JP6780985B2 - Terahertz wave generator and terahertz wave generation method - Google Patents

Description

The present invention relates to a terahertz wave generator and a terahertz wave generation method.

Non-Patent Document 1 describes a technique for enhancing a narrow-band terahertz wave radiated from a photoconducting antenna by optical pulse shaping. Further, Non-Patent Document 2 describes a technique for enhancing a narrow-band terahertz wave by combining a multi-optical pulse and a large-diameter photoconducting antenna. This non-patent document 2 describes the saturation and relaxation time of carriers in a light conducting antenna.

Yongqian Liu et. Al., “Enhancement of narrow-band terahertzradiation from photoconducting antennas by optical pulse shaping”, Optics Letters, Vol.21, No.21, pp.1762-1764 (1996) Sang-Gyu Park et.al., “High-Power Narrow-Band Terahertz Generation Using Large-Aperture Photoconductors”, IEEE Journal of Quantum Electronics, Vol.35, No.8, pp.1257-1268 (1999) Marc M. Wefers et. Al., “Space-Time Profiles of Shaped UltrafastOptical Waveforms”, IEEE Journal of Quantum electronics, vol.32, No.1, pp.161-172 (1996) Tingting Qi et. Al., “Collective Coherent Control: Synchronization of Polarization in Ferroelectric PbTiO3by Shaped THz Fields”, Physical Review Letters 102, 247603 (2009) K. Yamaguchi et. Al., “Terahertz Time-Domain Observation of SpinReorientation in Orthoferrite ErFeO3 through Magnetic Free InductionDecay”, Physical Review Letters, 110, 137204 (2013) Shayne M. Harrel et. Al., “Influence of free-carrier absorption onterahertz generation from ZnTe (110)”, Journal of applied physics, 107, 033526 (2010) T. Kampfrath et. Al., “Coherent terahertz control ofantiferromagnetic spin waves”, Nature Photonics, Vol.5, pp.31-34 (2011) M. Hacker, G. Stobrawa, T. Feurer, “Iterative Fourier transform algorithm for phase-only pulse shaping”, Optics Express, Vol. 9, No. 4, pp.191-199, 13 August 2001 Olivier Ripoll, Ville Kettunen, Hans Peter Herzig, “Review ofiterative Fourier transform algorithms for beam shaping applications”, OpticalEngineering, Vol. 43, No. 11, pp.2549-2556, November 2004

Conventionally, there has been known a technique of irradiating a terahertz wave generating element such as a photoconducting antenna with a single light pulse to generate a wide band terahertz wave. When an optical pulse is applied to a photoconducting antenna, the generated carrier propagates between the antennas, causing an instantaneous current to flow and generating a terahertz wave. However, if the energy of the optical pulse becomes too high, the number of carriers generated increases, and the space charge effect causes carrier repulsion (hereinafter, such a phenomenon is referred to as carrier saturation), and the generation efficiency of terahertz waves. Will go down. To solve such a problem, it is conceivable to divide the optical pulse in time. Since the carrier saturation state in the terahertz wave generating element is relaxed with the passage of time, if a plurality of optical pulses are sequentially irradiated at time intervals, the terahertz wave can be generated while avoiding carrier saturation. It becomes possible to output a strong terahertz wave (see, for example, Non-Patent Documents 1 and 2). However, according to the experiments of the present inventor, it has been found that it is difficult to sufficiently avoid carrier saturation simply by dividing the optical pulse in time.

An object of the present invention is to provide a terahertz wave generator and a terahertz wave generation method capable of more effectively avoiding carrier saturation in a terahertz wave generating element.

In order to solve the above-mentioned problems, the terahertz wave generator according to one form is optically coupled to the light source that outputs the first light pulse and the light source, and the center wavelength has a time difference from the first light pulse. A multi-pulse generator that generates a plurality of second light pulses that are different from each other and outputs each of the plurality of second light pulses on different optical paths according to the wavelength, and a multi-pulse generator that is optically coupled to the multi-pulse generator to generate a multi-pulse. It includes a terahertz wave generating element that generates a terahertz wave pulse by receiving a plurality of second optical pulses output from a generation unit.

Further, in the terahertz wave generation method according to one form, a plurality of second light pulses having a time difference from each other and having different center wavelengths are generated from the first light pulse, and each of the plurality of second light pulses is generated according to the wavelength. It includes a step of outputting on different optical paths (multi-pulse light generation step) and a step of inputting a plurality of second light pulses to a terahertz wave generating element to generate a terahertz wave pulse (terahertz wave generation step).

In these terahertz wave generators and terahertz wave generation methods, the first optical pulse output from the light source is time-divided, and a plurality of second optical pulses having a time difference from each other are generated. The central wavelengths of these second light pulses are different from each other, and they are output on different optical paths depending on the wavelength. As a result, in the terahertz wave generating element, the plurality of second light pulses are irradiated to different positions from each other. Therefore, the degree to which the carriers generated by the irradiation of the second light pulse affect the saturation of the carriers when the other second light pulses are irradiated, as compared with the case where a plurality of second light pulses are irradiated at the same position. Is small. That is, according to the above-mentioned terahertz wave generator and terahertz wave generation method, it is possible to more effectively avoid carrier saturation in the terahertz wave generation element.

In the above-mentioned terahertz wave generator, the multi-pulse generator outputs a plurality of second optical pulses having a time difference from each other and having different center wavelengths on a common optical path, and a time waveform shaper. It may have an optical dispersion element that separates each of the plurality of second light pulses into different optical paths according to the wavelength. Similarly, in the above-mentioned terahertz wave generation method, the multi-pulse generation step includes a time waveform shaping step of outputting a plurality of second light pulses having a time difference from each other and having different center wavelengths on a common optical path, and a plurality of first light pulses. It may include a light dispersion step that separates each of the two light pulses into different optical paths depending on the wavelength. As a result, the above-mentioned multi-pulse generation unit and multi-pulse generation step can be preferably realized. In this case, the time waveform shaper (time waveform shaping step) is different from each other by performing phase modulation of the first light pulse after the spectroscopy with the spectroscopic element (spectral step) that disperses the first light pulse for each wavelength. A spatial optical modulator (optical modulation step) that generates a plurality of second light pulses having different central wavelengths and an optical system (condensing step) that condenses a plurality of second optical pulses. Good. As a result, a configuration for generating a plurality of second optical pulses can be realized with a compact and simple configuration. Further, the light dispersion element may be a diffraction grating or a prism. Thereby, each of the plurality of second light pulses having different center wavelengths can be suitably separated according to the wavelength.

In the above-mentioned terahertz wave generator, the multi-pulse generator has a time difference between the spectroscopic element that disperses the first optical pulse and the phase modulation of the first optical pulse after the spectroscopy for each wavelength, and has a central wavelength. Generates a plurality of second light pulses different from each other and separates the optical paths of the plurality of second light pulses in a direction intersecting the spectral direction according to a wavelength, and a plurality of second light pulses in the spectral direction. It may have an optical system that collects light pulses. Similarly, in the above-mentioned terahertz wave generation method, the multi-pulse generation step has a time difference between the spectroscopic step for dispersing the first optical pulse and the phase modulation of the first optical pulse after the spectroscopy for each wavelength. A photomodulation step that generates a plurality of second light pulses having different central wavelengths and separates the optical paths of the plurality of second light pulses in a direction intersecting the spectral direction according to the wavelength, and a plurality of light modulation steps in the spectral direction. It may include a focusing step of condensing the second light pulse. Even with such a configuration, the above-mentioned multi-pulse generation unit (multi-pulse generation step) can be preferably realized. In this case, the spatial light modulator may phase-modulate the first optical pulse after spectroscopy, at least based on the grating pattern. As a result, in the spatial light modulator, the optical paths of a plurality of second light pulses can be suitably separated according to the wavelength.

In the above-mentioned terahertz wave generator and terahertz wave generation method, the time width of each second optical pulse may be 100 femtoseconds or less. Terahertz waves can be efficiently generated by such ultrashort optical pulses.

According to the terahertz wave generator and the terahertz wave generation method according to the present invention, carrier saturation in the terahertz wave generating element can be avoided more effectively.

It is a figure which shows roughly the structure of the terahertz wave generator which concerns on 1st Embodiment. It is a figure which shows the structural example of the time waveform shaper. It is a figure which shows the modulation surface of a spatial light modulation element. (A) The spectral waveform of a single pulse-like single light pulse is shown. (B) The time intensity waveform of the single light pulse is shown. (A) The spectral waveform of the modulated light when the rectangular wave-shaped phase spectrum modulation is applied in the spatial light modulation element is shown. (B) The time intensity waveform of the obtained light pulse is shown. It is a graph which shows the example of the wavelength band of three optical pulses generated from the single optical pulse which has a central wavelength of 800 nm and a bandwidth of ± 50 nm. It is a figure which shows the optical path of the optical pulse in the vicinity of a diffraction grating and a lens. It is a top view which shows the light conduction antenna element as an example of a terahertz wave generation element. It is a partially enlarged view of FIG. It is a side sectional view along the X-ray line of FIG. It is a side sectional view along the line XI-XI of FIG. As another example of the terahertz wave generating element, it is a top view which shows the light conduction antenna element. It is a partially enlarged view of FIG. It is a flowchart which shows the terahertz wave generation method which concerns on 1st Embodiment. It is a figure which shows the structure of the data creation part roughly. It is a block diagram which shows the internal structure of a phase spectrum design part and an intensity spectrum design part. It is a figure which shows the calculation procedure of the phase spectrum by the iterative Fourier transform method. It is a figure which shows the calculation procedure of the phase spectrum function in a phase spectrum design part. It is a figure which shows the calculation procedure of the spectrum intensity in the intensity spectrum design part. It is a figure which shows an example of the generation procedure of the target spectrogram in the target generation part. It is a figure which shows an example of the procedure for calculating an intensity spectrum function. It is a figure which shows (a) (b) the process of making a target spectrogram. It is a graph which shows the center wavelength of two domains included in each spectrogram whose evaluation value satisfied a predetermined condition for each corresponding target spectrogram, and the center wavelength interval between domains. (A) It is a graph which shows the spectral waveform when the combination of the center wavelengths of two domains of a target spectrogram is (800 nm, 800 nm). (B) It is a graph which shows the time intensity waveform of the output light obtained by Fourier transforming the spectral waveform of (a). (A) It is a graph which shows the spectral waveform when the combination of the center wavelengths of two domains of a target spectrogram is (802 nm, 798 nm). (B) It is a graph which shows the time intensity waveform of the output light obtained by Fourier transforming the spectral waveform of (a). (A) It is a figure which shows the target spectrogram used in the 2nd Example. (B) A spectrogram calculated based on the target spectrogram of (a). (A) It is a figure which shows the target spectrogram used in the 2nd Example. (B) A spectrogram calculated based on the target spectrogram of (a). (A) It is a graph which shows the spectral waveform calculated from the spectrogram of FIG. 26 (b). (B) It is a graph which shows the time intensity waveform of the output light obtained by Fourier transforming the spectral waveform of (a). (A) It is a graph which shows the spectral waveform calculated from the spectrogram of FIG. 27 (b). (B) It is a graph which shows the time intensity waveform of the output light obtained by Fourier transforming the spectral waveform of (a). It is a figure which shows the 1st modification. It is a figure which shows the 2nd modification. It is a figure which shows the condensing position of four light pulses in the 3rd modification. It is a figure which shows roughly the part of the structure of the terahertz wave generator which concerns on 4th modification. As a specific configuration example of the pulse stretcher, it is a figure which shows the structure of the grating pair type. It is a figure for demonstrating 5th modification (a)-(c). (A), (b) It is a figure for demonstrating the 6th modification. (A), (b) It is a figure for demonstrating the 6th modification. (A), (b) It is a figure for demonstrating the 7th modification. (A) It is a side view which shows a part of the structure of the terahertz wave generator which concerns on 2nd Embodiment. (B) It is a top view which shows a part of the structure of the terahertz wave generator which concerns on 2nd Embodiment. It is a perspective view which shows a part of the structure of the terahertz wave generator which concerns on 2nd Embodiment. It is a figure which shows an example of the modulation pattern of the spatial light modulation element in 2nd Embodiment. It is a flowchart which shows the terahertz wave generation method which concerns on 2nd Embodiment.

Hereinafter, embodiments of the terahertz wave generator and the terahertz wave generation method according to the present invention will be described in detail with reference to the accompanying drawings. In the description of the drawings, the same elements are designated by the same reference numerals, and duplicate description will be omitted.

(First Embodiment)
FIG. 1 is a diagram schematically showing a configuration of a terahertz wave generator 1A according to a first embodiment of the present invention. As shown in FIG. 1, the terahertz wave generator 1A of the present embodiment includes a light source 3, a multi-pulse generator 10A, and a terahertz wave generator 30.

The light source 3 outputs a coherent single light pulse P1. This single optical pulse P1 is an independent optical pulse that is not continuous with other optical pulses in terms of time, and is the first optical pulse in the present embodiment. Further, the single light pulse P1 includes light having a plurality of wavelengths. The light source 3 is composed of a solid-state laser such as an ultrashort pulse laser. The wavelength of the single optical pulse P1 is selected according to the constituent material of the terahertz wave generating element 30 described later. For example, when the terahertz wave generating element 30 is a photoconducting antenna using low temperature growth GaAs, the center wavelength of the single optical pulse P1 is set to, for example, 800 nm in order to efficiently excite carriers. The spectrum width of the single optical pulse P1 is set so that a Fourier limit pulse of 100 femtoseconds or less can be obtained as the multi-optical pulse P2 described later. For example, when the center wavelength of the single light pulse P1 is 800 nm, the full width at half maximum (FWHM) of the spectrum needs to be wider than 10 nm in order to obtain a Fourier limit pulse of 100 femtoseconds.

The multi-pulse generator 10A is optically coupled to the light source 3. The multi-pulse generation unit 10A generates a multi-optical pulse P2 composed of a plurality of optical pulses P21, P22 and P23 from a single optical pulse P1. These optical pulses P21 to P23 are the second optical pulses in this embodiment. The center wavelengths of the optical pulses P21 to P23 are different from each other. The central wavelength of the optical pulse P21 is the longest, and the central wavelength of the optical pulse P23 is the shortest. Further, the optical pulses P21 to P23 have a time difference from each other. In the example shown in FIG. 1, the optical pulse P21 is output earliest and the optical pulse P23 is output the latest, but the output order of the optical pulses P21 to P23 is arbitrary. The time width of each optical pulse P21 to P23 is preferably 100 femtoseconds or less, for example.

The multi-pulse generation unit 10A spatially divides each of the optical pulses P21 to P23 and outputs them on different optical paths L1 to L3 according to the wavelength. In one example, the optical paths L1 to L3 are arranged parallel to each other. The number of optical pulses output from the multi-pulse generation unit 10A is not limited and can be various. However, in order to effectively excite the molecular vibration in the terahertz wave generating element 30, there is an appropriate number of pulses. For example, when switching the molecular vibration on / off, the number of pulses required is equal to the number of switchings.

The multi-pulse generator 10A of the present embodiment includes a time waveform shaper 11, a diffraction grating 18 (light dispersion element), and a lens 19. The time waveform shaper 11 is optically coupled to the light source 3 and outputs optical pulses P21 to P23 having a time difference from each other and having different center wavelengths on a common (single) optical path L0. Here, FIG. 2 is a diagram showing a configuration example of the time waveform shaper 11. As shown in FIG. 2, the time waveform shaper 11 includes a diffraction grating 12, a lens 13, a Spatial Light Modulator (SLM) 14, a lens 15, a diffraction grating 16, and a data creation unit 17.

The diffraction grating 12 is a spectroscopic element in the present embodiment and is optically coupled to the light source 3. The SLM 14 is optically coupled to the diffraction grating 12 via the lens 13. The diffraction grating 12 disperses the single light pulse P1 for each wavelength component. As the spectroscopic element, another optical component such as a prism may be used instead of the diffraction grating 12. Further, although the reflection type diffraction grating 12 is shown in FIG. 2, the diffraction grating 12 may be a transmission type. The single light pulse P1 is obliquely incident on the diffraction grating 12 and is separated into a plurality of wavelength components. The light Lb containing the plurality of wavelength components is focused by the lens 13 for each wavelength component and imaged on the modulation surface of the SLM 14. The lens 13 may be a convex lens made of a light transmitting member or a concave mirror having a concave light reflecting surface.

The SLM 14 is an example of a phase mask, and in order to generate optical pulses P21 to P23 having a time difference from each other and having different center wavelengths from the single optical pulse P1, phase modulation and intensity modulation of the optical Lb are performed at the same time. Further, only the phase modulation of the optical Lb may be performed, or only the intensity modulation of the optical Lb may be performed. The SLM 14 is, for example, a phase modulation type. In one embodiment, SLM14 is of type LCOS (Liquid crystal on silicon). FIG. 3 is a diagram showing a modulation surface 14a of the SLM 14. As shown in FIG. 3, a plurality of modulation regions 14b are arranged along a certain direction B1 on the modulation surface 14a, and each modulation region 14b is in a direction B2 intersecting (for example, orthogonal to) the direction B1. It is extending. The direction B1 is the spectral direction of the diffraction grating 12. The modulation surface 14a acts as a Fourier transform surface, and each of the plurality of modulation regions 14b is incident with each corresponding wavelength component after spectroscopy. The SLM 14 modulates the phase and intensity of each incident wavelength component independently of other wavelength components in each modulation region 14b. Since the SLM 14 of the present embodiment is a phase modulation type, the intensity modulation is realized by the modulation pattern presented on the modulation surface 14a.

The SLM 14 receives a control signal based on the phase pattern created by the data creation unit 17. The modulation pattern is data for controlling the SLM 14, and is data including the intensity of the complex amplitude distribution or the intensity of the phase distribution. The modulation pattern is, for example, Computer-Generated Holograms (CGH). The details of the data creation unit 17 will be described later.

Each wavelength component of the modulated light Lc modulated by the SLM 14 is collected by the lens 15 at a point on the diffraction grid 16. The lens 15 at this time functions as a condensing optical system that condenses the modulated light Lc. The lens 15 may be a convex lens made of a light transmitting member or a concave mirror having a concave light reflecting surface. Further, the diffraction grating 16 functions as a combined wave optical system, and combines each wavelength component after modulation. That is, the plurality of wavelength components of the modulated light Lc are focused and combined with each other by these lenses 15 and the diffraction grating 16, and become optical pulses P21 to P23 propagating on the same optical path.

The region before the lens 15 (spectral region) and the region after the diffraction grating 16 (time domain) are in a Fourier transform relationship with each other, and the phase modulation in the spectral region is converted into a time intensity waveform in the time domain. Affect. Therefore, the optical pulses P21 to P23 have a desired time intensity waveform according to the modulation pattern of the SLM14. Here, FIG. 4A shows, as an example, the spectral waveforms (spectral phase G11 and spectral intensity G12) of the single pulse-shaped single light pulse P1, and FIG. 4B shows the time of the single light pulse P1. The intensity waveform is shown. Further, FIG. 5A shows, as an example, the spectral waveforms (spectral phase G21 and spectral intensity G22) of the modulated light Lc when the rectangular wave-shaped phase spectrum modulation is applied in SLM14, and FIG. 5B shows FIG. The time intensity waveform of the obtained optical pulse is shown. In FIGS. 4 (a) and 5 (a), the horizontal axis indicates the wavelength (nm), the left vertical axis indicates the intensity value (arbitrary unit) of the intensity spectrum, and the right vertical axis indicates the phase value of the phase spectrum. (Rad) is shown. Further, in FIGS. 4 (b) and 5 (b), the horizontal axis represents time (femtoseconds) and the vertical axis represents light intensity (arbitrary unit). In this example, by applying a rectangular wave-shaped phase spectrum waveform to the modulated light Lc, the single pulse of the single light pulse P1 is converted into a double pulse accompanied by higher-order light. The spectrum and waveform shown in FIG. 5 is an example, and the number of optical pulses and the time intensity waveform can be variously changed by combining various phase spectra and intensity spectra.

FIG. 6 is a graph showing an example of the wavelength band of three optical pulses P21 to P23 generated from a single optical pulse P1 having a center wavelength of 800 nm and a bandwidth of ± 50 nm. The horizontal axis of FIG. 6 represents a wavelength (unit: nm), and the vertical axis represents an intensity (arbitrary unit). As shown in FIG. 6, the bandwidth of each optical pulse P21 to P23 is, for example, 20 nm, and its full width at half maximum (FWHM) is, for example, 10 nm. The central wavelengths of the light pulses P21 to P23 are, for example, 840 nm, 800 nm, and 760 nm, respectively. In this example, the shortest pulse width of each optical pulse P21 to P23 is approximately 94 femtoseconds.

The phase mask of the time waveform shaper 11 is not limited to SLM14, and various other phase masks can be used. For example, the phase mask may be an acousto-optic modulator. Further, as the phase mask, for example, a fixed type can be used in addition to the electrically controllable SLM and the acousto-optic modulator. Further, although the transmissive type SLM14 is shown in FIG. 2, the SLM14 may be a reflective type (for example, X13138-02 manufactured by Hamamatsu Photonics Co., Ltd.) in order to improve the light utilization efficiency.

Further, the diffraction grating 12, the lens 13, and the SLM 14 have a configuration suitable for the spectrum of the single optical pulse P1. For example, when the full width at half maximum of the spectrum of the single light pulse P1 is 100 nm, the number of lines of the diffraction grating 12 is 300 lines / mm, the focal length of the lens 13 is 181.5 mm, the number of pixels of the SLM 14 is 1280 × 1024, and the number of pixels of the SLM 14 is 1280 × 1024. The pixel pitch is 12.5 μm. Further, in order to increase the terahertz wave generation efficiency, dispersion compensation may be performed in the time waveform shaper 11 so that the pulse widths of the optical pulses P21 to P23 are the shortest in the terahertz wave generating element 30. For example, assuming that the variances of the lens 15 and the lens 19 are β ₁ and β ₂ , the compensation variance β = − (β ₁ + β ₂ ) is added to each optical pulse P21 to P23 in the time waveform shaper 11. May be good.

See FIG. 1 again. The optical pulses P21 to P23 output from the time waveform shaper 11 are input to the diffraction grating 18. FIG. 7 is a diagram showing the optical paths of the optical pulses P21 to P23 near the diffraction grating 18 and the lens 19. The diffraction grating 18 spatially divides each of the optical pulses P21 to P23 output from the time waveform shaper 11 and propagating on the common optical path L0 into optical paths L1 to L3 according to the wavelength. The diffraction grating 18 may be a reflective type or a transmissive type. The light diameter D of the light pulses P21 to P23 is, for example, 4 mm, and the number of lattices d of the diffraction grating 18 is, for example, 20 lines / mm. The diffraction angle θ in the diffraction grating 18 is expressed as θ = sin ^-1 (λ / d). λ is the center wavelength of each light pulse P21 to P23.

The light pulses P21 to P23 that have passed through the diffraction grating 18 are focused on the focusing spots Q1 to Q3 by the lens 19, respectively. The lens 19 is, for example, a condensing optical system composed of one or a plurality of lenses. Alternatively, the lens 19 may be configured by a cylindrical mirror. The optical distance between the diffraction grating 18 and the lens 19 is equal to the focal length f of the lens 19. At this time, the interval Δh between the focused spots Q1 to Q3 is expressed as Δh = f · tan (θ). When the focal length f is 50 mm, the interval Δh is 40 μm. On the other hand, the diameter of the focused spots Q1 to Q3 (the diameter of the beam, which is within the range of 1 / e ² when Gaussian approximation is applied. E is the natural logarithm) φ is expressed as φ = 4λf / πD, and in this case, it is approximately It is 13 μm. Therefore, the interval Δd between the focused spots Q1 to Q3 becomes wider than the diameter φ of the focused spots Q1 to Q3, and the focused spots Q1 to Q3 are surely separated from each other.

See FIG. 1 again. The terahertz wave generation element 30 is optically coupled to the multi-pulse generation unit 10A, receives optical pulses P21 to P23 output from the multi-pulse generation unit 10A, and generates a terahertz wave pulse. Here, FIG. 8 is a plan view showing a light conducting antenna element 30A as an example of the terahertz wave generating element 30. 9 is a partially enlarged view of FIG. 8, FIG. 10 is a side sectional view taken along line XX of FIG. 9, and FIG. 11 is a side sectional view taken along line XI-XI of FIG.

As shown in FIGS. 10 and 11, the photoconducting antenna element 30A includes a substrate 31, a first semiconductor layer 32 provided on the substrate 31, and an insulating layer 33 provided on the first semiconductor layer 32. It has a pair of electrodes 34 provided on the insulating layer 33, and a protective film 37 covering the insulating layer 33 and the electrodes 34. A second semiconductor layer 35 having an insulating property having a bandgap larger than that of the first semiconductor layer is provided between the substrate 31 and the first semiconductor layer 32. The substrate 31 is, for example, a semi-insulating GaAs substrate having a substantially rectangular shape. The first semiconductor layer 32 is a GaAs layer epitaxially grown at a low temperature (200 ° C. to 300 ° C.) by, for example, molecular beam epitaxy (MBE). The insulating layer 33 is, for example, a SiN layer. The second semiconductor layer 35 is, for example, an AlGaAs layer. The electrode 34 is an ohmic electrode and is formed of, for example, AuGe, Au, or the like.

As shown in FIG. 8, the insulating layer 33 has an opening 33a. The opening 33a is formed in the central portion of the insulating layer 33 and has a rectangular shape when viewed from the normal direction of the surface of the light conductive antenna element 30A. Further, each electrode 34 has a wiring portion 34a, a plurality of antenna portions 34b (three in the present embodiment), and a pair of pad portions 34c. The wiring portion 34a connects the plurality of antenna portions 34b and the pad portion 34c to each other. The plurality of antenna portions 34b are arranged in an array along a predetermined direction B3, and are integrally formed with the wiring portion 34a. The plurality of antenna portions 34b are located inside the opening 33a and are in contact with the first semiconductor layer 32 (see FIG. 10). The plurality of antenna portions 34b of one electrode 34 and the plurality of antenna portions 34b of the other electrode 34 are arranged so as to face each other at a predetermined interval. The pair of pad portions 34c of one electrode 34 are located at two corner portions along one side on the insulating layer 33, respectively. The pair of pad portions 34c of the other electrode 34 are located at two corner portions along the other side on the insulating layer 33, respectively. These pad portions 34c are electrically connected to external wiring.

As shown in FIG. 9, the antenna portion 34b projects from the wiring portion 34a toward the other antenna portion 34b. More specifically, the antenna portion 34b projects from the central portion of the wiring portion 34a in the extending direction in a direction substantially orthogonal to the extending direction. The tip of the antenna portion 34b of one electrode 34 and the tip of the antenna portion 34b of the other electrode 34 face each other. The above-mentioned optical pulses P21 to P23 are incident on the region between these tips (antenna gap, width 5 μm) to form focused spots Q1 to Q3. The pattern of the antenna portion 34b is a dipole pattern that exhibits a rectangular shape when viewed from the normal direction of the surface of the photoconducting antenna element 30A, and the photoconducting antenna element 30A is a dipole antenna array in which a plurality of such dipole patterns are arranged. is there.

When the light conduction antenna element 30A is irradiated with the light pulses P21 to P23, the generated carriers propagate between the pair of electrodes 34, so that an instantaneous current flows and a terahertz wave is generated. For example, when the time interval of the optical pulses P21 to P23 is 1 picosecond, it is possible to excite a molecule or spin vibration having a resonance of 1 THz, which is the reciprocal of the pulse time interval. Since the carrier generation position is divided for each optical pulse P21 to P23, it is possible to efficiently generate a terahertz wave while sufficiently avoiding carrier saturation. The distance between the adjacent antenna portions 34b needs to be, for example, several tens of μm or less. That is, the size of the region of the generation position of all carriers is not less than the wavelength of the terahertz wave (for example, 300 μm or less if the frequency of the terahertz wave is 1 THz). With such an interval, the terahertz waves output from the focused spots Q1 to Q3 are regarded as outputs from one point light source.

Although an example in which all the antenna portions 34b form a dipole antenna has been described in FIGS. 8 and 9, the type of at least one antenna portion may be different from that of the other antenna portions. For example, at least one antenna unit may be a bowtie antenna, a spiral antenna, or a comb-shaped antenna. Alternatively, at least two of these antenna types may be mixed. Since the obtained terahertz wave spectrum differs depending on the antenna type, the terahertz wave spectrum can be controlled to an arbitrary shape by selecting the antenna type of each antenna unit as necessary.

FIG. 12 is a plan view showing a light conducting antenna element 30B as another example of the terahertz wave generating element 30. Further, FIG. 13 is a partially enlarged view of FIG. The configuration of the photoconducting antenna element 30B is the same as that of the photoconducting antenna element 30A described above, except for the electrode shape.

The pair of electrodes 36 included in the photoconducting antenna element 30B has a wiring portion 36a, an antenna portion 36b, and a pair of pad portions 36c, respectively. The shapes of the wiring portion 36a and the pair of pad portions 36c are the same as those of the wiring portion 34a and the pair of pad portions 34c of the photoconducting antenna element 30A described above. The antenna portion 36b extends in a predetermined direction B3 with a constant width, and is integrally formed with the wiring portion 36a. The antenna portion 36b is located inside the opening 33a and is in contact with the first semiconductor layer 32. The antenna portion 36b of one electrode 36 and the antenna portion 36b of the other electrode 36 are arranged in parallel with each other and opposed to each other at a predetermined interval.

As shown in FIG. 9, the above-mentioned optical pulses P21 to P23 are incident on the region between the antenna portions 36b to form focused spots Q1 to Q3 arranged in the predetermined direction B3. Then, by applying a bias voltage (for example, amplitude ± 10 V, modulation frequency 1 kHz) between the pair of electrodes 36, a terahertz wave is generated from each focused spot Q1 to Q3. The size of the region of each focused spot Q1 to Q3 is equal to or smaller than the wavelength of the terahertz wave. Therefore, the terahertz waves output from the focused spots Q1 to Q3 can be regarded as outputs from the point light source, so that the terahertz waves can be efficiently generated. The optical conductive antenna element 30B having such a pattern of the antenna portion 36b is referred to as a stripline antenna.

In the above-mentioned example of the terahertz wave generating element 30, low-temperature growth GaAs was exemplified as the semiconductor material of the first semiconductor layer 32, but any semiconductor material having high terahertz wave generation efficiency such as Semi-insulating (SI) GaAs and InGaAs can be used. Just do it.

Further, the light intensities of the light pulses P21 to P23 reaching the terahertz wave generating element 30 may be equal to each other or different from each other. In particular, depending on the application of terahertz waves such as coherent control, the light intensities of the light pulses P21 to P23 may be different from each other. Further, the light intensity of each light pulse P21 to P23 may be arbitrarily controlled, such as maximizing the light intensity of the light pulse that first reaches the terahertz wave generating element 30 in terms of time.

Here, a method of generating a terahertz wave according to the present embodiment will be described. This terahertz wave generation method can be realized by using the terahertz wave generator 1A of the present embodiment. FIG. 14 is a flowchart showing the terahertz wave generation method of the present embodiment.

First, in the first step (pulse light generation step) S1, a single light pulse P1 is output from the light source 3. Next, in the second step (multi-pulse light generation step) S2, a plurality of light pulses P21 to P23 having a time difference from each other and having different center wavelengths are generated from the single light pulse P1, and each of the light pulses P21 to P23 is generated. , Outputs on different optical paths depending on the wavelength. The second step S2 includes a time waveform shaping step S3. In the time waveform shaping step S3, the optical pulses P21 to P23 are output on the common optical path L0. The time waveform shaping step S3 includes a spectroscopic step S4, a light modulation step S5, and a focusing step S6. In the spectroscopy step S4, the single light pulse P1 is separated. In the optical modulation step S5, the phase modulation of the single optical pulse P1 after spectroscopy is performed for each wavelength to generate a plurality of optical pulses P21 to P23 having a time difference from each other and having different center wavelengths. In the focusing step S6, the optical pulses P21 to P23 are focused.

The second step S2 further includes a light dispersion step S7. In the optical dispersion step S7, the optical pulses P21 to P23 are separated into different optical paths L1 to L3 depending on the wavelength. Finally, in the third step (terahertz wave generation step) S8, the optical pulses P21 to P23 are input to the terahertz wave generating element 30 to generate the terahertz wave pulse.

The effects obtained by the terahertz wave generator 1A and the terahertz wave generation method according to the present embodiment described above will be described. For example, in the terahertz wave generator described in Non-Patent Document 2, in order to suppress the saturation of carriers in the terahertz wave generating element, the focused spot diameter is increased to, for example, about 4 mm, so that the terahertz wave generation efficiency can be suppressed. It ends up. Further, since the multiple Michelson interference system is used as the generation method of a plurality of optical pulses, there is a problem that it is not easy to change the number of optical pulses and the time interval, and the apparatus becomes large-scale. Furthermore, it has been reported that the power of terahertz waves is saturated when the energy of the optical pulse is high. It is presumed that the reason for this is that carrier saturation has not been sufficiently avoided.

On the other hand, in the terahertz wave generator 1A and the terahertz wave generation method of the present embodiment, the single optical pulse P1 output from the light source 3 is time-divided, and a plurality of optical pulses P21 to P23 having a time difference from each other are generated. .. The central wavelengths of these optical pulses P21 to P23 are different from each other, and are output on different optical paths L1 to L3 depending on the wavelength. As a result, the optical pulses P21 to P23 are irradiated to different positions in the terahertz wave generating element 30. Therefore, as compared with the case where the light pulses P21 to P23 are irradiated at the same position, the degree to which the carriers generated by the irradiation of each light pulse affect the saturation of the carriers when the other light pulses are irradiated is small. That is, according to the terahertz wave generator 1A of the present embodiment, it is possible to more effectively avoid carrier saturation in the terahertz wave generating element 30 and improve the terahertz wave generation efficiency.

Further, according to the present embodiment, since the time interval of the optical pulses P21 to P23 can be freely controlled, the degree of freedom and efficiency of spectrum selection can be further increased. Further, by controlling the time interval of the optical pulses P21 to P23, saturation can be avoided regardless of the relaxation time of the carrier peculiar to the terahertz wave generating element 30, so that the degree of freedom of selection of the terahertz wave generating element 30 can be increased. it can. For example, it is possible to adopt a terahertz wave generating element using SI-GaAs, which has a long relaxation time of about 100 picoseconds.

Further, according to the present embodiment, the degree of freedom in the shape of the light conducting antenna can be increased. For example, when a plurality of light pulses are focused on one point, a dipole antenna is used to efficiently generate a terahertz wave, and a focusing spot is formed in the antenna gap of the dipole antenna. On the other hand, in the present embodiment, since the optical pulses P21 to P23 are separated temporally and spatially and collected, an antenna having antenna gaps at a plurality of locations can be used, for example, a comb-shaped antenna. A light conducting antenna having a complicated shape can be used.

Further, according to the present embodiment, since the generation efficiency of terahertz waves is increased, it can be applied to coherent control such as spin control. In recent years, experiments in which a substance such as NiO is irradiated with a terahertz wave to control spin have attracted attention (for example, Non-Patent Document 7). For example, two terahertz wave pulses can be irradiated, and spin off / off can be controlled by the pulse interval. If the terahertz wave generation efficiency is improved by this embodiment, the spin can be efficiently excited. Further, since the pulse interval can be easily controlled by the time waveform shaper 11, it can be applied to quantum arithmetic and the like.

Further, as in the present embodiment, the multi-pulse generation unit 10A (second step S2) is a time waveform shaper that outputs optical pulses P21 to P23 having a time difference from each other and having different center wavelengths on a common optical path L0. It has 11 (time waveform shaping step S3) and a diffraction grating 18 (light dispersion step S7) that separates the optical pulses P21 to P23 output from the time waveform shaper 11 into different optical paths according to the wavelength. May be good. Thereby, the multi-pulse generation unit 10A can be preferably realized. In this case, the time waveform shaper 11 (time waveform shaping step S3) performs phase modulation of the diffraction grating 12 (spectral step S4) that disperses the single light pulse P1 and the single light pulse P1 after the spectroscopy for each wavelength. This includes an SLM14 (optical modulation step S5) that generates optical pulses P21 to P23 having a time difference from each other and having different center wavelengths, and an optical system (condensing step S6) that concentrates the optical pulses P21 to P23. It may be. As a result, a configuration for generating optical pulses P21 to P23 can be realized with a compact and simple configuration. Further, by using the diffraction grating 18 as the optical dispersion element, the optical pulses P21 to P23 having different center wavelengths can be suitably separated according to the wavelength.

Further, as in the present embodiment, the time width of the optical pulses P21 to P23 may be 100 femtoseconds or less. Terahertz waves can be efficiently generated by such ultrashort optical pulses.

In the present embodiment, the optical system (diffraction grating 12 and lens 13) that guides the single light pulse P1 to SLM 14 and the optical system (lens 15 and diffraction grating 16) that generate a plurality of optical pulses P21 to P23 are separately separated. Although provided, these optical systems may be common. In that case, the SLM 14 is preferably of the reflective type.

Here, the data creation unit 17 of the time waveform shaper 11 will be described in detail. As described above, the time waveform shaper 11 outputs a plurality of optical pulses having a time difference from each other and having different center wavelengths. The data creation unit 17 calculates a modulation pattern displayed on the SLM 14 in order to generate such a plurality of optical pulses, and provides the SLM 14 with data on the modulation pattern.

FIG. 15 is a diagram schematically showing the configuration of the data creation unit 17. The data creation unit 17 of the present embodiment is, for example, a personal computer; a smart device such as a smartphone or a tablet terminal; or a computer having a processor such as a cloud server. The data creation unit 17 is electrically connected to the SLM 14, calculates a phase modulation pattern for controlling the number of pulses of the optical pulse and the time intensity waveform, and provides the SLM 14 with a control signal SC including the phase modulation pattern. To do. The data creation unit 17 of the present embodiment outputs a phase pattern for phase modulation that gives the output light a phase spectrum for obtaining a desired waveform including a plurality of optical pulses, and an intensity spectrum for obtaining a desired waveform. A phase pattern including a phase pattern for intensity modulation given to the SLM 14 is presented to the SLM 14. Therefore, the data creation unit 17 includes an arbitrary waveform input unit 21, a phase spectrum design unit 22, an intensity spectrum design unit 23, and a modulation pattern generation unit 24. That is, the computer processor provided in the data creation unit 17 has the functions of the arbitrary waveform input unit 21, the function of the phase spectrum design unit 22, the function of the intensity spectrum design unit 23, and the function of the modulation pattern generation unit 24. To realize. Each function may be realized by the same processor or may be realized by different processors.

The computer processor can realize each of the above functions by the modulation pattern calculation program. Therefore, the modulation pattern calculation program operates the computer processor as an arbitrary waveform input unit 21, a phase spectrum design unit 22, an intensity spectrum design unit 23, and a modulation pattern generation unit 24 in the data creation unit 17. The modulation pattern calculation program is stored in a storage device (storage medium) inside or outside the computer. The storage device may be a non-temporary recording medium. Examples of the recording medium include flexible disks, recording media such as CDs and DVDs, recording media such as ROMs, semiconductor memories, and cloud servers.

The arbitrary waveform input unit 21 receives an input of a desired time intensity waveform from the operator. The operator inputs information about a desired time intensity waveform (for example, the number of pulses, pulse width, etc.) to the arbitrary waveform input unit 21. Information about the desired time intensity waveform is given to the phase spectrum design unit 22 and the intensity spectrum design unit 23. The phase spectrum design unit 22 calculates the corresponding phase spectrum based on the time intensity waveform. The intensity spectrum design unit 23 calculates the corresponding intensity spectrum based on the time intensity waveform. The modulation pattern generation unit 24 provides a phase modulation pattern (for example, a computer composite hologram) for giving the phase spectrum obtained by the phase spectrum design unit 22 and the intensity spectrum obtained by the intensity spectrum design unit 23 to the output light. calculate. Then, a control signal SC including the calculated phase modulation pattern is provided to the SLM 14, and the SLM 14 is controlled based on the control signal SC.

FIG. 16 is a block diagram showing the internal configurations of the phase spectrum design unit 22 and the intensity spectrum design unit 23. As shown in FIG. 16, the phase spectrum design unit 22 and the intensity spectrum design unit 23 include a Fourier transform unit 25, a function replacement unit 26, a waveform function correction unit 27, an inverse Fourier transform unit 28, and a target generation unit 29. .. The target generation unit 29 includes a Fourier transform unit 29a and a spectrogram correction unit 29b. The function of each of these components will be described in detail later.

Here, the desired time intensity waveform is represented as a function of the time domain and the phase spectrum is represented as a function of the frequency domain. Therefore, the phase spectrum corresponding to the desired time intensity waveform can be obtained, for example, by an iterative Fourier transform based on the desired time intensity waveform. FIG. 17 is a diagram showing a procedure for calculating a phase spectrum by the iterative Fourier transform method. First, the initial intensity spectrum function A ₀ (ω) and the phase spectrum function Ψ ₀ (ω), which are functions of the frequency ω, are prepared (processing number (1) in the figure). In one example, these intensity spectral functions A ₀ (ω) and phase spectral function Ψ ₀ (ω) represent the spectral intensity and spectral phase of the single light pulse P1, respectively. Next, a waveform function (a) in the frequency domain including the intensity spectrum function A ₀ (ω) and the phase spectrum function Ψ _n (ω) is prepared (processing number (2) in the figure).

The subscript n represents after the nth Fourier transform process. Before the first (first) Fourier transform process, the above-mentioned initial phase spectral function Ψ ₀ (ω) is used as the phase spectral function Ψ _n (ω). i is an imaginary number.

Subsequently, the Fourier transform from the frequency domain to the time domain is performed on the function (a) (arrow A1 in the figure). As a result, the waveform function (b) in the frequency domain including the time intensity waveform function b _n (t) and the time phase waveform function Θ _n (t) can be obtained (process number (3) in the figure).

Subsequently, the time intensity waveform function b _n (t) included in the above function (b) is replaced with the time intensity waveform function Target ₀ (t) based on the desired waveform (process numbers (4) and (5) in the figure. )).


Subsequently, the inverse Fourier transform from the time domain to the frequency domain is performed on the function (d) (arrow A2 in the figure). As a result, the waveform function (e) in the frequency domain including the intensity spectrum function B _n (ω) and the phase spectrum function Ψ _n (ω) can be obtained (processing number (6) in the figure).

Subsequently, in order to constrain the intensity spectrum function B _n (ω) included in the above function (e), it is replaced with the initial intensity spectrum function A ₀ (ω) (process number (7) in the figure).

After that, by repeating the above processes (1) to (7) a plurality of times, the phase spectrum shape represented by the phase spectrum function Ψ _n (ω) in the waveform function is changed to the phase spectrum shape corresponding to the desired time intensity waveform. Can be approached to. The finally obtained phase spectral function Ψ _IFTA (ω) is the basis of the modulation pattern for obtaining the desired time intensity waveform.

However, in the iterative Fourier method as described above, although the time intensity waveform can be controlled, there is a problem that the frequency component (band wavelength) constituting the time intensity waveform cannot be controlled. Therefore, the data creation unit 17 of the present embodiment calculates the phase spectrum function and the intensity spectrum function that are the basis of the modulation pattern by using the calculation method described below. FIG. 18 is a diagram showing a calculation procedure of the phase spectrum function in the phase spectrum design unit 22. First, the initial intensity spectrum function A ₀ (ω) and the phase spectrum function Φ ₀ (ω), which are functions of the frequency ω, are prepared (process number (1) in the figure). In one example, these intensity spectral functions A ₀ (ω) and phase spectral function Φ ₀ (ω) represent the spectral intensity and spectral phase of the single light pulse P1, respectively. Next, the first waveform function (g) in the frequency domain including the intensity spectrum function A ₀ (ω) and the phase spectrum function Φ ₀ (ω) is prepared (processing number (2-a)). However, i is an imaginary number.

Subsequently, the Fourier transform unit 25 of the phase spectrum design unit 22 performs a Fourier transform from the frequency domain to the time domain with respect to the function (g) (arrow A3 in the figure). As a result, a second waveform function (h) in the time domain including the time intensity waveform function a ₀ (t) and the time phase waveform function φ ₀ (t) is obtained (Fourier transform step, processing number (3)).

Subsequently, the function replacement unit 26 of the phase spectrum design unit 22 performs the desired waveform input by the arbitrary waveform input unit 21 to the time intensity waveform function b ₀ (t) as shown in the following mathematical formula (i). Substitute the time intensity waveform function Target ₀ (t) based on (processing number (4-a)).

Subsequently, the function replacement unit 26 of the phase spectrum design unit 22 replaces the time intensity waveform function a ₀ (t) with the time intensity waveform function b ₀ (t) as shown in the following mathematical formula (j). That is, the time intensity waveform function a ₀ (t) included in the above function (h) is replaced with the time intensity waveform function Target ₀ (t) based on a desired waveform (function replacement step, processing number (5)).

Subsequently, the waveform function correction unit 27 of the phase spectrum design unit 22 performs the second waveform function so that the spectrogram of the second waveform function (j) after replacement approaches the pre-generated target spectrogram according to the desired wavelength band. Fix it. First, the second waveform function (j) after replacement is subjected to time-frequency conversion to convert the second waveform function (j) into the spectrogram SG _{0, k} (ω, t) (processing in the figure). Number (5-a)). The subscript k represents the kth conversion process.

Here, the time-frequency conversion is a frequency filter process or a numerical calculation process (a process of multiplying a composite signal such as a time waveform while shifting the window function to derive a spectrum for each time). Is applied to convert to three-dimensional information consisting of time, frequency, and signal component strength (spectral strength). Further, in the present embodiment, the conversion result (time, frequency, spectral intensity) is defined as "spectrogram".

Examples of the time-frequency transform include a short-time Fourier transform (STFT) and a wavelet transform (Haar wavelet transform, Gabor wavelet transform, Mexican hat wavelet transform, Morley wavelet transform).

Further, the target spectrogram TargetSG ₀ (ω, t) generated in advance according to a desired wavelength band is read from the target generation unit 29. This target spectrogram TargetSG ₀ (ω, t) has approximately the same value as the target time waveform (time intensity waveform and frequency components constituting it), and is generated by the target spectrogram function of process number (5-b). ..

Next, the waveform function correction unit 27 of the phase spectrum design unit 22 performs pattern matching between the spectrogram SG _{0, k} (ω, t) and the target spectrogram Target SG ₀ (ω, t), and the degree of similarity (how much they match). Check). In the present embodiment, an evaluation value is calculated as an index indicating the degree of similarity. Then, in the subsequent processing number (5-c), it is determined whether or not the obtained evaluation value satisfies a predetermined end condition. If the condition is satisfied, the process proceeds to the process number (6), and if not, the process proceeds to the process number (5-d). In the process number (5-d), the time phase waveform function φ ₀ (t) included in the second waveform function is changed to an arbitrary time phase waveform function φ _{0, k} (t). The second waveform function after changing the time-phase waveform function is converted into a spectrogram again by a time-frequency conversion such as SFTT. After that, the above-mentioned processing numbers (5-a) to (5-c) are repeated. In this way, the second waveform function is modified so that the spectrogram SG _{0, k} (ω, t) gradually approaches the target spectrogram Target SG ₀ (ω, t) (waveform function modification step).

After that, the inverse Fourier transform unit 28 of the phase spectrum design unit 22 performs an inverse Fourier transform on the modified second waveform function (arrow A4 in the figure) to generate a third waveform function (k) in the frequency domain. (Inverse Fourier transform step, process number (6)).

The phase spectrum function Φ _{0, k} (ω) included in the third waveform function (k) becomes the desired phase spectrum function Φ _TWC-TFD (ω) finally obtained. This phase spectral function Φ _TWC-TFD (ω) is provided to the modulation pattern generation unit 24.

FIG. 19 is a diagram showing a procedure for calculating the spectral intensity in the intensity spectrum design unit 23. Since the processing numbers (1) to (5-c) are the same as the spectrum phase calculation procedure in the phase spectrum design unit 22 described above, the description thereof will be omitted. In the waveform function correction unit 27 of the intensity spectrum design unit 23, when the evaluation value indicating the degree of similarity between the spectrogram SG _{0, k} (ω, t) and the target spectrogram Target SG ₀ (ω, t) does not satisfy the predetermined end condition. , The time phase waveform function φ ₀ (t) included in the second waveform function is constrained by the initial value, and the time intensity waveform function b ₀ (t) is changed to an arbitrary time intensity waveform function b _{0, k} (t). (Processing number (5-e)). The second waveform function after changing the time intensity waveform function is converted into a spectrogram again by a time-frequency conversion such as STFT. After that, the processing numbers (5-a) to (5-c) are repeated. In this way, the second waveform function is modified so that the spectrogram SG _{0, k} (ω, t) gradually approaches the target spectrogram Target SG ₀ (ω, t) (waveform function modification step).

After that, the inverse Fourier transform unit 28 of the intensity spectrum design unit 23 performs an inverse Fourier transform on the modified second waveform function (arrow A4 in the figure) to generate a third waveform function (m) in the frequency domain. (Inverse Fourier transform step, process number (6)).

Subsequently, in the processing number (7-b), the filter processing unit of the intensity spectrum design unit 23 receives a single optical pulse P1 with respect to the intensity spectrum function B _{0, k} (ω) included in the third waveform function (m). Filtering is performed based on the intensity spectrum of (filtering step). Specifically, of the intensity spectrum obtained by multiplying the intensity spectrum function B _{0, k} (ω) by the coefficient α, the portion exceeding the cutoff intensity for each wavelength determined based on the intensity spectrum of the single optical pulse P1 is cut. To do. This is to prevent the intensity spectral function αB _{0, k} (ω) from exceeding the spectral intensity of the single light pulse P1 in all wavelength ranges. In one example, the cutoff intensity for each wavelength is set to match the intensity spectrum of the single light pulse P1 (in this embodiment, the initial intensity spectrum function A ₀ (ω)). In that case, as shown in the following equation (n), at frequencies where the intensity spectral function αB _{0, k} (ω) is larger than the intensity spectral function A ₀ (ω), the intensity spectral function A _TWC-TFD (ω) The value of the intensity spectrum function A ₀ (ω) is taken as the value of. Further, the intensity spectrum function αB _{0, k} (ω) is the intensity spectrum function A frequency is ₀ (omega) or less, the intensity spectrum function A _TWC-TFD intensity spectrum as a value of (omega) function αB _{0, k} (ω) The value of is taken in (processing number (7-b) in the figure).

This intensity spectral function A _TWC-TFD (ω) is provided to the modulation pattern generation unit 24 as the desired spectral intensity finally obtained.

The modulation pattern generation unit 24 includes the spectral phase indicated by the phase spectrum function Φ _TWC-TFD (ω) calculated by the phase spectrum design unit 22 and the intensity spectrum function A _TWC-TFD calculated by the intensity spectrum design unit 23. A phase modulation pattern (for example, a computer composite hologram) for giving the spectral intensity indicated by ω) to the optical pulses P21 to P23 is calculated (data generation step).

Here, FIG. 20 is a diagram showing an example of a procedure for generating the target spectrogram TargetSG ₀ (ω, t) in the target generation unit 29. Since the target spectrogram TargetSG ₀ (ω, t) indicates the target time waveform (time intensity waveform and frequency components (wavelength band components) that compose it), the target spectrogram is created by the frequency components (wavelength band components). It is a very important process to control. As shown in FIG. 20, the target generation unit 29 first has a spectral waveform (initial intensity spectral function A ₀ (ω) and initial phase spectral function Φ ₀ (ω)), and a desired time intensity waveform function Target _0. Enter (t). Further, the time function p ₀ (t) including the desired frequency (wavelength) band information is input (processing number (1)).

Next, the target generation unit 29 realizes the time intensity waveform function Target ₀ (t) by using, for example, the iterative Fourier transform method shown in FIG. 17 or the method described in Non-Patent Document 8 or 9. The phase spectrum function Φ _IFTA (ω) of is calculated (processing number (2)).

Subsequently, the target generation unit 29 uses the intensity spectrum function A _IFTA for realizing the time intensity waveform function Target ₀ (t) by the iterative Fourier transform method using the previously obtained phase spectrum function Φ _IFTA (ω). (Ω) is calculated (process number (3)). Here, FIG. 21 is a diagram showing an example of a procedure for calculating the intensity spectrum function A _IFTA (ω).

First, the initial intensity spectral function _{Ak = 0} (ω) and the phase spectral function Ψ ₀ (ω) are prepared (processing number (1) in the figure). Next, a waveform function (o) in the frequency domain including the intensity spectrum function _Ak (ω) and the phase spectrum function Ψ ₀ (ω) is prepared (processing number (2) in the figure).

The subscript k represents after the kth Fourier transform process. First in front of the Fourier transform processing (first time), the above initial strength spectral function A _{k = 0} (omega) is used as the intensity spectrum function A _k (omega). i is an imaginary number.

Subsequently, the Fourier transform from the frequency domain to the time domain is performed on the function (o) (arrow A5 in the figure). As a result, the waveform function (p) in the frequency domain including the time intensity waveform function b _k (t) can be obtained (process number (3) in the figure).

Subsequently, the time intensity waveform function b _k (t) included in the above function (p) is replaced with the time intensity waveform function Target ₀ (t) based on the desired waveform (process numbers (4) and (5) in the figure. )).

Subsequently, the inverse Fourier transform from the time domain to the frequency domain is performed on the function (r) (arrow A6 in the figure). As a result, the waveform function (s) in the frequency domain including the intensity spectrum function C _k (ω) and the phase spectrum function Ψ _k (ω) can be obtained (processing number (6) in the figure).

Subsequently, in order to constrain the phase spectrum function Ψ _k (ω) included in the above function (s), it is replaced with the initial phase spectrum function Ψ ₀ (ω) (processing number (7−a) in the figure).

Further, the intensity spectrum function C _k (ω) in the frequency domain after the inverse Fourier transform is filtered based on the intensity spectrum of the single optical pulse P1. Specifically, in the intensity spectrum represented by the intensity spectrum function C _k (ω), a portion exceeding the cutoff intensity for each wavelength determined based on the intensity spectrum of the single light pulse P1 is cut. In one example, the cutoff intensity for each wavelength is set to match the intensity spectrum of the single light pulse P1 (eg, the initial intensity spectrum function _{Ak = 0} (ω)). In that case, as shown in the following equation (u), at frequencies where the intensity spectrum function C _k (ω) is larger than the intensity spectrum function A _{k = 0} (ω), the value of the intensity spectrum function A _k (ω). The value of the intensity spectrum function A _{k = 0} (ω) is taken in as. Further, at frequencies where the intensity spectrum function C _k (ω) is equal to or less than the intensity spectrum function A _{k = 0} (ω), the value of the intensity spectrum function C _k (ω) is taken in as the value of the intensity spectrum function A _k (ω). (Processing number (7-b) in the figure).

The intensity spectrum function C _k (ω) included in the above function (s) is replaced with the intensity spectrum function A _k (ω) after filtering by the above equation (u).

After that, by repeating the above processes (1) to (7-b), the intensity spectrum shape represented by the intensity spectrum function _Ak (ω) in the waveform function is changed to the intensity spectrum shape corresponding to the desired time intensity waveform. Can be approached to. Finally, the intensity spectral function A _IFTA (ω) is obtained.

See FIG. 20 again. By calculating the phase spectrum function Φ _IFTA (ω) and the intensity spectrum function A _IFTA (ω) in the processing numbers (2) and (3) described above, the third waveform function (v) in the frequency domain including these functions Is obtained (processing number (4)).

The Fourier transform unit 29a of the target generation unit 29 Fourier transforms the above waveform function (v). As a result, the fourth waveform function (w) in the time domain is obtained (processing number (5)).

The spectrogram correction unit 29b of the target generation unit 29 converts the fourth waveform function (w) into the spectrogram SG _IFTA (ω, t) by time-frequency conversion (processing number (6)). Then, in the processing number (7), the target spectrogram TargetSG ₀ (ω, t) is modified by modifying the spectrogram SG _IFTA (ω, t) based on the time function p ₀ (t) including the desired frequency (wavelength) band information. t) is generated. For example, a characteristic pattern appearing in the spectrogram SG _IFTA (ω, t) composed of two-dimensional data is partially cut out, and the frequency component of the part is manipulated based on the time function p ₀ (t). Hereinafter, a specific example thereof will be described in detail.

For example, consider the case where a triple pulse having a time interval of 2 picoseconds is set as the desired time intensity waveform function Target ₀ (t). At this time, the spectrogram SG _IFTA (ω, t) gives the result as shown in FIG. 22 (a). In FIG. 22A, the horizontal axis represents time (unit: femtosecond) and the vertical axis represents wavelength (unit: nm). The spectrogram value is indicated by the lightness and darkness of the figure, and the brighter the spectrogram value, the larger the spectrogram value. In this spectrogram SG _IFTA (ω, t), triple pulses appear as domains D ₁ , D ₂ , and D ₃ divided on the time axis at 2-picosecond intervals. The center (peak) wavelengths of domains D ₁ , D ₂ , and D ₃ are 800 nm.

If it is desired to control only the time intensity waveforms of the optical pulses P21 to P23 (simply obtain a triple pulse), it is not necessary to operate these domains D ₁ , D ₂ , and D ₃ . However, if it is desired to control the frequency (wavelength) band of each pulse, it is necessary to operate these domains D ₁ , D ₂ , and D ₃ . That is, as shown in FIG. 22 (b), moving the domains D ₁ , D ₂ , and D ₃ independently of each other in the direction along the wavelength axis (vertical axis) is the configuration of each pulse. It means changing the frequency (wavelength band). Such a change in the constituent frequency (wavelength band) of each pulse is performed based on the time function p ₀ (t).

For example, when the time function p ₀ (t) is described so that the peak wavelength of domain D ₂ is translated at 800 nm and the peak wavelengths of domains D ₁ and D ₃ are translated by -2 nm and + 2 nm, respectively, the spectrogram SG _IFTA ( ω, t) changes to the target spectrogram TargetSG ₀ (ω, t) shown in FIG. 22 (b). For example, by applying such processing to the spectrogram, it is possible to create a target spectrogram in which the constituent frequencies (wavelength bands) of each pulse are arbitrarily controlled without changing the shape of the time intensity waveform.

As described above, the data creation unit 17 of the present embodiment generates the second waveform function (h) in the time domain by performing a Fourier transform on the first waveform function (g) in the frequency domain. , The second waveform function (h) is replaced with the time intensity waveform function Target ₀ (t) based on a desired waveform. After that, the inverse Fourier transform is performed on the second waveform function to generate the third waveform functions (k) and (m) in the frequency domain. Then, a modulation pattern is generated based on the phase spectrum function Φ _{0, k} (ω) of the third waveform function (k) and the intensity spectrum function B _{0, k} (ω) of the third waveform function (m). As a result, it is possible to preferably generate a modulation pattern for realizing a desired waveform including a plurality of optical pulses.

In addition, in the data creation unit 17 of the present embodiment, after replacing the time intensity waveform function Target ₀ (t) and before the inverse Fourier transform, the target spectrogram TargetSG ₀ (ω, t) is converted to the spectrogram of the second waveform function. Modify the second waveform function so that SG _{0, k} (ω, t) approaches. This target spectrogram TargetSG ₀ (ω, t) is pre-generated according to a desired wavelength band, and this processing corrects the wavelength band of the second waveform function to the desired wavelength band. Therefore, the third waveform functions (k) and (m) obtained by inverse Fourier transforming the second waveform function are also functions within the desired wavelength band. Then, as described above, the modulation pattern is based on the phase spectrum function Φ _{0, k} (ω) of the third waveform function ( _k ) and the intensity spectrum function B _{0, k} (ω) of the third waveform function (m). Is generated. From the above, according to the data creation unit 17 of the present embodiment, it is possible to individually control the wavelength components (frequency components) of a plurality of optical pulses.

Further, as in the present embodiment, the waveform function correction unit 27 calculates an evaluation value indicating the degree of similarity between the spectrogram SG _IFTA (ω, t) of the second waveform function and the target spectrogram TargetSG ₀ (ω, t). , The second waveform function may be modified so that the evaluation value satisfies a predetermined condition. For example, by such a method, the second waveform function can be accurately modified so that the spectrogram SG _IFTA (ω, t) of the second waveform function approaches the target spectrogram TargetSG ₀ (ω, t).

Further, as in the present embodiment, the waveform function correction unit 27 performs the time intensity waveform function b _{0, k} (t) or the time phase waveform function φ _{0, k} (t) in order to correct the second waveform function. You may change it. For example, by such a method, the second waveform function can be suitably modified so that the spectrogram SG _IFTA (ω, t) of the second waveform function approaches the target spectrogram TargetSG ₀ (ω, t).

The data creation unit 17 is not limited to this embodiment, and various changes can be made. For example, in the present embodiment, the phase spectrum design unit 22 calculates the phase spectrum function Φ _TWC-TFD (ω), the intensity spectrum design unit 23 calculates the intensity spectrum function A _TWC-TFD (ω), and generates a modulation pattern. The part 24 generates a modulation pattern based on both functions, and the modulation pattern generation part generates a modulation pattern in one of the phase spectrum function Φ _TWC-TFD (ω) and the intensity spectrum function A _TWC-TFD (ω). A modulation pattern may be generated based on this.

Further, as a time-frequency conversion when the waveform function correction unit 27 converts the second waveform function into the spectrogram SG _{0, k} (ω, t) (processing numbers (5-a) in FIGS. 18 and 19), this method is used. In the embodiment, the short-time Fourier transform (STFT) and the wavelet transform are illustrated. What is important in the time-frequency conversion process is to convert the time waveform into the spectrogram SG _{0, k} (ω, t) which is "time-frequency information". In the present embodiment, unlike the method of controlling only the time intensity waveform among the time waveforms (for example, Non-Patent Documents 8 and 9), the main purpose is to control the frequency component (band component) constituting the time waveform. This is because it is meaningful to extract time intensity information and frequency (band) information from the time waveform. That is, the time-frequency conversion is not limited to the RTM and wavelet transforms, and various conversion processes capable of extracting frequency information from the time waveform can be applied.

Further, in the present embodiment, the waveform function correction unit 27 uses evaluation values indicating the degree of similarity between the spectrogram SG _IFTA (ω, t) of the second waveform function and the target spectrogram Target SG ₀ (ω, t). Determine how close they are to each other (process numbers (5-c) in FIGS. 18 and 19). The target spectrogram TargetSG ₀ (ω, t) represents what kind of time intensity shape and frequency (band) information the desired time waveform contains, and serves as a so-called target value (design drawing). Therefore, the evaluation value of this embodiment can be one of the indexes indicating the waveform control accuracy.

On the other hand, since the spectrogram SG _IFTA (ω, t) contains two variables such as frequency ω and time t, it can be treated as an image. Therefore, to find out how much the spectrogram SG _IFTA (ω, t) and the target spectrogram Target SG ₀ (ω, t) match is a difference extraction work using various pattern matching methods in image analysis. I can think. Therefore, in addition to the method of using the evaluation value indicating the degree of similarity, for example, a method of extracting the feature amount of the image (contour / shape limited to the frequency or the time axis direction) and evaluating the degree of pattern matching is used. , A method of dividing an image into a plurality of parts and evaluating each part can also be applied.

Further, in the present embodiment, when the waveform function correction unit 27 does not satisfy the predetermined condition (the spectrogram SG _IFTA (ω, t) and the target spectrogram Target SG ₀ (ω, t) deviate from each other). In addition, the time phase waveform function φ _{0, k} (t) or the time intensity waveform function b _{0, k} (t) is changed to any other one (process number (5-d) in FIG. 18, process in FIG. 19). Number (5-e)). There are various methods for changing these functions φ _{0, k} (t) and b _{0, k} (t). The simplest method is to randomly change the functions φ _{0, k} (t) and b _{0, k} (t). Further, for example, a method of searching for a solution of a function φ _{0, k} (t) and b _{0, k} (t) according to a certain rule (according to a stochastic process) by a simulated annealing method or the like is also applicable. Alternatively, when an index of what kind of functions φ _{0, k} (t) and b _{0, k} (t) are used to improve the evaluation value can be obtained, the index may be used. For example, the magnitude of the evaluation value calculated by the processing number (5-a) and the judgment result in the processing number (5-c) can be changed by changing the functions φ _{0, k} (t) and b _{0, k} (t). You may give feedback at the time. Specifically, the spectrogram SG _IFTA (ω, t) and the target spectrogram TargetSG _{0 (ω,} t) on the basis of the difference value between, to create a new spectrogram NewSG ₀ (ω, t), the NewSG ₀ (ω , T) is transformed into a time waveform by inverse spectrogram conversion. For example, what kind of function φ _0, can be used for the time phase and time intensity functions of the time waveform obtained by inverse spectrogram conversion of NewSG ₀ (ω, t) obtained through such operations on the spectrogram _. Includes indicators (based on spectrogram calculations) as to whether the use of _k (t), b _{0, k} (t) improves the evaluation value. Therefore, a method of using this index as feedback for modifying the time phase waveform function φ _{0, k} (t) or the time intensity waveform function b _{0, k} (t) can be considered as appropriate.

Further, when the target generation unit 29 generates the target spectrogram TargetSG ₀ (ω, t), the phase spectrum function for realizing the time intensity waveform function Target ₀ (t) at the processing number (2) shown in FIG. 20. Calculate Φ _IFTA (ω). At this time, it has been described above that the iterative Fourier transform method shown in FIG. 17 or the method described in Non-Patent Document 8 or 9 can be used, but these are the methods for calculating the phase spectrum function Φ _IFTA (ω). The phase spectrum function Φ _IFTA (ω) obtained analytically or approximately may be used so as to realize the generation of the time intensity waveform function Target ₀ (t).

Further, when the target generation unit 29 generates the target spectrogram TargetSG ₀ (ω, t), the intensity spectrum function for realizing the time intensity waveform function Target ₀ (t) at the processing number (3) shown in FIG. 20. Calculate A _IFTA (ω). At this time, an example of calculating the intensity spectrum function A _IFTA (ω) using the improved iterative Fourier transform method shown in FIG. 19 has been described above, but the method of calculating the intensity spectrum function A _IFTA (ω) has been described above. Is not limited to this, and an intensity spectrum function A _IFTA (ω) obtained analytically or approximately may be used so as to realize the generation of the time intensity waveform function Target ₀ (t).

A modulation pattern (for example, a computer composite hologram) is created based on the data creation method of the present embodiment using a computer existing in a remote location such as a cloud server, and data related to the created modulation pattern is transmitted to the user. You may. In that case, the data creation unit 17 of this embodiment becomes unnecessary.

(First Example)
Calculations based on the data creation method of the above embodiment were performed, and it was confirmed that the time waveform could be controlled including the control of the frequency (wavelength) band. In this calculation, a single pulse having a wavelength band of 5 nm in full width at half maximum was set as a single optical pulse P1, and a double pulse with an interval of 2 picoseconds was set as a plurality of optical pulses. In this case, the target spectrogram TargetSG (ω, t) contains two domains. In this embodiment, five target spectrograms TargetSG in which the two domains are translated in the wavelength axis direction (that is, the frequency (wavelength) bands constituting each pulse are changed) by the method shown in FIG. 22 (b). (Ω, t) was prepared. Specifically, the combinations of the center wavelengths (peak wavelengths) of the two domains are (800 nm, 800 nm), (801 nm, 799 nm), (802 nm, 798 nm), (803 nm, 797 nm), and (804 nm, 796 nm), respectively. Five types of target spectrograms TargetSG ₀ (ω, t) were prepared. Then, the phase spectral function Φ _TWC-TFD (ω) was calculated using the method shown in FIG. 18, and the intensity spectral function A _TWC-TFD (ω) was calculated using the method shown in FIG.

FIG. 23 shows the center wavelengths of two domains included in each spectrogram SG _{0, k} (ω, t) whose evaluation value satisfies a predetermined condition for each corresponding target spectrogram TargetSG ₀ (ω, t). It is a graph which shows the center wavelength interval between domains. The vertical axis shows the center wavelength of each domain, and the horizontal axis shows the center wavelength interval between domains. Further, the graphs G31 and G32 are straight lines connecting the center wavelengths of one and the other domains included in the target spectrometer TargetSG ₀ (ω, t), respectively, and the graphs G33 and G34 are the spectrogram SG _{0, k} (ω, t). ) Is an approximate curve connecting the center wavelengths of one and the other domains. From this result, it was shown that the time waveform including the frequency (wavelength) band can be controlled when the wavelength band difference of adjacent pulses is, for example, within 4 nm. That is, it was shown that it is possible to change to an arbitrary wavelength band within the range of the full width at half maximum (5 nm) of the wavelength band of the single optical pulse P1. In other words, when moving the domain of the target spectrogram TargetSG (ω, t) in the wavelength axis direction, it is desirable to move within the wavelength band of the single optical pulse P1.

FIG. 24 (a) was obtained using the method shown in FIGS. 18 and 19 with the combination of the center wavelengths of the two domains of the target spectrum TargetSG (ω, t) set to (800 nm, 800 nm). It is a graph which shows the spectral waveform (spectral phase G41 and spectral intensity G42). FIG. 24 (b) is a graph showing the time intensity waveforms of a plurality of optical pulses obtained by Fourier transforming the spectral waveform of FIG. 24 (a). Further, FIG. 25 (a) is obtained by setting the combination of the center wavelengths of the two domains of the target spectrum SG (ω, t) as (802 nm, 798 nm) and using the method shown in FIGS. 18 and 19. It is a graph which shows the spectrum waveform (spectral phase G51 and spectrum intensity G52). FIG. 25 (b) is a graph showing the time intensity waveforms of a plurality of optical pulses obtained by Fourier transforming the spectral waveform of FIG. 25 (a).

Comparing FIG. 24 (a) and FIG. 25 (a), when the central wavelengths of each domain are 800 nm and 800 nm (FIG. 24 (a)), the phase spectrum (G41) is stepped and has an intensity. The phase spectrum (G41) is folded back at a wavelength near the tail of the spectrum (G42). On the other hand, when the central wavelengths of each domain are 802 nm and 798 nm (FIG. 25 (a)), the step of the phase spectrum (G51) becomes smoother as the wavelength near the tail of the spectral intensity (G52) approaches, and , The phase spectrum (G51) is not folded back. From this, it can be seen that there is a clear difference in the phase spectrum in order to control the frequency (wavelength) bands of a plurality of optical pulses.

Further, when FIG. 24 (b) and FIG. 25 (b) are compared, it can be seen that the same time intensity waveform can be obtained regardless of the difference in the control of the frequency (wavelength) bands of the plurality of optical pulses.

(Second Example)
Subsequently, an embodiment in which seven optical pulses are generated and the wavelength bands of the pulses are different from each other will be described. 26 (a) and 27 (a) show the target spectrogram TargetSG (ω, t) used in this example. FIG. 26 (a) shows a case where the wavelength bands of each pulse are not controlled (equalized), and FIG. 27 (a) shows a case where the wavelength bands of each pulse are different from each other. Further, FIGS. 26 (b) and 27 (b) show the spectrogram SG _{0, k} (ω, t) calculated based on the target spectrogram Target SG (ω, t) of FIGS. 26 (a) and 27 (a), respectively. t). In these figures, the horizontal axis indicates time (unit: femtosecond), and the vertical axis indicates wavelength (unit: nm). The spectrogram value is indicated by the lightness and darkness of the figure, and the brighter the spectrogram value, the larger the spectrogram value. In this embodiment, the target spectrogram Target SG (ω, t) and the spectrogram SG _{0, k} (ω, t) include the same number of domains D _{1 to} D ₇ as the number of pulses.

28 (a) shows the spectral waveforms (spectral phase G61 and spectral intensity G62) calculated from the time waveform (second waveform function) corresponding to the spectrogram SG _{0, k} (ω, t) of FIG. 26 (b). It is a graph which shows. FIG. 28 (b) is a graph showing the time intensity waveforms of a plurality of optical pulses obtained by Fourier transforming the spectral waveform of FIG. 28 (a). Further, FIG. 29 (a) shows a spectral waveform (spectral phase G71 and spectral intensity G72) calculated from a time waveform (second waveform function) corresponding to the spectrogram SG _{0, k} (ω, t) of FIG. 27 (b). ) Is a graph. FIG. 29 (b) is a graph showing the time intensity waveforms of a plurality of optical pulses obtained by Fourier transforming the spectral waveform of FIG. 29 (a).

Comparing FIGS. 28 (a) and 29 (a), when the center wavelengths of the domains are equal to each other (FIG. 28 (a)) and when the center wavelengths of the domains are different from each other (FIG. 29 (a)). It can be seen that there is a clear difference in the phase spectrum. On the other hand, when FIG. 28 (b) and FIG. 29 (b) are compared, it can be seen that the same time intensity waveform can be obtained regardless of the difference in the control of the frequency (wavelength) bands of the plurality of optical pulses.

(First modification)
In the first embodiment, the multi-pulse generator 10A has a lens 19 in order to align the optical paths L1 to L3 of the optical pulses P21 to P23, but as shown in FIG. 30, the diffraction grating 18 and the lens 19 A cylindrical mirror 41 and a periscope array 42 may be further provided between them. In that case, the plurality of optical pulses P21 to P23 divided for each wavelength by the diffraction grating 18 are incident on the periscope array 42 via the cylindrical mirror 41. Then, after the interval between the optical paths of the optical pulses P21 to P23 is adjusted by the periscope array 42, the optical pulses P21 to P23 are focused by the lens 19. In such a configuration, by controlling the reflection angle of the periscope array 42, the condensing position of each of the light pulses P21 to P23 can be suitably controlled as in FIG. 7. In this modification as well, the diffraction grating 18 may be either a reflective type or a transmissive type, and the cylindrical mirror 41 may be a lens. Further, the lens 19 may be a cylindrical mirror.

(Second modification)
In the first embodiment, the diffraction grating 18 is used as an optical dispersion element for separating the optical pulses P21 to P23 output from the time waveform shaper 11 into different optical paths L1 to L3. FIG. 31 shows. As shown, the prism 43 may be used instead of the diffraction grating 18. Even in such a form, the optical pulses P21 to P23 having different center wavelengths can be suitably separated according to the wavelength. However, since the dispersion angle of the prism 43 for each wavelength is smaller than that of the diffraction grating 18, the diffraction grating 18 is selected when the terahertz wave generator needs to be miniaturized.

(Third modification example)
In the first embodiment shown in FIG. 7, the light pulse P21 having the longest center wavelength is output earliest, and the light pulse P23 having the shortest center wavelength is output the latest. Therefore, as shown in FIGS. 7 and 9, the focused spots Q1 to Q3 corresponding to the optical pulses P21 to P23 are formed in this order in time. That is, the light pulses that are spatially adjacent to each other at the condensing position are also temporally adjacent to each other.

However, from the viewpoint of avoiding carrier saturation, it is more desirable that temporally adjacent optical pulses are focused as spatially separated as possible. FIG. 32 is a diagram showing the focusing positions of the four optical pulses P21 to P24 in one modification. In this modification, the output timings of the optical pulses P21 to P24 are devised so that the focusing positions of the optical pulses that are adjacent in time are not adjacent to each other. That is, when the central wavelength of the optical pulse P21 is λ1, the central wavelength of the optical pulse P22 is λ2, the central wavelength of the optical pulse P23 is λ3, and the central wavelength of the optical pulse P24 is λ4, the relationship of λ1> λ2> λ3> λ4. Satisfy. Then, the optical pulse P23 is output first, then the optical pulse P21 is output, then the optical pulse P24 is output, and finally the optical pulse P22 is output. In this case, among the condensing spots Q1 to Q4 arranged in a predetermined direction, the condensing spot Q3 is formed first, the condensing spot Q1 is formed next, the condensing spot Q4 is formed next, and finally the condensing spot Q4 is formed. Light spot Q2 is formed. As described above, since the focusing positions of the light pulses that are adjacent in time are not adjacent to each other, the saturation of the carriers in the terahertz wave generating element 30 can be avoided more effectively, and the terahertz wave can be generated more efficiently.

(Fourth modification)
FIG. 33 is a diagram schematically showing a part of the configuration of the terahertz wave generator 1B according to the fourth modification of the above embodiment. As shown in FIG. 33, the terahertz wave generator 1B according to the present modification further includes a pulse stretcher (pulse expander) 44 arranged on the optical path between the light source 3 and the time waveform shaper 11. The pulse stretcher 44 is optically coupled to the light source 3 to extend the pulse time width by giving different delays for each wavelength component of the single light pulse P1. Specifically, the pulse stretcher 44 generates a chirp light pulse P11 having wavelength components different in time by giving a linear dispersion to the single light pulse P1.

After that, the chirp optical pulse P11 is input to the time waveform shaper 11. The time waveform shaper 11 divides each wavelength component of the chirp optical pulse P11 to generate a plurality of optical pulses having different center wavelengths from each other. At the same time, the time waveform shaper 11 gives a different delay time for each of the plurality of optical pulses and compresses the time width of each optical pulse. As a result, a plurality of optical pulses P21 to P23 having a time difference from each other and having different center wavelengths are generated.

FIG. 34 is a diagram showing a grating pair type configuration as a specific configuration example of the pulse stretcher 44. The pulse stretcher 44 includes a pair of diffraction gratings 44a and 44b and a mirror 44c. A single optical pulse P1 is input to the diffraction grating 44a. The single light pulse P1 is separated into each wavelength component on the diffraction grating 44a. Each of the dispersed wavelength components is input to the diffraction grating 44b, and the optical paths of each are aligned in parallel. After that, these wavelength components are reflected by the mirror 44c and are again coupled to each other via the diffraction gratings 44b and 44a. In this process, the optical path lengths of each wavelength component are different (the long wavelength side is short and the short wavelength side is long). As a result, the optical pulse after the combined wave is given a different delay for each wavelength component, and the pulse time width. Is extended. The pulse stretcher 44 is not limited to such a configuration, and for example, a prism pair may be used, or a glass rod having an appropriate length may be used.

(Fifth modification)
35 (a) to 35 (c) are diagrams for explaining a fifth modification of the above embodiment. In this modification, the polarization direction of the multi-terahertz wave pulse is controlled by devising the antenna structure of the terahertz wave generating element 30. In one of the three sets of antenna portions of the dipole antenna shown in FIG. 35 (a), the opposite direction of the pair of antenna portions 34b1 is different from that of the other set of antenna portions 34b2. That is, in the other set, the positions of the pair of antenna portions 34b2 in the predetermined direction B3 coincide with each other, so that the pair of antenna portions 34b2 face each other in the direction B4 orthogonal to the predetermined direction B3. On the other hand, in one set, the positions of the pair of antenna portions 34b1 in the predetermined direction B3 are deviated from each other, and the protruding length is longer than that of the other set. As a result, the pair of antenna portions 34b1 face each other in the predetermined direction B3. Then, a focusing spot Q3 of the optical pulse P23 is formed in the antenna gap of the pair of antenna portions 34b1 facing each other in the predetermined direction B3, and the focusing spot of the optical pulses P21 and P22 is formed in the antenna gap of the other set of antenna portions 34b2. It is assumed that Q1 and Q2 are formed.

In this case, as shown in FIG. 35 (c), the polarization directions TP1 and TP2 of the terahertz waves T1 and T2 generated by the optical pulses P21 and P22 are along the direction B4, and the polarization direction TP3 of the terahertz wave T3 generated by the optical pulse P23. Is along the predetermined direction B3. Then, when the time waveform shaper 11 first outputs the optical pulse P21, then outputs the optical pulse P22, and finally outputs the optical pulse P23, the first output terahertz wave T1 and then the next output. The polarization direction of the terahertz wave T2 is along the direction B4, and the polarization direction of the finally output terahertz wave T3 is along the predetermined direction B3. Further, when the time waveform shaper 11 first outputs the optical pulse P23, then outputs the optical pulse P22, and finally outputs the optical pulse P21, the polarization direction of the first output terahertz wave T3 is a predetermined direction. The polarization directions of the next output terahertz wave T2 and the last output terahertz wave T1 along B3 are along the direction B4. As described above, according to this modification, by controlling the delay time of each optical pulse P21 to P23, the polarization directions of a plurality of terahertz waves output in order from the terahertz wave generating element 30 are independently controlled. can do.

(6th modification)
36 and 37 are diagrams for explaining a sixth modification of the above embodiment. In this modification, the antenna structure of the terahertz wave generating element 30 is further devised, and a moving stage that enables the relative position between the optical path of the optical pulses P21 to P23 and the terahertz wave generating element 30 is further provided to provide terahertz. Increase the degree of freedom in controlling the polarization direction of the wave.

The dipole antenna shown in FIGS. 36 (a) and 37 (a) has six sets of antenna portions. In the three sets, the positions of the pair of antenna portions 34b1 in the predetermined direction B3 are deviated from each other, and the protruding length is longer than that of the other sets. As a result, the pair of antenna portions 34b1 face each other in the predetermined direction B3. Further, in the remaining three sets, the positions of the pair of antenna portions 34b2 in the predetermined direction B3 coincide with each other, and the pair of antenna portions 34b2 face each other in the direction B4 orthogonal to the predetermined direction B3.

As shown in FIG. 36A, it is assumed that the focused spots Q1 to Q3 of the three optical pulses P21 to P23 are formed in the antenna gap of the first three sets of antenna portions 34b1. In this case, as shown in FIG. 36B, the polarization directions TP1 to TP3 of the terahertz waves T1 to T3 generated by the optical pulses P21 to P23 all follow the predetermined direction B3. On the other hand, by moving the relative positions of the optical paths of the optical pulses P21 to P23 and the terahertz wave generating element 30, a focused spot Q1 is formed in the antenna gap of the antenna portion 34b1 as shown in FIG. 37 (a). It is assumed that the focusing spots Q2 and Q3 are formed in the antenna gap of the antenna portion 34b2. In this case, as shown in FIG. 37 (b), the polarization direction TP1 of the terahertz wave T1 generated by the optical pulse P21 is along the predetermined direction B3, and the polarization direction TP2 of the terahertz waves T2 and T3 generated by the optical pulses P22 and P23. TP3 will follow direction B4. As described above, in this modification, by changing the relative position between the optical path of the optical pulses P21 to P23 and the terahertz wave generating element 30, the polarization directions of the plurality of terahertz waves output in order from the terahertz wave generating element 30 are changed. Can be controlled.

In this modification, the order of polarization of the multi-terahertz wave pulse can also be controlled by controlling the time delay of each optical pulse P21 to P23. Further, by changing the positions of the focused spots Q1 to Q3 of the optical pulses P21 to P23, the combination of polarized light of the multiterahertz wave pulse can be controlled. For example, Non-Patent Document 4 shows that in coherent control of PbTiO ₃ , crystal vibration can be efficiently excited by rotating the polarization of the last terahertz wave pulse among a plurality of terahertz wave pulses by 90 °. .. Therefore, this modification is particularly effective for coherent control of crystals and the like.

(7th modification)
FIG. 38 is a diagram for explaining a seventh modification of the above embodiment. In this modification, the sign of the electric field of the multi-terahertz wave pulse is controlled by devising the antenna structure of the terahertz wave generating element 30. In the dipole antenna shown in FIG. 38 (a), the two antenna portions 34b3 of one electrode and the two antenna portions 34b3 of the other electrode are alternately arranged in a predetermined direction B3. Then, a first antenna gap is formed by the first antenna portion 34b3 of one electrode and the first antenna portion 34b3 of the other electrode, and the second antenna portion 34b3 of one electrode and the second antenna portion of the other electrode are formed. A second antenna gap is formed by the antenna portion 34b3 of 1, and a third antenna gap is formed by the second antenna portion 34b3 of one electrode and the second antenna portion 34b3 of the other electrode. A focusing spot Q1 is formed in the first antenna gap, a focusing spot Q2 is formed in the second antenna gap, and a focusing spot Q3 is formed in the third antenna gap.

When an optical pulse is applied to a dipole antenna, the sign of the electric field of the generated terahertz wave changes depending on the direction of current flow by the carrier. Here, in FIG. 38A, it is assumed that the electric field of the terahertz wave oscillates in the positive direction when the current flows in one direction of the predetermined direction B3 (arrow B31 in the drawing). At this time, the electric fields of the terahertz wave pulses T1 and T3 oscillate mainly in the positive direction. On the other hand, the terahertz wave pulse T2 oscillates mainly in the negative direction. As described above, according to this modification, the vibration direction of the multi-terahertz wave pulse in each electric field can be freely controlled. Further, as in the fifth modification and the sixth modification described above, the polarization direction of the multi-terahertz wave pulse can be controlled at the same time by devising the facing directions of the dipole antennas.

Non-Patent Document 5 reports that the spin of ErFeO ₃ , which is an antiferromagnet, can be controlled by using a terahertz wave pulse. In Non-Patent Document 5, a half-cycle (an electric field that oscillates in either plus or minus, that is, a single peak) terahertz wave is used to give an impulse-like impact force to the spin. It has been shown that the excitation F-mode (0.377 THz) and AF-mode (0.673 THz) resonances can be distinguished. It is also known that the spin direction is also important in magnetization control of hard disk drives and the like. Therefore, in spin control, it is desired to control the vibration direction of the electric field of the terahertz wave. This variant is useful for such spin coherent control.

(Second Embodiment)
39 and 40 are diagrams showing a part of the configuration of the terahertz wave generator 1C according to the second embodiment of the present invention. 39 (a) is a side view, FIG. 39 (b) is a top view, and FIG. 40 is a perspective view. Unlike the first embodiment, the terahertz wave generator 1C of the present embodiment does not include the diffraction grating 18. That is, the time waveform shaper 11 and the lens 19 are photocoupled without passing through the diffraction grating 18.

Specifically, the SLM14 of the present embodiment has a time difference from each other and has a central wavelength by performing phase modulation of the single light pulse P1 after spectroscopy by the diffraction grid 12 for each wavelength, as in the first embodiment. Generates a plurality of light pulses P21 to P23 different from each other. In addition, the SLM14 of the present embodiment separates the optical paths of the plurality of optical pulses P21 to P23 in a direction (arrow B3 in the figure) that intersects the spectral direction (arrow B4 in the figure) according to the wavelength. Such a function of the SLM 14 is preferably realized when the modulation pattern displayed on the SLM 14 includes a diffraction grating pattern in which the gratings are arranged along the direction B3. That is, the optical pulses P21 to P23 are output at different angles according to their center wavelengths due to the action of the diffraction grating pattern. The lens 15 collects a plurality of light pulses P21 to P23 in the spectral direction B4. In the diffraction grating 16, the optical paths of the light pulses P21 to P23 coincide with each other by the diffraction action in the direction B4, but remain different from each other because the diffraction action does not occur in the direction B3. After that, the light pulses P21 to P23 are focused by the lens 19 to form three focused spots Q1 to Q3 arranged along the direction B3 in the terahertz wave generating element 30. In one example, the distance from the SLM 14 to the lens 19 and the distance from the lens 19 to the terahertz wave generating element 30 are both equal to the focal length f of the lens 19.

FIG. 41 is a diagram showing an example of the modulation pattern of the SLM 14 in the present embodiment. In this embodiment, the modulation pattern of the SLM 14 is divided into three regions 14c to 14e in the direction B4. The region 14c is a region corresponding to a long wavelength region having a central wavelength of 840 nm (that is, the band of the optical pulse P21), and the region 14d is a region corresponding to a medium wavelength region having a central wavelength of 800 nm (that is, the band of the optical pulse P22). The region 14e is a region corresponding to a short wavelength region having a central wavelength of 760 nm (that is, a band of the optical pulse P23). Regions 14c and 14e include a phase pattern of a blazed diffraction grating whose phase changes smoothly from 0 (rad) to π (rad).

Then, the blazed diffraction grating outputs the light pulse P21 to the optical path inclined in one direction in the direction B4 in the region 14c, and outputs the light pulse P23 to the optical path inclined in the other direction in the direction B4 in the region 14e. To do. The region 14d does not include a blazed diffraction grating, and the optical pulse P22 is output in a direction intermediate between these optical paths.

In one embodiment, the number of lines of the diffraction gratings 12 and 16 is 300 lines / mm, and the focal lengths of the lenses 13 and 15 are 181.5 mm. The number of lines of the phase pattern of the blazed diffraction grating of the SLM 14 is, for example, 3.33 lines / mm, the light diameter of the single light pulse P1 is 16 mm, and the focal length of the lens 19 is f = 500 mm. In this case, the diameters of the focused spots Q1 to Q3 are 30 μm, and the distance between the focused spots Q1 to Q3 is 66 μm. Therefore, also in the present embodiment, as in the first embodiment, the distance between the focused spots Q1 to Q3 is wider than the diameter of the focused spots Q1 to Q3, and the focused spots Q1 to Q3 are surely separated from each other. Can be done. As a result, carrier saturation can be suppressed and the efficiency of terahertz wave generation can be improved. Also in this embodiment, it is desirable that the light pulses that are adjacent in time are focused as spatially separated as possible, as in the case of the third modification described above.

Here, a method of generating a terahertz wave according to the present embodiment will be described. This terahertz wave generation method can be realized by using the terahertz wave generator 1C of the present embodiment. FIG. 42 is a flowchart showing the terahertz wave generation method of the present embodiment.

First, in the first step (pulse light generation step) S11, the single light pulse P1 is output from the light source 3. Next, in the second step (multi-pulse light generation step) S12, a plurality of light pulses P21 to P23 having a time difference from each other and having different center wavelengths are generated from the single light pulse P1, and each of the light pulses P21 to P23 is generated. , Outputs on different optical paths depending on the wavelength. The second step S12 includes a spectroscopic step S13, a light modulation step S14, and a focusing step S15. In the spectroscopy step S13, the single light pulse P1 is separated. In the optical modulation step S14, by performing phase modulation of the single optical pulse P1 after spectroscopy for each wavelength, a plurality of optical pulses P21 to P23 having a time difference from each other and having different center wavelengths are generated, and the optical pulses P21 to P21 The optical path of P23 is separated into a direction B3 that intersects the spectral direction B4 according to the wavelength. In the focusing step S15, the optical pulses P21 to P23 are focused in the spectral direction B4. Finally, in the third step (terahertz wave generation step) S16, the optical pulses P21 to P23 are input to the terahertz wave generating element 30 to generate the terahertz wave pulse.

According to the terahertz wave generator 1C and the terahertz wave generation method of the present embodiment, similarly to the terahertz wave generator 1A described above, carrier saturation in the terahertz wave generator 30 is more effectively avoided, and terahertz waves are generated. Efficiency can be improved.

The terahertz wave generator and the terahertz wave generation method according to the present invention are not limited to the above-described embodiments, and various other modifications are possible. For example, the above-described embodiments and modifications may be combined with each other according to the required purpose and effect. Further, in each of the above embodiments, a terahertz wave generating element using a semiconductor as a terahertz wave generating medium has been exemplified, but the present invention is also effective for a terahertz wave generating element using a nonlinear crystal. That is, in semiconductors, carrier saturation due to an increase in laser power becomes a problem, but in nonlinear crystals as well, when laser power increases, saturation of terahertz wave output due to two-photon absorption of the crystal itself becomes a problem (for example, non-patented). Document 6). Therefore, the present invention can be applied even when the terahertz wave generating element has a non-linear crystal. Non-linear crystals include, for example, ZnTe, Gase, GaP, DAST, DSTMS and the like.

1A to 1C ... Terahertz wave generator, 3 ... Light source, 10A ... Multi-pulse generator, 11 ... Time waveform shaper, 12, 16 ... Diffraction grating, 13, 15 ... Lens, 14a ... Modulation surface, 14b ... Modulation region, 17 ... Data creation unit, 18 ... Diffraction grating, 19 ... Lens, 21 ... Arbitrary waveform input unit, 22 ... Phase spectrum design unit, 23 ... Intensity spectrum design unit, 24 ... Modulation pattern generation unit, 25 ... Fourier transform unit, 26 ... Function substitution unit, 27 ... Wave function correction unit, 28 ... Inverse Fourier transform unit, 29 ... Target generation unit, 29a ... Fourier transform unit, 29b ... Spectrogram correction unit, 30 ... Terahertz wave generator, 30A, 30B ... Photoconduction Antenna element, 31 ... substrate, 32 ... first semiconductor layer, 33 ... insulating layer, 33a ... opening, 34, 36 ... electrode, 34a, 36a ... wiring section, 34b, 36b ... antenna section, 34c, 36c ... pad section , 35 ... second semiconductor layer, 37 ... protective film, 41 ... cylindrical mirror, 42 ... periscope array, 43 ... prism, 44 ... pulse stretcher, 44a, 44b ... diffraction grating, 44c ... mirror, B3 ... predetermined direction, B4 ... Spectral direction, D ... light diameter, f ... focal distance, L0 to L3 ... optical path, P1 ... single optical pulse, P2 ... multi optical pulse, P21 to P24 ... optical pulse, Q1 to Q4 ... focused spot, SC ... control signal , T1 to T3 ... Terahertz wave, TP1 to TP3 ... Polarization direction.

Claims (6)

A light source that outputs the first light pulse and
Optically coupled to the light source, a plurality of second light pulses having a time difference from each other and having different center wavelengths are generated from the first light pulse, and each of the plurality of second light pulses is generated according to the wavelength. A multi-pulse generator that outputs on different optical paths,
A terahertz wave generating element that is optically coupled to the multi-pulse generating unit and receives the plurality of second optical pulses output from the multi-pulse generating unit to generate a terahertz wave pulse.
Bei to give a,
The multi-pulse generator
A time waveform shaper that outputs the plurality of second optical pulses having a time difference from each other and having different center wavelengths on a common optical path, and a time waveform shaper.
An optical dispersion element that separates each of the plurality of second light pulses output from the time waveform shaper into different optical paths according to the wavelength.
A terahertz wave generator. The time waveform shaper is
A spectroscopic element that disperses the first optical pulse and
A spatial light modulator that generates a plurality of second light pulses having a time difference from each other and having different center wavelengths by performing phase modulation of the first light pulse after spectroscopy for each wavelength.
An optical system that collects the plurality of second light pulses, and
The terahertz wave generator according to claim 1 . The terahertz wave generator according to claim 1 or 2 , wherein the light dispersion element is a diffraction grating or a prism. The terahertz wave generator according to any one of claims 1 to 3 , wherein the time width of each second optical pulse is 100 femtoseconds or less. From the first light pulse, a plurality of second light pulses having a time difference from each other and having different center wavelengths are generated, and each of the plurality of second light pulses is output on different optical paths according to the wavelength. Steps and
A terahertz wave generation step in which a plurality of second optical pulses are input to a terahertz wave generating element to generate a terahertz wave pulse, and a terahertz wave generation step.
Only including,
The multi-pulse light generation step
A time waveform shaping step that outputs the plurality of second optical pulses having a time difference from each other and having different center wavelengths on a common optical path, and
An optical dispersion step that separates each of the plurality of second light pulses into different optical paths according to the wavelength.
How to generate terahertz waves , including . The time waveform shaping step is
The spectroscopic step of dispersing the first light pulse and
An optical modulation step of generating the plurality of second optical pulses having a time difference from each other and having different center wavelengths by performing phase modulation of the first optical pulse after spectroscopy for each wavelength.
A condensing step for condensing the plurality of second light pulses, and
5. The method for generating a terahertz wave according to claim 5 .