Letter pubs.acs.org/journal/apchd5
Terahertz Radiation by Subpicosecond Magnetization Modulation in the Ferrimagnet LiFe5O8 Yuto Kinoshita,† Noriaki Kida,*,† Masato Sotome,† Tatsuya Miyamoto,† Yusuke Iguchi,‡ Yoshinori Onose,‡ and Hiroshi Okamoto† †
Department of Advanced Materials Science, University of Tokyo, 5-1-5 Kashiwa-no-ha, Chiba 277-8561, Japan Department of Basic Science, University of Tokyo, 3-8-1 Komaba, Tokyo 153-8902, Japan
Downloaded via DURHAM UNIV on July 28, 2018 at 10:48:32 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
‡
ABSTRACT: We report the terahertz radiation in the polar and ferrimagnetic insulating oxide LiFe5O8 upon irradiation of a femtosecond laser pulse at room temperature. By the resonant optical excitation of an originally spin-forbidden d−d transition in Fe3+ ions, we observed the emission of terahertz waves, which is attributable to the magnetic-dipole radiation originating from the light-induced subpicosecond magnetization modulation. By mapping out the emission of the terahertz waves, we successfully visualized the vector images of in-plane magnetic domains in magnetic fields. The terahertzradiation imaging presented here is useful as a new tool to observe ferromagnetic or ferrimagnetic domains. KEYWORDS: ferrimagnets, terahertz radiation imaging, magnetic domains, magnetization modulation
F
recognized as the dominant process of the terahertz radiation. On the other hand, in an insulating paramagnet, Tb3Ga5O12 garnet,15,16 irradiation of a circularly polarized femtosecond laser pulse induces a subpicosecond magnetization modulation via the inverse Faraday effect, which gives rise to light helicitydependent terahertz radiation. More recently, terahertz radiation originating from the magnetization modulation of Dy3+ ions via excitations of the f−f electronic transitions was observed in an insulating antiferromagnet DyFeO3.17 An alternative way to effectively induce the magnetization modulation is to optically excite an originally spin-forbidden d− d transition in solids, which includes a spin-reversal process and therefore necessarily causes a magnetization modulation. For this subject, we focus on a polar and ferrimagnetic insulating oxide, LiFe5O8 ferrite. At room temperature, the crystal structure of LiFe5O8 is a cubic spinel with a space group of P4332 (point group of 432), as illustrated in Figure 1a.18 According to the X-ray diffraction studies at 296 K,18 Fe3+ ions occupy six-coordinated octahedral or four-coordinated tetrahedral sites, and Li+ ions occupy six-coordinated octahedral sites, which are shown in red, blue, and green colors, respectively, in Figure 1a. Six-coordinated and four-coordinated Fe3+ ions in the high-spin state (S = 5/2) are ordered antiparallel with a 3:2 ratio,19 resulting in ferrimagnetism at room temperature. The magnetic easy axis is the [111]-direction of the crystal, and the ferrimagnetic transition temperature is 928 K.20 Left and right panels of Figure 1b show the schematics of the energy level diagrams of six-coordinated and four-coordinated Fe3+ ions,
erromagnets can be widely used in various information storage devices such as magnetoresistive random access memories and magneto-optical memories. The driving frequency of magnetization reversals is usually determined by a frequency of spin precessions, i.e., Larmor frequency, which is typically ∼100 GHz in ferromagnets. Recent development of the information-processing technology requires ultrafast manipulation of the magnetization at a frequency higher than 1 THz.1 In this context, ultrafast demagnetization upon irradiation of a femtosecond laser pulse has been intensively studied.1 In ferromagnetic metals such as Ni, a pump-and-probe magneto-optical measurement was used to detect the lightinduced demagnetization on the subpicosecond time scale,2−4 which is understood by the heating of the spin systems via the energy transfer from excited hot electrons.2,5 Such a subpicosecond magnetization modulation would generate terahertz electromagnetic waves via the magnetic-dipole radiation, which is expressed as E THz(t ) = −
μ0 ∂ 2 ⎛ r⎞ e × m ⎜t − ⎟ ⎝ 4πrc ∂t 2 c⎠
(1)
Here, ETHz is the radiated terahertz electric field, e is a unit vector directing from the position of the magnetic-dipole moment m to the detecting position, r is a distance between the position of m and the detecting position, μ0 is the permeability in a vacuum, and c is the velocity of light. So far, light-induced terahertz radiation has been observed not only in metallic magnets6−8 but also in semiconducting or insulating magnets including diluted magnetic semiconductors and antiferromagnetic insulators.9−14 In metallic magnets such as Ni6 and permalloy,8 the magnetic-dipole radiation mechanism is indeed © 2016 American Chemical Society
Received: April 15, 2016 Published: June 6, 2016 1170
DOI: 10.1021/acsphotonics.6b00272 ACS Photonics 2016, 3, 1170−1175
ACS Photonics
■
Letter
RESULTS AND DISCUSSION
Optical Spectra in Terahertz and Near-Infrared to Visible Regions. First, we show the optical spectra of LiFe5O8. By using standard terahertz time-domain spectroscopy (THzTDS),21 we obtain the transmission T spectrum in the terahertz frequency region, which is shown in Figure 1c. In this experiment, we use an 850 μm thick crystal. We also evaluate the refractive index n and absorption coefficient α spectra, which are shown in Figure 1d and e, respectively; detailed evaluation procedures in the THz-TDS are described in ref 22. As can be seen in Figure 1d and e, n is a nearly constant (∼4.3) and α is negligibly small (∼50 cm−1) below 2.5 THz. Next, we show in Figure 1f T and reflectance R spectra in the near-infrared to visible region (0.5−4 eV) by red and blue lines, respectively. In these measurements, we use a 110 μm thick crystal. We observe a sharp decrease of T and a gradual increase of R above 2 eV, which are due to the presence of the chargetransfer (CT) transition from O 2p to Fe 3d orbitals located at the higher energy side.23 Figure 1g and h show n and α spectra by blue lines, respectively; n and α spectra are derived by the application of the Kramers−Kronig transformation to the measured R spectrum. The shoulder structure is observed at around 3.2 eV, which is in agreement with the peak energy (∼3.17 eV) in the ellipsometry measurements previously reported.23 We reproduce the n spectrum below 2.5 eV by using the Sellmeier relationship given by n =
1+
S0λ 0 2 1 − (λ 0 / λ)2
as shown by the dotted pink line in Figure 1g. Here, λ is the wavelength, and the parameters are S0 = 6.98 × 1013 m−2 and λ0 = 247 nm. By using these S0 and λ0 values, we calculate the spectrum of the optical group index ng represented by
Figure 1. Schematics of (a) the crystal structure of LiFe5O818 and (b) the energy level of 3d electrons of Fe3+. (c) Transmittance T, (d) refractive index n, and (e) absorption coefficient α spectra in the terahertz region. (f) T and reflectance R spectra in the near-infrared to visible region. (g) n spectrum derived from Kramers−Kronig (KK) transformation with use of the measured R spectrum. The dotted pink line indicates the fitting result using the Sellmeier relationship, from which the spectrum of the optical group refractive index ng shown by the green line is obtained. (h) α spectra obtained from KK transformation and derived from measured T and R are shown by blue and red lines, respectively. All the measurements were done in the Eω ∥ (111)-plane at room temperature.
dn
n − λ dλ , which is shown by the green line in Figure 1g.
The value of ng for the femtosecond laser pulses with 1.55 eV (800 nm) (indicated by the vertical arrow) used in terahertz radiation experiments is ∼2.60. In the α spectrum (blue line) in Figure 1h, one can discern a shoulder structure at around 2.4 eV. According to the previous circular dichroism24 and linear Kerr effect25 measurements, it was attributed to the d−d transition (6A1g → 4A1g or 4Eg) of six-coordinated Fe3+ ions. Below 1.6 eV, we can derive the α spectrum directly from the 1
(
measured R and T data by using the relation α =− d ln
T (1 − R )2
)
where d is the sample thickness (110 μm). The evaluated α spectrum is shown in Figure 1h by the red line. Two shoulder structures are discerned at ∼1.3 and ∼1.6 eV. According to the Tanabe−Sugano diagram for 3d5 systems,26 these structures can be attributed to other d−d transitions such as 6A1g → 4T1g, 4 T2g of six-coordinated Fe3+ or 6A1 → 4T1, 4T2 of fourcoordinated Fe3+. Terahertz Radiation. Figure 2b shows the magnetization M as a function of the magnetic field H (M−H curve) of the crystal used in the terahertz radiation experiments. H is applied along the [110̅ ] direction of the crystal. Because of the isotropic nature of the magnetization within the (111) plane, the remanent magnetization is almost equal to zero and the saturation field is ±60 mT. In the terahertz radiation experiments, we use a 110 μm thick crystal, which is the same as that used in the optical measurements in the nearinfrared and visible regions mentioned above. The magnetic field H = ±60.5 mT (indicated by vertical arrows in the M−H curve in Figure 2b), which exceeds the saturation field, is
respectively; 3d-orbitals split into doubly degenerate eg (e) orbitals and triply degenerate t2g (t2) orbitals by the octahedral (tetrahedral) crystal fields. In the present study, we observe the emission of the terahertz waves in LiFe5O8 upon irradiation of the femtosecond laser pulses at room temperature. We find that the phase of the radiated terahertz wave is inverted when the direction of the magnetization is reversed. In addition, we reveal that the direction of the electric field of the terahertz waves does not depend on the polarization direction of the incident femtosecond laser pulses, but is perpendicular to the direction of the magnetization. These results suggest that the emission of the terahertz waves originates from the magnetic-dipole radiation via a light-induced subpicosecond magnetization modulation. Furthermore, we show that the magnitude and direction of the magnetization correspond to the measured amplitude and phase of the terahertz waves, respectively. With use of this oneto-one correspondence, we propose a new scheme for visualizing magnetic domains in magnetic fields. 1171
DOI: 10.1021/acsphotonics.6b00272 ACS Photonics 2016, 3, 1170−1175
ACS Photonics
Letter
second pulse. The calculated Δt is 2.8 ps, which is in good agreement with the interval of two pulses experimentally observed. Terahertz-Radiation Mechanism. In order to reveal the terahertz-radiation mechanism, we first measure the dependence of the terahertz electric-field amplitudes on the polarization direction (the azimuthal angle) of the incident femtosecond laser pulses. The experimental setup is schematically shown in Figure 3a. We apply a magnetic field H of 60.5 Figure 2. (a) Electric-field waveforms of terahertz radiation in LiFe5O8 for a magnetic field H of ±60.5 mT. Insets show schematics of terahertz radiation, which depends on the direction of the magnetization M. (b) M−H curve (solid line) of the same sample used in terahertz radiation experiments. Terahertz electric field at 0 ps as a function of H is shown by circles. Vertical arrows indicate H values used in terahertz radiation experiments shown in (a).
applied along the [11̅0] direction. Electric fields Eω of the incident femtosecond laser pulses are set parallel to the [11̅0] direction, while the electric fields ETHz parallel to the [112̅] direction of the radiated terahertz wave are detected, that is, Eω ⊥ ETHz, as schematically shown in the inset of Figure 2a. Figure 2a shows the measured terahertz electric-field waveforms for H = ±60.5 mT in which the time origin is arbitrary. A nearly single-cycle electric field pulse with a pulse width of ∼1 ps is detected in common. Noticeably, the phase of the radiated terahertz wave is inverted when the direction of M is reversed, as schematically shown in the insets of Figure 2a. The peak magnitude of the electric field is about 1/3000 as large as that of the terahertz wave radiated from the typical terahertz emitter: the (110)-oriented ZnTe crystal.21 In order to confirm the relationship between the terahertz radiation and M, we measured H dependence of the peak amplitudes of the electric field at 0 ps [ETHz(0)], which is shown by circles in Figure 2b. ETHz(0)−H curve is related to the M−H curve. This clearly indicates that the phase of ETHz(0) corresponds to the direction of M. In order to characterize the phase-matching condition between a femtosecond laser pulse and a radiated terahertz wave, we calculate the coherent length lc, which is expressed as
lc =
λ 1 2 |ng − n THz|
Figure 3. (a) Schematics of the experimental setup for azimuthal angle θ dependence of terahertz radiation in LiFe5O8. (b) Vector basis to obtain the polarization of the terahertz electric field ETHz from measured values. θ dependence of (c) angle φ of the emitted terahertz wave relative to the X-axis and (d) difference of the electric-field amplitude at 0 and 0.4 ps in the terahertz waveform along the X-axis (EX) and Y-axis (EY).
mT along the [110̅ ] direction of the crystal. We define the laboratory coordinate as X (horizontal) and Y (vertical) axes. We rotate the polarization direction of the incident femtosecond laser pulses using a half-wave plate; the angle θ is defined as the angle of the electric field of femtosecond laser pulses relative to the X-axis. In order to determine the X- and Yaxis components of the radiated terahertz waves, we perform vector analysis. In this analysis, we use two wire grid polarizers, WG1 and WG2, in front of the detector. With respect to the Xaxis, the angle of WG2 is set at 0° and the angle of WG1 is set at +45° or −45°, as illustrated in Figure 3a. Then, we detect the electric field of the corresponding terahertz wave, E+45° or E−45°. In this setup, the X- and Y-axis components of the terahertz waves, EX and EY, are expressed as follows.
(2)
Here, λ and nTHz are the wavelength and the refractive index of the terahertz waves, respectively. Using eq 2 with the measured ng (∼2.60) at 1.55 eV (Figure 1g) and nTHz (∼4.3) at 1 THz (Figure 1d), we obtain lc ≈ 90 μm. The penetration depth (1/ α) of the incident femtosecond laser pulses is 20.8 μm at 1.55 eV (Figure 1h). Thus, we can consider that the terahertz waves radiated within 20.8 μm depth from the crystal surface are transmitted through the residual part of the crystal and then detected. In Figure 2a, the second pulse is discerned at ∼2.8 ps in addition to the single-cycle pulse at around 0 ps. A possible origin of the second pulse is a multiple reflection of the generated terahertz wave between the top and back sample/air interfaces. The multiple reflection of the incident femtosecond laser pulses can be neglected because of the short penetration depth (20.8 μm). By assuming the multiple reflection of the terahertz wave, we calculate the expected interval time Δt (=2nTHzd/c) between the single-cycle pulse at 0 ps and the
EX ∝ E+45 ° − E−45 °
(3)
EY ∝ E+45 ° + E−45 °
(4)
The relationship between EX, EY, E+45°, and E−45° is schematically shown in Figure 3b. Then, the angle φ between the electric field of the radiated terahertz wave and X-axis is given by φ = arctan(EY/EX). In our study, we use as a measure the difference of the electric-field amplitude at 0 ps and at 0.4 ps in the terahertz waveform shown in Figure 2a. The estimated EX 1172
DOI: 10.1021/acsphotonics.6b00272 ACS Photonics 2016, 3, 1170−1175
ACS Photonics
Letter
and EY values are shown as a function of θ in Figure 3d by red and blue circles, respectively. EX values are nearly equal to zero, while EY values are finite and almost constant. Accordingly, φ is ∼90° as shown in Figure 3c. In addition, EY and φ values are independent of θ. These results indicate that the polarization of the radiated terahertz waves is parallel to the Y-axis and perpendicular to the direction of the magnetization. On the basis of the results presented above, we discuss the terahertz-radiation mechanism in LiFe5O8. In noncentrosymmetric media such as ZnTe, the dominant mechanism of the light-induced terahertz radiation is optical rectification, in which a difference frequency generation occurs from the corresponding nonlinear polarization P(2).21,27 In this case, P(2) is described by the second-order nonlinear optical susceptibility tensor χ(2). LiFe5O8 belongs to the point group of 432 and has no inversion symmetry, so that it is polar. However, all the components of χ(2) are zero in the point group 432.28 Thus, the optical rectification due to the crystal nonlinearity is forbidden in LiFe5O8. Next, we consider the possibility of the optical rectification due to magnetic nonlinearity, which is recognized as one of the terahertz-radiation mechanisms in ferromagnetic metals.7 Since the magnetization is fixed to the [11̅0] direction, the magnetic point group is 22′2′.29 In this case, the χ(2) zzz component becomes nonzero.30 However, χ(2) zzz is not an origin for the observed terahertz radiation, because the θ-dependence characteristic of χ(2) zzz is not observed (Figure 3c). Thus, the terahertz radiation observed in LiFe5O8 cannot be explained by the optical rectification processes originating from the crystal or magnetic nonlinearity. As described above, we demonstrate that the electric field of the radiated terahertz wave is perpendicular to the direction of the magnetization (Figure 3c), and the phase of the radiated terahertz wave is inverted by the reversal of the magnetization direction (Figure 2). These results are well explained by the magnetization modulation expressed by eq 1. Thus, we can conclude that the mechanism of the observed terahertz radiation is magnetic-dipole radiation via the light-induced subpicosecond modulation of the magnetization. The excitation density Iex used in the study of LiFe5O8 is 4 × 10−6 photon/Fe atom. This value is 2 orders of magnitude lower than Iex (= 9 × 10−4 photon/Ni atom) used in terahertzradiation experiments on Ni,6 which is obtained by taking into account optical constants and penetration depth (80 nm) at 1.51 eV31 and pulse energy used in terahertz-radiation experiments (1 mJ/cm2).6 Thus, thermally induced demagnetization is unlikely to be an origin of the terahertz radiation. A plausible microscopic origin of subpicosecond modulation of the magnetization is the resonant excitation of the d−d transition. As mentioned above, the absorption of the d−d transition in Fe3+ ions located at ∼1.5 eV (Figure 1h) can be assigned to either 6A1g → 4T1g, 4T2g of six-coordinated Fe3+ or 6 A1 → 4T1, 4T2 of four-coordinated Fe3+. In the ground state, one unpaired electron exists in all the 3d-orbitals, as schematically shown in Figure 1b, so that all d−d transitions include a reversal process of a spin. Therefore, the resonant excitation of a d−d transition should modify the magnetization, resulting in terahertz radiation according to eq 1. In order to discuss in more detail the dynamics of the magnetization after the initial spin flip by the d−d transition, further studies such as the measurements of the excitation photon-energy dependence of the terahertz-radiation efficiency and the analyses of the electric-field waveform of the terahertz radiation are necessary.
Magnetic Domain Imaging Using Emission of Terahertz Waves. In this section, we show the results of the magnetic domain imaging in LiFe5O8 using terahertz radiation. As seen in Figure 2, the phase of the terahertz electric field depends on the direction of the in-plane magnetization. This makes us expect that the in-plane magnetic domain can be visualized by mapping out the electric fields of the terahertz waves. Furthermore, the direction of the magnetization can be determined from the electric-field direction of the radiated terahertz wave evaluated with the vector-imaging technique illustrated in Figure 3a. This scheme has been used to visualize the ferroelectric domains and polarization directions in several kinds of ferroelectrics.32−34 Figure 4a shows the optical image of the crystal; a flat surface area in the crystal, which is extracted from the whole image, is
Figure 4. (a) Optical image of the LiFe5O8 crystal. The flat surface region of the crystal is surrounded by dotted lines. (b) Terahertz radiation images for a magnetic field H of ±60.5 mT. (c) Terahertz vector images for an H of ±60.5 mT. The direction and length of each arrow indicate the direction and magnitude of the magnetization, respectively.
surrounded by dotted lines. In the terahertz-radiation imaging experiments, θ is set to 0° and H = ±60.5 mT is applied along the [110̅ ] direction of the crystal. We measured E+45° and E−45° at various positions of the sample by a raster scan. The spatial resolution of the experiments is determined by the spot diameter of the incident femtosecond laser pulses. In the present experimental setup, the spot diameter is 25 μm. We show in the upper and lower panels of Figure 4b the terahertzradiation E+45° and E−45° images, respectively, which are obtained from the difference of the electric-field amplitude at 0 ps and at 0.4 ps in the terahertz waveform. The left and right panels show the results for H = −60.5 and 60.5 mT, respectively. The scale bar represents the magnitudes of the electric fields. Using these four images, we perform the vector analyses with use of eqs 3 and 4 and evaluate the magnitudes of EX and EY, from which vector images can be obtained. Since the direction of the magnetization is perpendicular to that of the radiated terahertz electric fields (Figure 3c), we rotate EX and EY vector images by 90°, which are shown in Figure 4c and d for H = −60.5 and 60.5 mT, respectively. The direction and length of each vector exhibit the direction and magnitude of the magnetization, respectively. The magnetization is almost 1173
DOI: 10.1021/acsphotonics.6b00272 ACS Photonics 2016, 3, 1170−1175
ACS Photonics
Letter
(20.8 μm), the excitation density is estimated to be 4 × 10−6 photon/Fe atom. The electromagnet was used to apply the magnetic field of 60.5 mT along the [11̅0] direction of the crystal, which reaches the saturation of the magnetization (see Figure 2b). In this experiment, the [110̅ ] direction of the crystal was set parallel to the X-axis (horizontal). The electric field of the femtosecond laser pulse was rotated by a half-wave plate with the angle θ; θ was defined by the angle of the polarization of the femtosecond laser pulse relative to the X-axis. Using two wire grids, we detected X- and Y-axis components of the radiated terahertz waves in transmission geometry by the standard photoconducting sampling technique with the LTGaAs detector.21 In the terahertz-radiation imaging experiments, we monitored the peak-to-peak amplitude of the terahertz wave at 0 and 0.4 ps [ETHz(0.4) − ETHz(0)] and obtained images by measuring ETHz(0.4) − ETHz(0) at each position by the raster scan method. The sample holder was attached to a twodimensional moving stage and moved along the X- and Y-axes. Further experimental details are described in refs 32−34. All the experiments were performed at room temperature.
directed in the [110̅ ] direction, i.e., M || [110̅ ]; slight variations of tilting angles from the [11̅0] direction (within 10°) are considered to appear from the inaccuracy of the direction of the applied magnetic field H. We discern no magnetization at the edge areas of the crystal. This comes from the fact that the edge areas are not normal to the surface, resulting in the decrease of the terahertz wave. So far, various methods have been developed to visualize the magnetic domains.35 Among them, magneto-optical means with the use of the Faraday (Kerr) effect due to the presence of the off-diagonal components of the dielectric constant are widely used for the magnetic-domain imaging over a wide area of a sample. However, it is difficult to determine the direction of the in-plane magnetization, i.e., the magnetization perpendicular to the light-k vector. On the other hand, the terahertz-radiationimaging method presented here can easily detect the orientations of in-plane magnetizations and the presence of magnetic domains by measuring the phase information on the terahertz waves.
■
CONCLUSION We successfully detected the emission of the terahertz waves in a polar and ferrimagnetic insulating oxide, LiFe5O8 ferrite, by the irradiation of a femtosecond laser pulse at 1.55 eV (800 nm). On the basis of the measurements of magnetic-field and azimuthal angle dependences of the emitted terahertz waves, we concluded that the terahertz-radiation mechanism is the magnetic-dipole radiation via light-induced subpicosecond magnetization modulation. Furthermore, we developed a new method for visualizing magnetic domains by mapping out the emission of the terahertz waves. This method can easily determine the magnetization vector of the crystal and, thus, would be useful to characterize the magnetic properties of various magnets.
■
AUTHOR INFORMATION
Corresponding Author
*E-mail (N. Kida):
[email protected]. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS This work was partly supported by a Grant-in-Aid by MEXT (Nos. 25247058, 25600072, and 15K13330). Y.K. and M.S. were supported by Japan Society for the Promotion of Science (JSPS) through Program for Leading Graduate Schools (MERIT) and JSPS Research Fellowships for Young Scientists.
■
■
METHODS (111)-Oriented single crystals of LiFe5O8 were grown by the flux method, details of which were reported elsewhere.36 We used two samples. One is a 110 μm thick single crystal for the measurements of terahertz radiation and optical spectroscopy in the near-infrared to visible region (0.5−4 eV). The other is an 850 μm thick single crystal for terahertz time-domain spectroscopy. The surfaces of all the samples were polished with Al powder. We measured the reflectance and transmittance spectra in the energy region 0.5−4 eV using a grating monochromator. In the terahertz frequency region, we measured the transmittance spectrum using standard terahertz time-domain spectroscopy.21 The terahertz emitter was a 2 mm thick (110)oriented ZnTe crystal, and the terahertz detector was a lowtemperature-grown GaAs (LT-GaAs). From the measured transmittance spectrum, we numerically calculated the refractive index and absorption coefficient spectra; details of our calculation procedure were reported in ref 22. In the terahertz-radiation experiments, we used femtosecond laser pulses delivered from a mode-locked Ti:sapphire laser (photon energy of 1.55 eV, repetition rate of 80 MHz, and pulse width of 100 fs). The laser pulses were focused on the sample at normal incidence with a spot diameter of 25 μm. The laser power was fixed at 55 mW, which corresponds to ∼100 μJ/cm2 per pulse. By taking into account the optical constants shown in Figure 1g and h and the penetration depth at 1.55 eV
REFERENCES
(1) Kirilyuk, A.; Kimel, A. V.; Rasing, T. Ultrafast optical manipulation of magnetic order. Rev. Mod. Phys. 2010, 82, 2731−2784. (2) Braurepaire, E.; Merle, J.-C.; Daunois, A.; Bigot, J.-Y. Ultrafast spin dynamics in ferromagnetic Nickel. Phys. Rev. Lett. 1996, 76, 4250−4253. (3) Koopmans, B.; van Kampen, M.; Kohlhepp, J. T.; de Jonge, W. J. M. Ultrafast magneto-optics in nickel: Magnetism or optics? Phys. Rev. Lett. 2000, 85, 844−847. (4) Hohlfeld, J.; Matthias, E.; Knorren, R.; Bennemann, K. H. Nonequilibrium magnetization dynamics of nickel. Phys. Rev. Lett. 1997, 78, 4861−4864. (5) Bigot, J.-V.; Vomir, M.; Beaurepaire, E. Coherent ultrafast magnetism induced by femtosecond laser pulses. Nat. Phys. 2009, 5, 515−520. (6) Beaurepaire, E.; Turner, G. M.; Harrel, S. M.; Beard, M. C.; Bigot, J.-Y.; Schmuttenmaer, C. A. Coherent terahertz emission from ferromagnetic films excited by femtosecond laser pulses. Appl. Phys. Lett. 2004, 84, 3465−3467. (7) Hilton, D. J.; Averitt, R. D.; Meserole, C. A.; Fisher, G. L.; Funk, D. J.; Thompson, J. D.; Taylor, A. J. Terahertz emission via ultrashortpulse excitation of magnetic metal films. Opt. Lett. 2004, 29, 1805− 1807. (8) Shen, J.; Fan, X.; Chen, Z.; Decamp, M. F.; Zhang, H.; Xiao, J. Q. Damping modulated terahertz emission of ferromagnetic films excited by ultrafast laser pulses. Appl. Phys. Lett. 2012, 101, 072401. (9) Nishitani, J.; Kozuki, K.; Nagashima, T.; Hangyo, M. Terahertz radiation from coherent antiferromagnetic magnons excited by femtosecond laser pulses. Appl. Phys. Lett. 2010, 96, 221906.
1174
DOI: 10.1021/acsphotonics.6b00272 ACS Photonics 2016, 3, 1170−1175
ACS Photonics
Letter
(10) Higuchi, T.; Kanda, N.; Tamaru, H.; Gonokami, M. K. Selection rules for light-induced magnetization of a crystal with threefold symmetry: the case of antiferromagnetic NiO. Phys. Rev. Lett. 2011, 106, 047401. (11) Heróux, J. B.; Ino, Y.; Gonokami, M. K.; Hashimoto, Y.; Katsumoto, S. Terahertz radiation emission from GaMnAs. Appl. Phys. Lett. 2006, 88, 221110. (12) Rungsawang, R.; Perez, R. F.; Oustinov, D.; Gómez, J.; Kolkovsky, V.; Karczewski, G.; Wojtowicz, T.; Madéo, J.; Jukam, N.; Dhillon, S.; Tignon, J. Terahertz radiation from magnetic excitations in diluted magnetic semiconductors. Phys. Rev. Lett. 2013, 110, 177203. (13) Zhan, H.; Deibel, J.; Laib, J.; Sun, C.; Kono, J.; Mittleman, D. M.; Munekata, H. Temperature dependence of terahertz emission from InMnAs. Appl. Phys. Lett. 2007, 90, 012103. (14) Kida, N.; Tonouchi, M. Terahertz radiation from magnetoresistive Pr0. 7Ca0. 3MnO3 thin films. Appl. Phys. Lett. 2001, 78, 4115− 4117. (15) Gorelov, S. D.; Mashkovich, E. A.; Tsarev, M. V.; Bakunov, M. I. Terahertz Cherenkov radiation from ultrafast magnetization in terbium gallium garnet. Phys. Rev. B: Condens. Matter Mater. Phys. 2013, 88, 220411. (16) Bakunov, M. I.; Mikhaylovskiy, R. V.; Bodrov, S. B. Probing ultrafast optomagnetism by terahertz Cherenkov radiation. Phys. Rev. B: Condens. Matter Mater. Phys. 2012, 86, 134405. (17) Mikhaylovskiy, R. V.; Huisman, T. J.; Popov, A. I.; Zvezdin, A. K.; Rasing, T.; Pisarev, R. V.; Kimel, A. V. Terahertz magnetization dynamics induced by femtosecond resonant pumping of Dy3+ subsystem in the multisublattice antiferromagnet DyFeO3. Phys. Rev. B: Condens. Matter Mater. Phys. 2015, 92, 094437. (18) Smolentsev, A. I.; Meshalkin, A. B.; Podberezskaya, N. V.; Kaplun, A. B. Refinement of LiFe5O8 crystal structure. J. Struct. Chem. 2008, 49, 953−956. (19) White, G. O.; Patton, C. E. Magnetic properties of lithium ferrite microwave materials. J. Magn. Magn. Mater. 1978, 9, 299−317. (20) Rezlescu, N.; Rezlescu, L.; Craus, M. L.; Rezlescu, E. LiFe5O8 and BaFe12O19 fine particles crystallised in a glassy matrix. Cryst. Res. Technol. 1999, 7, 829−836. (21) Tonouchi, M. Cutting-edge terahertz technology. Nat. Photonics 2007, 1, 97−105. (22) Sotome, M.; Kida, N.; Takeda, R.; Okamoto, H. Terahertz radiation induced by coherent phonon generation via impulsive stimulated Raman scattering in paratellurite. Phys. Rev. A: At., Mol., Opt. Phys. 2014, 90, 033842. (23) Moskvin, A. S.; Pisarev, R. V. Optical spectroscopy of charge transfer transitions in multiferroic manganites, ferrites, and related insulators. Low Temp. Phys. 2010, 36, 489−510. (24) Gridnev, V. N.; Krichevtsov, B. B.; Pavlov, V. V.; Pisarev, R. V. Magnetization-odd nonreciprocal reflection of light from the magnetoelectric-ferromagnet LiFe5O8. JETP Lett. 1997, 65, 68−73. (25) Visnovsky, S.; Krishnan, R. Polar Kerr rotation spectra in yttrium iron garnet and lithium ferrite: A comparative study. Appl. Phys. 1979, 18, 243−247. (26) Tanabe, Y.; Sugano, S. On the absorption spectra of complex ions II. J. Phys. Soc. Jpn. 1954, 9, 766−779. (27) Bass, M.; Franken, P. A.; Ward, J. F.; Weinreich, G. Optical rectification. Phys. Rev. Lett. 1962, 9, 446−448. (28) Shen, Y. R. The Principles of Nonlinear Optics; Wiley,: New York, 1984. (29) Mercier, M.; Velleaud, G.; Puvinel, J. Second order magnetoelectric effect in LiFe5O8. Physica B+C 1977, 86−88B, 1089−1090. (30) Birss, R. R. Macroscopic symmetry in space-time. Rep. Prog. Phys. 1963, 26, 307−360. (31) Johnson, P. B.; Christy, R. W. Optical constants of transition metals: Ti, V, Cr, Mn, Fe, Co, Ni, and Pd. Phys. Rev. B 1974, 9, 5056− 5070. (32) Sotome, M.; Kida, N.; Horiuchi, S.; Okamoto, H. Visualization of ferroelectric domains in a hydrogen-bonded molecular crystal using emission of terahertz radiation. Appl. Phys. Lett. 2014, 105, 041101.
(33) Kinoshita, Y.; Kida, N.; Sotome, M.; Takeda, R.; Abe, N.; Saito, M.; Arima, T.; Okamoto, H. Visualization of ferroelectric domains in boracite using emission of terahertz radiation. Jpn. J. Appl. Phys. 2014, 53, 09PD08. (34) Sotome, M.; Kida, N.; Horiuchi, S.; Okamoto, H. Terahertz radiation imaging of ferroelectric domain topography in roomtemperature hydrogen-bonded supramolecular ferroelectrics. ACS Photonics 2015, 2, 1373−1383. (35) Hubert, A.; Schafer, R. Magnetic Domains: the Analysis of Magnetic Microstructures; Springer: Berlin, 2008. (36) Beregi, E.; Sterk, E.; Pál, E.; Farkas-Jahnke, M. Crystal defects in flux grown lithium ferrite, LiFe5O8 single crystals. Acta Phys. Acad. Sci. Hung. 1979, 47, 263−273.
1175
DOI: 10.1021/acsphotonics.6b00272 ACS Photonics 2016, 3, 1170−1175