to the Raman-active phonon [41].

In the following, we present a comparative analysis of the lattice dynamics modeled using ionic RS, IRRS, and photonic RS, under the THz pump excitation conditions of the experiment.

To compare these models with the experimental data, we first extract the oscillatory components from the measured pump-probe time traces (shown in Figure 4(e)-(Figure 4h) in the main text). This is achieved by isolating the oscillatory signals via subtraction of a slowly varying background by fitting the background with an exponential function η(tpp)=c⋅(1−e−tpp/τ)\eta(t_{pp})=c\cdot(1-e^{-t_{pp}/\tau}), where cc represents the amplitude scaling factor and τ\tau the characteristic time constant. Figure S3 shows both the extracted oscillatory components and the exponential fits. The time constant is fixed to τpp=1.58\tau_{pp}=1.58 ps for all the fits and cc is set to 0.002 for Pump 1, 0.0039 for Pump 2, 0.029 for Pump 3, and 0.029 for Pump 4.

Four panels showing experimental data and ionic, IRRS, and photonic RS models

Figure S3. (a) Extracted oscillatory component from THz pump-induced ΔR/R\Delta R/R dynamics utilizing an exponential fitting procedure. The gray solid lines in these panels represent the fit using the function η(tpp)=(1−e−tpp/τ)\eta(t_{pp})=(1-e^{-t_{pp}/\tau}), with a time constant τpp\tau_{pp} of 1.58 ps. Colored solid lines depict experimental data after subtracting the fit. (b)-(d) Calculated oscillatory dynamics of ΔR/R\Delta R/R employing the ionic RS model (b) the IRRS model (c) and the photonic RS model (d).

In the numerical models, we computed the temporal dynamics of QRQ_R using THz pump pulses that were experimentally measured by electro-optic sampling, see Sec. 1 for details. The spectrum of these pulses is shown in Figure 4(a)-(Figure 4c) of the main text. The oscillatory component is modeled as ΔR/R∝QR\Delta R/R\propto Q_R, as discussed in Ref. [42]. For the ionic RS model, the phonon parameters employed in the calculation are discussed in the main text, while the coefficients cij,Rc_{ij,R} were phenomenologically set to have equal magnitudes due to the absence of precise information about their specific values. The results are shown in Figure S3.

For the IRRS model, we numerically calculate the QRQ_R dynamics with a driving force F=∑ibiQIR,iE(t)F=\sum_i b_iQ_{\mathrm{IR},i}E(t). In these calculations, we assumed equal values for the bib_i parameters due to the lack of specific information for the α′\alpha'-NaV2_2O5_5.

In the photonics RS model, we numerically solved Eq. (6), with F=RE(t)2F=RE(t)^2, as in Ref. [13, 15]. The results from the photonics RS model are depicted in Figure S3.

As shown in Figure S3, the amplitude of coherent oscillations in the experimentally observed dynamics is significantly suppressed when excited by a pump containing lower frequency components (Pump 1). This suppression is consistent with the ionic RS model, as the THz pump spectrum lies below the frequency of the IR-active phonons. In contrast, the ΔR/R\Delta R/R response predicted by the IRRS and photonic RS models (Figure S3) exhibits pronounced oscillatory dynamics with a significant amplitude when driven with Pump 1.

This behavior is better shown in Figure S4, where the spectrum of the oscillatory component of ΔR/R\Delta R/R experimentally measured, shown in Figure S4, is compared with the spectrum of the computed driving force F\mathcal{F} for the different driving mechanisms, see Figure S4-(Figure S4).

Fourier spectra and spectral amplitudes for different Raman scattering mechanisms

Figure S4. (a) Fourier spectra of the oscillatory component of ΔR/R\Delta R/R dynamics from Figure S3. (b)-(d) Spectral amplitude of the calculated driving force FF corresponding to different Raman scattering mechanisms. The black dashed lines mark the phonon resonance frequency at 1.85 THz. (e)-(h) Spectral amplitudes at ΩR=1.85\Omega_R=1.85 THz extracted from panels (a)-(d), plotted as a function of Pump 1 through Pump 4.

The spectral amplitude at the phonon peak frequency (1.85 THz) for Pump 1 to Pump 4 is shown in Figure S4 and compared with the spectral amplitude of the driving forces, see Figure S4-(Figure S4).

While the ionic RS model aligns closely with the resonant behavior observed in the experiment, the IRRS and photonic RS models do not accurately reproduce it. This reinforces the conclusion that the observed dynamics are more likely driven by nonlinear phononics, rather than the mechanisms proposed by the IRRS or photonic RS models.

Model Parameters

In Table 1, we present the values of the constants AA and BB in the expression: ΔR/R(tpp)=A⋅QR(tpp)+B⋅η(tpp)\Delta R/R(t_{pp}) = A \cdot Q_R(t_{pp}) + B \cdot\eta(t_{pp}) for each pump in Figure 4(i)-(Figure 4n) of the main manuscript. The constant AA, governing the oscillation amplitude is kept unchanged for all pumps, while the constant BB is adjusted to reproduce the experimentally observed slowly varying background.

AB
Pump 1703⋅10−33 \cdot10^{-3}
Pump 2705⋅10−35 \cdot10^{-3}
Pump 3702.8⋅10−22.8 \cdot10^{-2}
Pump 47038⋅10−38 \cdot10^{-3}

Table 1.

Relation Between Conductivity and Effective Charge

In this section, we discuss how the phonon effective charge can be estimated from the real part of the optical conductivity σ1(ω)\sigma_1(\omega). In particular, the ionic contribution to the conductivity, associated with the presence of an IR-active phonon mode, can be generally written as [43]

σ1ion(ω)∝iωZeff2(ω+iγIR)2−ωIR2,(7)\sigma_1^{ion}(\omega) \propto i\omega\frac{Z_{eff}^{2}}{(\omega+i\gamma_{IR})^{2}-\omega_{IR}^{2}}, \tag*{(7)}

where tensor indices are neglected for simplicity. Here ZeffZ_{eff} is the phonon effective charge, ωIR\omega_{IR} and γIR\gamma_{IR} are the IR-active phonon frequency and effective phonon broadening, respectively. Assuming that γIR≪ωIR\gamma_{IR} \ll\omega_{IR}, the real part of the ionic conductivity is given by

σ1ion(ω)∝Zeff22γIRω2(ω2−ωIR2)2+4γIR2ω2.(8)\sigma_1^{ion}(\omega) \propto Z_{eff}^{2}\frac{2\gamma_{IR}\omega^{2}}{(\omega^{2}-\omega_{IR}^{2})^{2}+4\gamma_{IR}^{2}\omega^{2}}. \tag*{(8)}

Since phonon peaks are usually rather sharp, σ1ion\sigma_1^{ion} is controlled by its ω=ωIR\omega=\omega_{IR} value, i.e.

σ1ion(ωIR)∝Zeff2γIR.(9)\sigma_1^{ion}(\omega_{IR}) \propto\frac{Z_{eff}^{2}}{\gamma_{IR}}. \tag*{(9)}

When the electronic contribution to the conductivity can be neglected, as it happens e.g. in insulating compounds with phonon peaks below the optical gap, σ1(ωIR)\sigma_1(\omega_{IR}) is well approximated by (9), and the effective charge can be estimated as:

Zeff∝σ1(ωIR)γIR.(10)Z_{eff} \propto\sqrt{\sigma_1(\omega_{IR})\gamma_{IR}}. \tag*{(10)}

Point Group and Product of Irreducible Representations for the Nonlinear Phononic Excitation

In this section, we examine the symmetry properties of THz-driven phonons relevant to the nonlinear phononic excitation mechanism of the observed Raman-active mode. The exact lattice symmetry of the low-temperature charge-order phase of α′\alpha^\prime-NaV2_2O5_5 remains a subject of debate. While infrared spectroscopy and Raman scattering studies have proposed that the LT phase is described either by the centrosymmetric space group C2/cC2/c or the non-centrosymmetric Structure - Fmm2Fmm2 [29, 44], X-ray diffraction suggests that the crystal structure may instead correspond to the monoclinic C23−A112C_2^3 - A112 space group, hosting multiple charge order configurations [45]. In the following, we analyze the symmetry selection rules for the nonlinear phononic excitation mechanism for all proposed space group symmetries. In the THz range of our experiment, the IR-active phonons along the bb-axis are labeled as IR1 (at 111 cm−1^{-1}), IR2 (at 132 cm−1^{-1}), and IR3 (at 180 cm−1^{-1}). The modes IR1 and IR2 correspond to zone-folded phonons that emerge in the LT phase due to the formation of a lattice superstructure.

  • Centrosymmetric Structure - C2/cC2/c, C2hC2h (2/m2/m): The folded phonon modes excited by the THz field exhibit are BuB_u, with displacement coordinates confined to the abab plane [29, 44]. The observed folded Raman-active phonon has AgA_g symmetry and, therefore, it can be detected via transient anisotropic reflectivity, as it is active in the abab Raman channel [44]. According to the symmetry multiplication rules, Ag=Bu×BuA_g = B_u \times B_u. Therefore the nonlinear phononic excitation mechanism is supported by this crystal symmetry.

  • Non-centrosymmetric Structure - Fmm2Fmm2, C2vC_{2v} (mm2mm2): The THz-excited IR-active folded phonon modes, IR1\mathrm{IR}_1 and IR2\mathrm{IR}_2, are proposed to exhibit symmetries of (2B1+2B22B_1 + 2B_2), making them IR-active along both the crystallographic aa (B1B_1) and bb (B2B_2) axes, as discussed in Ref.[30]. The excited IR3\mathrm{IR}_3 phonon mode is characterized by B2B_2 symmetry and is exclusively IR-active along the bb axis[25, 30, 46]. For the Fmm2Fmm2 space group, the coherent phonon observed experimentally is Raman-active in the abab channel and possesses A2A_2 symmetry. Notably, this mode is degenerate in frequency with a phonon mode of A1A_1 symmetry, which is also purely Raman-active but observed in the aaaa, bbbb, and cccc channels [29]. Since the THz pump is polarized along the bb axis, the IR1\mathrm{IR}_1, IR2\mathrm{IR}_2, and IR3\mathrm{IR}_3 phonon modes are primarily excited with B2B_2 symmetry. Applying multiplication rules yields B2×B2=A1B_2 \times B_2 = A_1, rather than A2A_2. Thus, multiplication rules for this crystal symmetry do not support the nonlinear phononic mechanism.

  • Monoclinic Structure - C23C_2^3-A112A112: In this case, the point group lacks the two mirror planes present in the C2vC_{2v} symmetry [45]. Consequently, the B1B_1 and B2B_2 modes, as well as the A1A_1 and A2A_2, in this lower symmetry structure transform with the same irreducible representation, AA and BB respectively. Consequently, all phonon modes in this lower-symmetry structure are predicted to exhibit both infrared (IR) and Raman activity. In the Schoenflies notation, C23C_2^3 corresponds to a monoclinic cell with point group C2C_2. For this point group, symmetry multiplication rules yield B×B=AB \times B = A, meaning that the indirect excitation of a phonon with AA symmetry can be obtained by driving the BB phonon modes. Therefore the nonlinear phononic excitation mechanism is supported by this crystal symmetry.

Beyond the ambiguities in the crystal symmetry of α′\alpha'-NaV2_2O5_5, charge ordering introduces further complexity. Polarization-resolved infrared spectroscopy experiments have revealed anomalies in the symmetry selection rules for zone-folded IR-active phonon modes, which have been attributed to charge ordering [25, 46]. Specifically, it has been proposed that multiple charge-order configurations substantially alter the short-range crystal structure of α′\alpha'-NaV2_2O5_5, leading to local deviations from its average symmetry - a conclusion further supported by X-ray studies [47]. This structural complexity makes the definition of selection rules in the low-temperature charge-ordered phase challenging, as well as the assignment of symmetry to the zone-folded phonons. The divergence between local and average crystal symmetries may also imply that conventional symmetry composition rules for nonlinear phononics cannot be directly applied.

References

  1. [1]P. Salén, M. Basini, S. Bonetti, J. Hebling, M. Krasilnikov, A. Y. Nikitin, G. Shamuilov, Z. Tibai, V. Zhaunerchyk, and V. Goryashko, “Matter manipulation with extreme terahertz light: Progress in the enabling thz technology,” Physics reports 836, 1–74 (2019).DOI
  2. [2]Y. Zhang, K. Li, and H. Zhao, “Intense terahertz radiation: generation and application,” Frontiers of Optoelectronics 14, 4–36 (2021).DOI
  3. [3]A. Zong, B. R. Nehen, S.-C. Lin, J. A. Spies, and M. Zuerch, “Emerging ultrafast techniques for studying quantum materials,” Nature Reviews Materials 8, 224–240 (2023).DOI
  4. [4]A. de la Torre, D. M. Kennes, M. Claassen, S. Gerber, J. W. McIver, and M. A. Sentef, “Colloquium: Nonthermal pathways to ultrafast control in quantum materials,” Rev. Mod. Phys. 93, 041002 (2021).
  5. [5]D. N. Basov, R. D. Averitt, and D Hsieh, “Towards properties on demand in quantum materials,” Nature materials 16, 1077–1088 (2017).DOI
  6. [6]A. S. Disa, J. Curtis, M. Fechner, A. Liu, A. von Hoegen, M. Först, T. F. Nova, P. Narang, A. Maljuk, A. V. Boris, B. Keimer, and A. Cavalleri, “Photo-induced high-temperature ferromagnetism in YTiO₃,” Nature 617, 73–78 (2023).DOI
  7. [7]A. S. Disa, T. F. Nova, and A. Cavalleri, “Engineering crystal structures with light,” Nature Physics 17, 1087–1092 (2021).DOI
  8. [8]M. Budden, T. Gebert, M. Buzzi, G. Jotzu, E Wang, T. Matsuyama, G. Meier, Y. Laplace, D. Pontiroli, M. Riccò, F. Schlawin, D. Jaksch, and A. Cavalleri, “Evidence for metastable photo-induced superconductivity in K₃C₆₀,” Nature Physics 17, 611–618 (2021).DOI
  9. [9]M. Buzzi, D. Nicoletti, M. Fechner, N. Tancogne-Dejean, M. A. Sentef, A. Georges, T. Biesner, E. Uykur, M. Dressel, A. Henderson, T. Siegrist, J. A. Schlueter, K. Miyagawa, K. Kanoda, M.-S. Nam, A. Ardavan, J. Coulthard, J. Tindall, F. Schlawin, D. Jaksch, and A. Cavalleri, “Photomolecular high-temperature superconductivity,” Physical Review X 10, 031028 (2020).
  10. [10]X. Li, T. Qiu, J. Zhang, E. Baldini, J. Lu, A. M. Rappe, and K. A. Nelson, “Terahertz field-induced ferroelectricity in quantum paraelectric srtio3,” Science 364, 1079–1082 (2019).DOI
  11. [11]Y.H. Wang, H. Steinberg, P. Jarillo-Herrero, and N. Gedik, “Observation of floquet-bloch states on the surface of a topological insulator,” Science 342, 453–457 (2013).
  12. [12]M. Basini, M. Pancaldi, B. Wehinger, M. Udina, V. Unikandannunni, T. Tadano, M.C. Hoffmann, A.V. Balatsky, and S. Bonetti, “Terahertz electric-field-driven dynamical multiferroicity in srtio3,” Nature , 1–6 (2024).
  13. [13]D. M Juraschek and S. F. Maehrlein, “Sum-frequency ionic raman scattering,” Physical Review B 97, 174302 (2018).DOI
  14. [14]D. M. Juraschek, D. S. Wang, and P. Narang, “Sum-frequency excitation of coherent magnons,” Physical Review B 103, 094407 (2021).
  15. [15]S. Maehrlein, A. Paarmann, M. Wolf, and T. Kampfrath, “Terahertz sum-frequency excitation of a raman-active phonon,” Phys. Rev. Lett. 119, 127402 (2017).DOI
  16. [16]C. L Johnson, B. E. Knighton, and J. A. Johnson, “Distinguishing nonlinear terahertz excitation pathways with two-dimensional spectroscopy,” Physical Review Letters 122, 073901 (2019).DOI
  17. [17]F. Giorgianni, M. Udina, T. Cea, E. Paris, M. Caputo, M. Radovic, L. Boje, J. Sakai, C. W Schneider, and S. L. Johnson, “Terahertz displacive excitation of a coherent raman-active phonon in v2o3,” Communications Physics 5, 103 (2022).DOI
  18. [18]D. Afanasiev, J. R. Hortensius, B. A. Ivanov, A. Sasani, E. Bousquet, Y. M. Blanter, R. V. Mikhaylovskiy, A. V. Kimel, and A. D. Caviglia, “Ultrafast control of magnetic interactions via light-driven phonons,” Nature materials 20, 607–611 (2021).
  19. [19]F. Giorgianni, B. Wehinger, S. Allenspach, N. Colonna, C. Vicario, P. Puphah, E. Pomjakushina, B. Normand, and Ch. Rüegg, “Ultrafast frustration breaking and magnetophononic driving of singlet excitations in a quantum magnet,” Physical Review B 107, 184440 (2023).DOI
  20. [20]E. A. Mashkovich, K. A. Grishunin, R. M Dubrovnin, A. K. Zvezdin, R. V. Pisarev, and A. V. Kimel, “Terahertz light-driven coupling of antiferromagnetic spins to lattice,” Science 374, 1608–1611 (2021).DOI
  21. [21]C. Vaswani, J. H. Kang, M. Mootz, L. Luo, X. Yang, C. Sundahl, D. Cheng, C. Huang, R. H. J. Kim, Z. Liu, et al., “Light quantum control of persisting higgs modes in iron-based superconductors,” Nature Communications 12, 258 (2021).
  22. [22]L. Schwarz, B. Fauseweh, N. Tsuji, N. Cheng, N. Bitner, H. Krull, M. Berciu, G.S. Uhrig, A.P. Schmÿder, S. Kaiser, et al., “Classification and characterization of nonequilibrium higgs modes in unconventional superconductors,” Nature communications 11, 287 (2020).
  23. [23]H. Chu, M.-J. Kim, K. Katsumi, S. Kovalev, R. D. Dawson, L. Schwarz, N. Yoshikawa, G. Kim, D. Putzky, Z. Zhong Li, et al., “Phase-resolved higgs response in superconducting cuprates,” Nature communications 11, 1793 (2020).DOI
  24. [24]B. Keimer and J. E. Moore, “The physics of quantum materials,” Nature Physics 13, 1045–1055 (2017).DOI
  25. [25]T. Room, D. Hüvonen, U. Nagel, Y.-J. Wang, and R. K. Kremer, “Low-energy excitations and dynamic dzyaloshinskii-moriya interaction in α′- NaV2O5 studied by far-infrared spectroscopy,” Physical Review B 69, 144410 (2004).
  26. [26]P. Lemmens, M. Fischer, G. Els, G. Güntherodt, A. S. Mishchenko, M. Weiden, R. Hauptmann, C. Geibel, and F. Steglich, “Magnetic bound states in the quarter-filled ladder system α′- NaV2O5,” Physical Review B 58, 14159 (1998).
  27. [27]T. Ohama, H. Yasuoka, M. Isobe, and Y. Ueda, “Mixed valency and charge ordering in α′-NaV2O5,” Physical Review B 59, 3299 (1999).
  28. [28]H. Smolinski, C. Gros, W. Weber, U. Peuchert, G. Roth, M. Weiden, and C. Geibel, “NaV2O5 as a quarter-filled ladder compound,” Physical Review Letters 80, 5164 (1998).
  29. [29]M. N. Popova, A. B. Sushkov, S. A. Klimin, E. P. Chukalina, B. Z. Malkin, M. Isobe, and Y. Ueda, “Lattice vibrations of α′-NaV2O5 in the low-temperature phase: An alternative interpretation to magnetic bound states,” Physical Review B 65, 144303 (2002).
  30. [30]M. N. Popova, A. B. Sushkov, E. P. Chukalina, E. A. Romanov, M. Isobe, and Y. Ueda, “Charge ordering and lattice dimerization in α′-NaV2O5: One or two phase transitions?” Physics of the Solid State 44, 1450–1454 (2002).
  31. [31]M. J. Konstantinović, J. C. Irwin, M. Isobe, and Y. Ueda, “Low-energy excitations in NaV2O5,” Physical Review B 65, 012404 (2001).DOI
  32. [32]T. Suemoto, M. Nakajima, K. Kazumi, H. Nagase, M. Aiba, M. Isobe, and Y. Ueda, “Ultrafast dynamics of photoinduced phenomena in the spin ladder system NaV2O5,” Journal of Physics: Conference Series 21, 15 (2005).
  33. [33]M. Nakajima, K. Kazumi, M. Isobe, Y. Ueda, and T. Suemoto, “Time-resolved raman spectroscopy applied to the photoinduced phenomena in NaV2O5,” Journal of Physics: Conference Series 21, 201 (2005).
  34. [34]S. Luther, H. Nojiri, M. Motokawa, M. Isobe, and Y. Ueda, “Direct observation of the spin gap in NaV2O5 by submillimeter wave esr,” journal of the physical society of japan 67, 3715–3717 (1998).
  35. [35]E. Y. Sherman, M. Fischer, P. Lemmens, P. H. M. Van Loosdrecht, and G. Güntherodt, “Electron-phonon and spin-phonon coupling in NaV2O5: Charge fluctuations effects,” Europhysics Letters 48, 648 (1999).
  36. [36]M. Isobe, C. Kagami, and Y. Ueda, “Crystal growth of new spin-peierls compound NaV2O5,” Journal of crystal growth 181, 314–317 (1997).
  37. [37]See the Supplemental Material for additional details on the experimental setup, the analytical models based on nonlinear phononics and nonlinear photonics, and discussion on symmetry selection rules for nonlinear phononics excitation [38–47].
  38. [38]F. Giorgianni, J. Sakai, and S. Lupi, “Overcoming the thermal regime for the electric-field driven mott transition in vanadium sesquioxide,” Nature Communications 10 (2019), 10.1038/s41467-019-09137-6.
  39. [39]C. Vicario, A. Trisorio, S. Allenspach, C. Ruegg, and F. Giorgianni, “Narrow-band and tunable intense terahertz pulses for mode-selective coherent phonon excitation,” Applied Physics Letters 117, 101101 (2020).
  40. [40]F. Blanchard, G. Sharma, X. Ropagnol, L. Razzari, R. Marondotti, and T. Ozaki, “Improved terahertz two-color plasma sources pumped by high intensity laser beam,” Opt. Express 17, 6044–6052 (2009).
  41. [41]G. Khalsa, Nicole A. Benedek, and J. Moses, “Ultrafast control of material optical properties via the infrared resonant raman effect,” Phys. Rev. X 11, 021067 (2021).
  42. [42]T. Huber, M. Ranke, A. Ferrer, L. Huber, and S. L. Johnson, “Coherent phonon spectroscopy of non-fully symmetric modes using resonant terahertz excitation,” Applied Physics Letters 107 (2015).
  43. [43]O. Bistoni, P. Barone, E. Cappelluti, L. Benfatto, and F. Mauri, “Giant effective charges and pyroelectricity in gapped graphene,” 2D Materials 6, 045015 (2019).
  44. [44]M.J. Konstantinović, Z.V. Popović, A.N. Vasil’ev, M. Isobe, and Y. Ueda, “First evidence for charge ordering in naV2O5 from Raman spectroscopy,” Solid State Communications 112, 397–402 (1999).
  45. [45]K. Ohwada, Y. Fujii, Y. Katsuki, J. Muraoka, H. Nakao, Y. Murakami, H. Sawa, E. Ninomiya, M. Isobe, and Y. Ueda, “Charge-order pattern of the low-temperature phase from a monoclinic single domain of nav2o5 uniquely determined by resonant x-ray scattering,” Phys. Rev. Lett. 94, 106401 (2005).
  46. [46]A. Damascelli, C. Presura, D. van der Marel, J. Jegoudez, and A. Revcolevschi, “Optical spectroscopic study of the interplay of spin and charge in α′-NaV2O5,” Phys. Rev. B 61, 2535–2552 (2000).
  47. [47]B. D. Gaulin, M. D. Lumsden, R. K. Kremer, M. A. Lumsden, and H. Dabkowska, “Two dimensional ordering and fluctuations in α′ − naV2O5,” Phys. Rev. Lett. 84, 3446–3449 (2000).
  48. [48]H. Kamiotka, S. Saito, M. Isobe, Y. Ueda, and T. Suemoto, “Coherent magnetic oscillation in the spin ladder system α′-NaV₂O₅,” Physical review letters 88, 127201 (2002).
  49. [49]T. P. Martin and L. Genzel, “Ionic raman scattering and ionic frequency mixing,” physica status solidi (b) 61, 493–502 (1974).DOI
  50. [50]A. A. Melnikov, K. N. Boldyrev, Y. G. Selivanov, V. P. Martovitskii, S. V. Chekalin, and E. A. Ryabov, “Coherent phonons in a Bi₂Se₃ film generated by an intense single-cycle thz pulse,” Physical Review B 97, 214304 (2018).DOI
  51. [51]C. Presura, D. van der Marel, A. Damascelli, and R. K. Kremer, “Low-temperature ellipsometry of α′-NaV₂O₅,” Phys. Rev. B 61, 15762–15765 (2000).
  52. [52]A. Damascelli, D. van der Marel, M. Grüninger, C. Presura, T. T. M. Palstra, J. Jegoudez, and A. Revcolevschi, “Direct two-magnon optical absorption in α′-NaV₂O₅: “charged” magnon,” Phys. Rev. Lett. 81, 918–921 (1998).DOI
  53. [53]R. Merlin, “Generating coherent thz phonons with light pulses,” Solid State Communications 102, 207–220 (1997), highlights in Condensed Matter Physics and Materials Science.DOI

Paper details

vlnc:W_35satd2y
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