Spectroscopic measurement device and spectroscopic measurement method

Abstract

To provide a spectroscopic measurement device and method capable of acquiring a spectrum with a high efficiency and at a high speed.SOLUTION: A spectroscopic measurement device 1 includes: a light source part 3 for outputting a first light pulse at a first repetition frequency and outputting a second light pulse at a second repetition frequency different from the first repetition frequency; a polarization beam splitter 7 for forming a light pulse train having a group delay swept between a pump pulse and a probe pulse by superimposing the first light pulse on the second light pulse; a photodetector 97 for detecting the intensity of observation light emitted from a sample 93 by irradiating the sample with the light pulse train; and a computer 13 for acquiring a time domain interferogram on the basis of the intensity of the observation light and Fourier-transforming the time domain interferogram. The light source part 3 modulates a difference between the first repetition frequency and the second repetition frequency on the basis of the oscillation coherence life and group delay of the sample excited by the irradiation of the pump pulse.SELECTED DRAWING: Figure 3

Description

[Technical Field]

[0001] The present invention relates to a spectroscopic measurement device and a spectroscopic measurement method. [Background technology]

[0002] Dual-comb coherent anti-Stokes Raman scattering (DC-CARS) spectroscopy is known as one of the spectroscopic techniques for noninvasively identifying the chemical characteristics of a target object (see, for example, Non-Patent Document 1). In DC-CARS spectroscopy, a pair of ultrashort pulse lasers with slightly different repetition rates is spatially overlapped to generate an optical pulse train in which the group delay between the pump pulse and the probe pulse is automatically swept. This optical pulse train is then irradiated onto a sample. The pump pulse excites the sample's molecular vibrations, and the probe pulse excites the sample, obtaining a time-domain interferogram. The Raman spectrum of the sample can be obtained by Fourier transforming the time-domain interferogram. DC-CARS spectroscopy can rapidly investigate the characteristics of molecular vibrations over a wide frequency band (fingerprint region), and is therefore applied not only to basic research in chemistry and medicine but also to industrial fields such as pharmaceuticals and semiconductor testing. [Prior art documents] [Non-patent literature]

[0003] [Non-Patent Document 1] T. Ideguchi, T. Nakamura, S. Takizawa, M. Tamamitsu, S. Lee, K. Hiramatsu, V. Ramaiah-Badarla, J. Park, Y. Kasai, T. Hayakawa, S. Sakuma, F. Arai, and K. Goda, “Microfluidic single-particle coherent chemical analyzer with dual-comb Raman spectroscopy,” Opt. Lett. 43, p.4057-4060 (2018). Summary of the Invention [Problem to be solved by the invention]

[0004] However, in conventional DC-CARS spectroscopy, the duty cycle of spectrum acquisition is only <1% due to the mismatch between the laser pulse interval (>1 nm) and the coherence lifetime of the sample's molecular vibrations (~3 ps). This means that more than 99% of the energy of the laser pulse irradiated to the sample is wasted, resulting in a low spectrum acquisition rate and a low signal-to-noise ratio (SN ratio).

[0005] The present invention has been made in view of the above-mentioned problems, and an object of the present invention is to provide a spectroscopic measurement device and a spectroscopic measurement method that are capable of acquiring a spectrum with high efficiency and high speed. [Means for solving the problem]

[0006] A spectroscopic measurement device according to the present invention includes a light source unit that outputs first optical pulses at a first repetition rate and second optical pulses at a second repetition rate different from the first repetition rate, a superposition unit that forms pulse pairs consisting of pump pulses and probe pulses by superposing the first optical pulses and the second optical pulses and generates an optical pulse train in which the group delay between the pump pulses and the probe pulses is swept, a photodetector that detects the intensity of observation light emitted from a sample when the sample is irradiated with the optical pulse train, and a computer that acquires a time-domain interferogram based on the intensity of the observation light and converts the time-domain interferogram into a frequency domain to acquire a spectrum. The light source unit modulates the frequency difference between the first repetition rate and the second repetition rate based on the coherence lifetime and the group delay of vibrations in the sample excited by the irradiation of the pump pulses.

[0007] A spectroscopic measurement method according to the present invention includes: a light source unit outputs a first optical pulse at a first repetition rate and a second optical pulse at a second repetition rate different from the first repetition rate; a superposition unit superimposes the first optical pulse and the second optical pulse to form a pulse pair consisting of a pump pulse and a probe pulse; a light pulse train in which the group delay between the pump pulse and the probe pulse is swept; a photodetector detects the intensity of observation light emitted from a sample when the sample is irradiated with the optical pulse train; a computer acquires a time-domain interferogram based on the intensity of the observation light; and a computer converts the time-domain interferogram into a frequency domain to acquire a spectrum. When the light source unit outputs the first optical pulse and the second optical pulse, the frequency difference between the first repetition rate and the second repetition rate is modulated based on the coherence lifetime and the group delay of vibrations in the sample excited by the irradiation of the pump pulse. [Effects of the Invention]

[0008] According to the present invention, the spectrum of the observation light can be obtained efficiently and quickly by modulating the frequency difference between the first repetition frequency and the second repetition frequency based on the group delay between the pump pulse and the probe pulse and the coherence lifetime of the vibration of the sample excited by the irradiation of the pump pulse. [Brief explanation of the drawings]

[0009] [Figure 1] FIG. 1 is an explanatory diagram showing the conceptual difference between conventional DC-CARS spectroscopy (a) and the spectroscopic measurement method according to this embodiment (b). [Figure 2] FIG. 2 is a schematic diagram showing a pump pulse and a probe pulse incident on a sample. [Figure 3] 1 is a schematic diagram illustrating a configuration of a spectroscopic measurement device according to an embodiment of the present invention. [Figure 4] 3A and 3B are schematic diagrams illustrating the configurations of a first light source and a second light source. [Figure 5] 10 is a flowchart showing a spectrum calculation process executed by a computer according to the present embodiment. [Figure 6] 10 is a flowchart illustrating a group delay calculation process. [Figure 7A] 1 is a graph showing an example of a two-color interference signal measured at different frequencies. [Figure 7B] 10 is a graph showing a spectrum obtained by Fourier transforming a two-color interference signal and a mask function. [Figure 7C] 10 is a graph showing a two-color interference signal reconstructed by inverse Fourier transform of the masked spectrum. [Figure 7D] 10 is a graph showing the time dependence of the phase calculated from the reconstructed two-color interference signal. [Figure 7E] 10 is a graph showing the time dependence of group delay calculated using two-color interference. [Figure 8] 10 is a graph showing the results of a proof-of-principle experiment using the spectroscopic measurement device according to the present embodiment. [Figure 9] 10 is a graph showing the results of quantitative analysis using the spectroscopic measurement device according to the present embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0010] Hereinafter, an embodiment of the present invention will be described with reference to the drawings. First, with reference to FIG. 1, the conceptual difference between the conventional DC-CARS spectroscopy (a) and the spectroscopic measurement method according to this embodiment (denoted as "Quasi-DC-CARS" in FIG. 1) (b) will be described.

[0011] In conventional DC-CARS spectroscopy (Fig. 1a), a pair of ultrashort pulse lasers with slightly different repetition rates are used, each with a repetition rate of f rep1 , f rep2 is fixed, the frequency difference Δf rep (=f rep2 -f rep1) is constant. By spatially overlapping a pair of ultrashort pulse lasers, an optical pulse train consisting of a pump pulse and a probe pulse is generated. As shown in Figure 2, the probe pulse is irradiated onto the sample with a time τ (group delay) after the pump pulse. f rep1 <f rep2 Then, the refresh time is 1 / Δf rep The group delay τ varies from zero to 1 / f rep1 Based on the highest repetition rate of currently available lasers, rep1 ≒f rep2 At a frequency of ≈1 GHz, the group delay τ exceeded the coherence lifetime CL (~3 ps) of the molecular vibrations of the sample for more than 99% of the spectrum acquisition time (refresh time), resulting in a duty cycle of just under 1%.

[0012] In contrast, in the spectroscopic measurement method according to this embodiment, as shown in FIG. 1b, the frequency difference Δf is set so that the group delay τ is equal to or less than the coherence lifetime CL of molecular vibration. rep (=f rep2 -f rep1 ) at high speed to make the duty cycle of spectrum acquisition almost 100%. Specifically, the frequency difference Δf rep The frequency difference Δf rep The frequency f mod When modulated with , the group delay τ is the refresh time 1 / (2f mod ) is swept between zero and the coherence lifetime CL. The frequency difference Δf rep Ideally, it would be desirable to rapidly modulate τ like a square wave function and vary the group delay τ like a triangular wave function. However, in reality, the variation in the group delay τ is distorted due to the high-speed modulation. Therefore, in this embodiment, as described below, the group delay τ is measured using two-color interference, and the obtained group delay τ is used to calibrate the time-domain interferogram.

[0013] <Configuration of spectroscopic measurement device> Next, the configuration of an apparatus for implementing the spectroscopic measurement method according to this embodiment will be described with reference to Figures 3 and 4. As shown in Figure 3, the spectroscopic measurement apparatus 1 according to this embodiment includes a light source unit 3, a mirror 6, a polarizing beam splitter 7, a scattering signal acquisition unit 9, a two-color interference signal generation unit 11, and a computer 13.

[0014] The light source unit 3 emits light at a first repetition frequency f rep1 and a first light source 4A that outputs a first optical pulse at a second repetition frequency f rep2 a second light source 4B that outputs a second light pulse, a half-wave plate 5A that adjusts the polarization direction of the first light pulse from the first light source 4A, and a half-wave plate 5B that adjusts the polarization direction of the second light pulse from the second light source 4B.

[0015] The polarization direction of the first optical pulse output from half-wave plate 5A and the polarization direction of the second optical pulse output from half-wave plate 5B are perpendicular to each other, which makes the polarization direction of the probe pulse perpendicular to the polarization direction of the pump pulse, as shown in Figure 2, thereby eliminating the influence of the pump pulse scattered from the sample.

[0016] The first light pulse from half-wave plate 5A is reflected by mirror 6 and output to polarizing beam splitter 7. The second light pulse from half-wave plate 5B is also output to polarizing beam splitter 7.

[0017] 4, the first light source 4A and the second light source 4B have the same configuration, and include a function generator (FG) 41, a diode-pumped solid-state (DPSS) laser 42, a first mirror 43a, a second mirror 43b, a third mirror 43c, a fourth mirror 43d, a gain medium 44, a piezoelectric element 45, a beam splitter 46, a signal generator 47, a phase shifter 48, a photodetector 49, a mixer 50, a low-pass filter 51, a servo controller 52, and an amplifier 53. Each of the first light source 4A and the second light source 4B is, for example, a mode-locked laser light source operating at a repetition rate of approximately 1 GHz.

[0018] The DPSS laser 42 is selected according to the type of gain medium 44, and outputs a laser beam of a predetermined wavelength according to the current applied from the FG 41. For example, if the gain medium 44 is a titanium sapphire (Ti:Sapphire) crystal, a laser beam with a wavelength of 532 nm can be used as the DPSS laser 42.

[0019] The first mirror 43a and the second mirror 43b are concave mirrors, and are arranged so that their respective reflective surfaces face each other. A gain medium 44 is arranged between the first mirror 43a and the second mirror 43b. The third mirror 43c and the fourth mirror 43d are planar mirrors, and are arranged so that their respective reflective surfaces face each other. The reflective surface of the third mirror 43c also faces the reflective surface of the second mirror 43b, and the reflective surface of the fourth mirror 43d also faces the reflective surface of the first mirror 43a. A piezoelectric element 45 is attached to the third mirror 43c, making the position of the third mirror 43c adjustable.

[0020] The laser light from the DPSS laser 42 passes through the first mirror 43a and enters the gain medium 44. When the laser light that passed through the first mirror 43a is irradiated onto the gain medium 44, the gain medium 44 is pumped, and the amplified laser light is emitted by stimulated emission. The laser light emitted from the gain medium 44 is reflected by the second mirror 43b, the third mirror 43c, the fourth mirror 43d, and the first mirror 43a in that order, and is again irradiated onto the gain medium 44. In this way, the laser circulates along the bowtie ring optical path, and the laser light of the target wavelength passes through the second mirror 43b and enters the beam splitter 46.

[0021] The beam splitter 46 is disposed on the output side of the second mirror 43b and splits the laser from the second mirror 43b into two beams. One beam from the beam splitter 46 is sent to the outside, and the other beam is incident on a photodetector 49 for feedback control. The photodetector 49 detects the intensity of the incident beam and photoelectrically converts the detected intensity to obtain an electrical signal. The electrical signal obtained by the photodetector 49 is output to a mixer 50.

[0022] A signal generator 47 generates a target waveform signal. The waveform signal generated by the signal generator 47 is phase-adjusted by a phase shifter 48 and then output to a mixer 50. The mixer 50 mixes the electrical signal from the photodetector 49 with the waveform signal from the phase shifter 48 and outputs the difference between the two signals as a feedback signal. High-frequency noise is removed from the feedback signal by a low-pass filter 51, and the feedback signal is output to a servo controller 52. The cutoff frequency of the low-pass filter 51 is, for example, 1 GHz.

[0023] The servo controller 52 outputs a drive signal to the amplifier 53 to drive the piezoelectric element 45 in accordance with the input feedback signal. The amplifier 53 amplifies the drive signal and outputs it to the piezoelectric element 45. The piezoelectric element 45 adjusts the position of the third mirror 43c in accordance with the input drive signal. Adjusting the position of the third mirror 43c changes the optical path length. In this embodiment, for example, feedback control is performed using proportional-integral (PI) control so that the difference between the waveform signal from the signal generator 47 and the electrical signal corresponding to the laser output is minimized.

[0024] In this embodiment, the optical path length is set so that the repetition frequency of the optical pulses emitted from each of the first light source 4A and the second light source 4B is approximately 1 GHz, but the optical path length is finely adjusted by feedback-controlling the position of the third mirror 43c to obtain the target repetition frequency.

[0025] The repetition frequency of one of the first light source 4A and the second light source 4B is fixed, and the repetition frequency of the other is modulated. In the following, the first repetition frequency f rep1 is a fixed value, and the second repetition frequency f rep2 is modulated.

[0026] In the first light source 4A, the first repetition frequency f rep1 is fixed.

[0027] In the second light source 4B, the second repetition frequency f is modulated by the modulation function g(t) as shown in equation (1) so that the group delay τ is equal to or less than the coherence lifetime CL of molecular vibration. rep2 The first repetition frequency f rep1 (i.e., around the first repetition frequency f rep1 The modulation function g(t) is the frequency difference Δf rep (=f rep2 -f rep1 ) is a square wave function corresponding to the frequency difference Δf rep Equation (2) is satisfied so that the positive and negative fluctuations of are equal.

number

[0028] To achieve this modulation, the second light source 4B not only feedback controls the position of the third mirror 43c as described above, but also adjusts the strength of the current applied to the DPSS laser 42 by the FG 41 to rapidly control the intensity of the laser that excites the gain medium 44. This causes the refractive index of the gain medium 44 to change due to the nonlinear optical Kerr effect, which in turn changes the optical path length in the gain medium 44. In this way, the second repetition frequency f rep2 For example, the second repetition frequency f rep2 at a frequency of 50 kHz mod It can be modulated by.

[0029] Returning to Figure 3, the polarizing beam splitter 7 functions as a superposition unit that forms a pulse pair consisting of a pump pulse and a probe pulse by superposing the first and second optical pulses output from the light source unit 3, and generates an optical pulse train in which the group delay between the pump pulse and the probe pulse is swept. The polarizing beam splitter 7 also splits this optical pulse train into a first beam 8A for irradiating the sample 93 and a second beam 8B for measuring the group delay. The first beam 8A is sent to the scattering signal acquisition unit 9, and the second beam 8B is sent to the two-color interference signal generation unit 11.

[0030] The scattering signal acquisition unit 9 includes a compensator 90, a long-pass filter 91, a first objective lens 92, a second objective lens 94, a short-pass filter 95, a polarizer 96, a photodetector 97, and a low-pass filter 98.

[0031] The compensator 90 compensates for the group velocity dispersion of each of the pump pulse and probe pulse of the first beam 8A incident from the polarizing beam splitter 7. The compensator 90 minimizes the pulse width of each of the pump pulse and probe pulse at the position of the sample 93. By compensating for the group velocity dispersion of the pump pulse and the probe pulse in this way, Raman scattering is more likely to occur in the sample 93, and a spectrum can be acquired more reliably.

[0032] The compensator 90 has a pair of chirp mirrors, a first chirp mirror 90a and a second chirp mirror 90b, which are arranged so that their mirror surfaces are parallel to each other. The first beam 8A incident on one end of the compensator 90 is propagated while being repeatedly reflected between the first chirp mirror 90a and the second chirp mirror 90b, and is sent out from the other end of the compensator 90 toward the long-pass filter 91.

[0033] Note that various devices can be used for the compensator 90 as long as they can compensate for the group velocity dispersion of the pump pulse and probe pulse of the first beam 8A and reduce the pulse width. For example, a diffraction grating pair, a photonic crystal fiber (PCF), a chirped fiber Bragg grating (FBG), etc. may be used. While these, including the chirped mirror pair, compensate for group velocity dispersion, i.e., up to second-order dispersion, a compensator 90 such as a prism pair or grism pair that compensates for higher-order (third-order or higher) dispersion may also be used.

[0034] The long-pass filter 91 has a cutoff wavelength set so as not to pass light at the short-wavelength tail of the spectrum of the incident first beam 8A. The long-pass filter 91 cuts out light of a wavelength similar to that of the anti-Stokes light from the sample 93 from the first beam 8A, making it easier to detect the anti-Stokes light and enabling a more reliable spectrum to be obtained. The cutoff wavelength of the long-pass filter 91 is, for example, 780 nm.

[0035] The first beam 8A that has passed through the long-pass filter 91 is incident on a first objective lens 92. The first objective lens 92 and the second objective lens 94 are arranged opposite each other so that their focal positions coincide, and a sample 93 is placed at that focal position. The first objective lens 92 and the second objective lens 94 can be, for example, an achromatic lens with a focal length of 7.5 mm.

[0036] The first objective lens 92 focuses the incident pump pulse and probe pulse and irradiates them onto the sample 93. When the pump pulse is irradiated onto the sample 93, light of a certain frequency contained in the pump pulse becomes a substantial pump pulse that excites the substance with respect to the molecules of the sample 93, and light of a certain frequency different from this pump pulse becomes Stokes light, which is equivalent to these being irradiated onto the sample 93. As a result, molecular vibrations are induced in the sample 93 at a vibration frequency corresponding to the frequency difference between the pump pulse and the Stokes light.

[0037] When the probe pulse is irradiated onto the sample 93 after the pump pulse, the probe pulse interacts with the induced molecular vibration, shifting the frequency of the light contained in the probe pulse and emitting scattered light. This scattered light includes anti-Stokes light, whose frequency is increased relative to the frequency of the probe pulse by the frequency of the molecular vibration induced by the substantial pump pulse and Stokes light, and Stokes light, whose frequency is decreased relative to the frequency of the probe pulse by the frequency of the molecular vibration induced by the substantial pump pulse and Stokes light.

[0038] The intensities of the anti-Stokes light and Stokes light scattered from the sample 93 change depending on the delay time (group delay) of the probe pulse relative to the pump pulse. Therefore, a time-domain interferogram of the anti-Stokes light is acquired by irradiating the sample 93 with the first beam 8A, in which the group delay between the pump pulse and the probe pulse has been swept. The pump pulse and the probe pulse are broadband optical pulses and have a wide range of wavelength components. Therefore, pump pulses of various frequencies induce vibrations in various molecules, and anti-Stokes light and Stokes light corresponding to these molecular vibrations are emitted. Therefore, the acquired time-domain interferogram includes a spectrum over a wide wavenumber range.

[0039] The second objective lens 94 collimates the anti-Stokes light and Stokes light from the sample 93 together with the first beam 8A that has passed through the sample 93. These collimated light beams are incident on a short-pass filter 95.

[0040] The short-pass filter 95 has a cutoff wavelength set to a wavelength shorter than or equal to the cutoff wavelength of the long-pass filter 91. Most of the wavelength components of the pump pulse and probe pulse of the first beam 8A that are shorter than the cutoff wavelength are cut off by the long-pass filter 91. Therefore, most of the pump pulse and probe pulse that have passed through the sample 93 are cut off by the short-pass filter 95.

[0041] Moreover, since the Stokes light is a specific frequency component of the probe pulse frequency-shifted to the lower frequency side, most of it is cut by the short-pass filter 95. On the other hand, most of the anti-Stokes light, whose frequency is increased relative to the probe pulse, passes through the short-pass filter 95. In this way, of the light scattered from the sample 93, only the anti-Stokes light passes through the short-pass filter 95 as the observation light. The anti-Stokes light that has passed through the short-pass filter 95 is incident on the polarizer 96. The cutoff wavelength of the short-pass filter 95 is, for example, 780 nm.

[0042] The polarizer 96 passes the anti-Stokes light having the same polarization direction as the probe pulse. The anti-Stokes light that has passed through the polarizer 96 is incident on the photodetector 97.

[0043] The photodetector 97 receives the anti-Stokes light that has passed through the short-pass filter 95 and the polarizer 96, detects the intensity, and converts the detected intensity into an electrical signal. The obtained electrical signal is output to the low-pass filter 98. The photodetector 97 may be, for example, an avalanche photodiode, a photomultiplier tube PIN photodiode, or the like, but is not particularly limited thereto.

[0044] The low-pass filter 98 is connected to the photodetector 97 via a conductor and removes high-frequency noise from the electrical signal output from the photodetector 97. The low-pass filter 98 is also connected to the computer 13 via another conductor, and the electrical signal from which high-frequency noise has been removed by the low-pass filter 98 is output to the computer 13. The cutoff frequency of the low-pass filter 98 is, for example, 530 MHz.

[0045] The two-color interference signal generating unit 11 is a device for generating a two-color interference signal required to measure the group delay between the pump pulse and the probe pulse, and as shown in FIG. 3, includes a polarizer 110, a diffraction grating 112, a lens 114, photodetectors 116a and 116b, and low-pass filters 118a and 118b.

[0046] The polarizer 110 passes light of a specific polarization direction from the second beam 8B from the polarizing beam splitter 7. The specific polarization direction is preferably set at an angle such that the pump pulse and the probe pulse pass in equal amounts (i.e., 45° or −45° with respect to the polarization direction of the probe pulse or the pump pulse). The second beam 8B that has passed through the polarizer 110 is sent to the diffraction grating 112.

[0047] Diffraction grating 112 is an optical element that disperses second beam 8B into two beams with different wavelengths (with frequencies f1 and f2, respectively). These two beams are sent to lens 114. Diffraction grating 112 is disposed at the focal position (position of focal length f) of lens 114. For example, a diffraction grating having 1200 grooves per mm can be used as diffraction grating 112.

[0048] The lens 114 focuses the two beams from the diffraction grating 112 onto the photodetectors 116a and 116b, respectively. The lens 114 may be, for example, an achromatic lens with a focal length f of 200 mm.

[0049] The photodetectors 116a and 116b are arranged on the opposite side of the lens 114 from the diffraction grating 112, and are spaced apart from the lens 114 by a focal length f. The photodetectors 116a and 116b detect the intensities of the two beams from the lens 114, respectively, and convert the intensities into electrical signals. The photodetectors 116a and 116b may be, for example, photodetectors with an effective area of ​​0.126 mm 2 A photodetector of the type can be employed.

[0050] Low-pass filter 118a is connected to photodetector 116a via a conductor and removes high-frequency noise from the electrical signal output from photodetector 116a. Low-pass filter 118b is connected to photodetector 116b via a conductor and removes high-frequency noise from the electrical signal output from photodetector 116b.

[0051] The low-pass filters 118a and 118b are also connected to the computer 13 via respective conductors. The electrical signals from which high-frequency noise has been removed by the low-pass filters 118a and 118b are output to the computer 13. The cutoff frequencies of the low-pass filters 118a and 118b are, for example, 600 MHz.

[0052] The computer 13 includes an A / D converter 131 , a memory 133 , and a processor 135 .

[0053] The A / D converter 131 samples the electrical signal of the anti-Stokes light output from the low-pass filter 98 and the electrical signals output from the low-pass filters 118a and 118b, and converts these electrical signals into digital signals. The resulting digital signals are sequentially stored in memory 133 together with the measurement time t. In this way, a time-domain interferogram (CARS interferogram) of the anti-Stokes light and two-color interference signals measured at frequencies f1 and f2 are acquired.

[0054] The memory 133 stores programs executed by the processor 135 and data required when the programs are executed.

[0055] The processor 135 performs a Fourier transform on the time-domain interferogram of the anti-Stokes light to obtain a spectrum in accordance with a program stored in the memory 133. Specifically, the processor 135 calculates the group delay τ between the pump pulse and the probe pulse from the two-color interference signal measured at frequencies f1 and f2, calibrates the time-domain interferogram using this group delay τ, and performs a Fourier transform on the calibrated time-domain interferogram to calculate a spectrum. Details of the method for calculating the spectrum will be described later (see FIGS. 5 to 7E).

[0056] 3 may be configured without the half-wave plates 5A and 5B. In this case, if a non-polarizing beam splitter is used instead of the polarizing beam splitter 7, the spectroscopic measurement method according to this embodiment can be realized.

[0057] 3, the long-pass filter 91 is provided on the input side of the first objective lens 92, and the short-pass filter 95 is provided on the output side of the second objective lens 94, but the arrangement of the long-pass filter 91 and the short-pass filter 95 is not particularly limited. It is sufficient that the combination of the long-pass filter 91 and the short-pass filter 95 blocks transmission of the pump pulse and the probe pulse from the sample 93. For example, the short-pass filter 95 may be provided on the input side of the first objective lens 92, and the long-pass filter 91 may be provided on the output side of the second objective lens 94.

[0058] <Spectrum calculation process> Next, the spectrum calculation process executed by the computer 13 will be described with reference to FIGS. 5 to 7E.

[0059] As shown in Fig. 5, first, the A / D converter 131 simultaneously acquires a CARS interferogram of the anti-Stokes light scattered from the sample 93 and two-color interference signals measured at different frequencies f1 and f2 (step 502). If the two-color interference signals measured at frequencies f1 and f2 are s1(t) and s2(t), respectively, they are given by equation (3). Also, Fig. 7A shows an example of s1(t) and s2(t).

number

[0060] c in Equation (3) j (t) is a real number b j (t) and phase Φ j Using (t), it is defined as in equation (4).

number

[0061] As is clear from FIG. 7A, equations (3) and (4), s j (t) is the frequency f j The amplitude of the light is cos[Φ j The waveform fluctuates periodically, as shown in (t)]. The sample 93 is irradiated with a train of optical pulses obtained by the interference of the first and second optical pulses. If the two waves are in phase (zero phase difference), the amplitude increases; if they are out of phase (π phase difference), they cancel each other out, resulting in a minimum amplitude. In other words, the fluctuation in the amplitude of the interference wave (Equation (4)) corresponds to the interval (group delay) between the pump pulse and the probe pulse.

[0062] Therefore, the processor 135 executes a process of calculating the group delay from the two-color interference signals s1(t) and s2(t) acquired in step 502 (step 504 and FIG. 6). As shown in FIG. 6, in the group delay calculation process, the processor 135 first performs a Fourier transform on each of the two-color interference signals s1(t) and s2(t) (step 602). When equation (3) is Fourier transformed, the spectrum s Fj (f) is obtained.

number

[0063] In the frequency domain, -f j , 0, +f j The spectrum s1(t) is obtained by Fourier transforming it. F1 An example of (f) is shown below.

[0064] Next, the processor 135 applies a mask function to the spectrum to extract the spectrum of the required frequency components (step 604). Here, as shown in FIG. 7B, among the three peaks of the spectrum, +f j Extract only the spectrum corresponding to the masked spectrum s Frecj(f) is expressed as in equation (6): As shown in equation (6) and Fig. 7B, masking is to make the spectral intensity of unnecessary frequency components zero, that is, to remove the unnecessary frequency components.

number

[0065] Next, the processor 135 performs an inverse Fourier transform on the masked spectrum to reconstruct the two-color interference signal (step 606). Frecj By inverse Fourier transforming (f), the reconstructed two-color interference signal s is obtained as shown in equation (7). recj (t) is obtained.

number

[0066] The processor 135 then processes the reconstructed two-color interference signal s recj (t) to phase Φ j (t) is calculated (step 608). According to Euler's formula, the real part and imaginary part of equation (7) are expressed as equations (8) and (9), respectively. FIG. 7C shows the reconstructed two-color interference signal s rec1 An example of (t) is shown in FIG. rec1 The real and imaginary parts of (t) are shown by the solid and dotted lines, respectively.

number

[0067] The reconstructed two-color interference signal s recj Dividing the imaginary part of (t) by the real part gives equation (10), from which the phase Φ j (t) is given by equation (11). Fig. 7D shows an example of the phase Φ1(t) and the phase Φ2(t) calculated by equation (11).

number

[0068] Next, the processor 135 calculates the frequency f j and the phase Φ calculated in step 608 j (t), the group delay is calculated (step 610). j (angular frequency ω j =2πf j ) the two-color interference signal s measured i Phase Φ of (t) j (t) is expressed as ω j can be expressed as a Taylor expansion around a certain angular frequency ω0 close to

number

[0069] φ in equation (12) (1) (t) represents the group delay, where φ (n) If we ignore the terms n ≥ 2, the group delay φ (1) (t) is given by equation (13).

number

[0070] FIG. 7E shows the first repetition frequency f rep1 and the second repetition frequency f rep2 The group delay (true value; dotted line) obtained from the measured repetition rate is fixed at 1000.200 MHz and 1000.190 MHz, respectively. The group delay (solid line) calculated from equation (13) using two-color interference is shown. The error between the calculated group delay and the true value is 17.5 fs, which is only 0.146 cm in the Raman spectral region. -1 Since the error is equivalent to the error of

[0071] Once the group delay is calculated in step 610, processor 135 uses the calculated group delay to calibrate the CARS interferogram acquired in step 502 (step 506 of FIG. 5).

[0072] The CARS interferogram acquired in step 502 and the group delay calculated in step 610 are respectively expressed as E CARS =E CARS (t), τ=d g (t). The measurement time t is expressed as the inverse function d g -1 That is, the measurement time t is expressed as a function of the group delay τ.

number

[0073] From equation (14), the CARS interferogram E CARS Converting (t) into a function of group delay τ gives the calibrated CARS interferogram E´ as shown in equation (15). CARS This corresponds to resampling the intensity of the anti-Stokes light as the observed light so that the group delay τ is at equal intervals.

number

[0074] The processor 135 then generates the calibrated CARS interferogram E' CARS (τ) is Fourier transformed to obtain the spectrum (step 508), which completes the spectrum calculation process. [Example]

[0075] Next, as an example, the results of an experiment using the spectroscopic measurement device 1 will be described with reference to FIGS.

[0076] In Fig. 8, liquid toluene was used as a sample, and the second repetition rate f was set so that the group delay between the pump pulse and the probe pulse was less than the coherence lifetime of molecular vibration (~3 ps). rep2 at a frequency of 50 kHz mod The results of a proof-of-principle experiment in which a spectrum was acquired by modulating the signal with

[0077] FIG. 8a shows the reconstructed two-color interference signal s obtained by dispersing the second beam 8B at wavelengths 767 nm and 816 nm using a diffraction grating 112. rec1 (t) and s rec2 (t). The inset in Fig. 8a shows the reconstructed two-color interference signal s rec1 (t) and s rec2 (t) (color 1 and color 2). Fig. 8b shows the reconstructed two-color interference signal s rec1 (t) and s rec2 Figure 8c shows the group delay calculated from (t), and Figure 8d shows the CARS interferogram calibrated using the calculated group delay. Figure 8d shows the spectrum obtained by Fourier transform of the calibrated CARS interferogram.

[0078] From Figure 8d, the spectrum obtained is -1 , 786cm -1 , 1004cm -1 , and 1210 cm -1 It can be seen that a Raman peak appears at 100,000 spectra / s. The spectral acquisition rate in this experiment was 10 times higher than the highest reported value for conventional DC-CARS spectroscopy (10,000 spectra / s) and twice higher than the highest reported value for Fourier transform coherent anti-Stokes Raman scattering (FT-CARS) spectroscopy (50,000 spectra / s). At 100,000 spectra / s, the group delay sweep range was 0.71 ps, as shown in Figure 8b, resulting in a group delay of 23.4 cm. -1 corresponds to a spectral resolution of

[0079] Although the spectral resolution deteriorates due to the shortened sweep range of the group delay, the energy efficiency for measuring the time-domain interferogram is nearly 100%. This not only increases the spectral acquisition rate as described above, but also allows the acquisition of spectral features that could not be obtained by conventional DC-CARS spectroscopy (e.g., 532 cm -1 and 1210 cm -1 It is clear that the Raman peaks at the nuclei of the nuclei are also recognizable, improving the sensitivity.

[0080] 9A shows the peak at approximately 1004 cm for the conventional DC-CARS spectroscopy and the spectroscopic measurement method according to this embodiment (hereinafter referred to as "Quasi-DC-CARS spectroscopy"). -1 It shows a peak at about 1780 cm -1 The graph shows the signal-to-noise ratio (SNR) of toluene spectra with a standard deviation of 0.01 at various spectral acquisition rates. Fitting of the measured values ​​reveals that for both spectroscopy methods, the SNR is proportional to the spectral acquisition rate (with a slope of approximately -1 / 2) in logarithmic terms, and that the SNR decreases as the spectral acquisition rate increases.

[0081] In addition, Quasi-DC-CARS spectroscopy has a higher signal-to-noise ratio than conventional DC-CARS spectroscopy; for example, the signal-to-noise ratio at 100,000 spectra / second is 20 times higher (assuming that conventional DC-CARS spectroscopy is capable of 100,000 spectra / second). In other words, the signal-to-noise ratio of Quasi-DC-CARS spectroscopy at 100,000 spectra / second is equivalent to that of conventional DC-CARS spectroscopy at 200 spectra / second.

[0082] Figure 9b shows the sample concentration dependence of the S / N ratio when toluene in an ethanol solution is used as the sample in Quasi-DC-CARS spectroscopy. Fitting of the measured values ​​reveals that the S / N ratio increases quadratically as the concentration increases. This makes it possible to estimate the sample concentration from the S / N ratio.

[0083] The spectra acquired for each sample concentration shown in Figure 9b are shown in Figure 9c. Figures 9b and 9c show that with Quasi-DC-CARS spectroscopy, a spectrum acquisition rate of 100,000 spectra / second can obtain spectra with sufficient S / N ratio even at low concentrations.

[0084] As described above, according to this embodiment, the first repetition frequency f is determined based on the group delay between the pump pulse and the probe pulse and the coherence lifetime CL of the molecular vibration of the sample 93. rep1 and the second repetition frequency f rep2Frequency difference Δf rep By modulating the signal at high speed, the spectrum of the observed light can be obtained efficiently and quickly.

[0085] In particular, the frequency difference Δf is set so that the group delay is equal to or less than the coherence lifetime CL of the molecular vibration of the sample 93. rep By modulating the ion beam (Fig. 1b), the energy efficiency for measuring the time-domain interferogram becomes nearly 100%, which increases the spectrum acquisition rate and improves the sensitivity. The time change of the spectrum obtained by the method of this embodiment is expected to enable non-invasive and highly sensitive observation of the mechanism of chemical reactions that develop on the order of microseconds.

[0086] It should be noted that various modifications can be made to this embodiment without departing from the spirit of the present invention. For example, in the above-described embodiment, the spectrum of anti-Stokes light from the sample is acquired, but it is also possible to acquire the spectrum of Stokes light from the sample.

[0087] Furthermore, in the above-described embodiment, forward scattered light from the sample is used as the observation light, but the spectrum may be acquired using back scattered light (reflected light) as the observation light.

[0088] Furthermore, the above-described embodiment is not limited to DC-CARS spectroscopy, but can be applied to any spectroscopy in which a sample is irradiated with a train of optical pulses in which the group delay between the pump pulse and the probe pulse is swept, the sample is excited by the pump pulse, and the sample is observed by the probe pulse. [Explanation of symbols]

[0089] 1 Spectrometer 3 Light source section 4A 1st light source 4B 2nd light source 7 Polarizing Beam Splitter 8A First Beam 8B Second beam 9 Scattered signal acquisition section 11 Two-color interference signal generator 13. Computer 97, 116a, 116b Photodetectors 112 Diffraction Grating 131 A / D converter 133 memory 135 processors

Claims

1. a light source unit that outputs a first optical pulse at a first repetition frequency and a second optical pulse at a second repetition frequency different from the first repetition frequency; a beam splitter as a superposition unit that forms a pulse pair consisting of a pump pulse and a probe pulse by superposing the first optical pulse and the second optical pulse, generates an optical pulse train in which the group delay between the pump pulse and the probe pulse is swept, and splits the optical pulse train into a first beam for irradiating a sample and a second beam for measuring the group delay; a photodetector for detecting the intensity of observation light emitted from the sample when the first beam is irradiated onto the sample; a computer that acquires a time domain interferogram based on the intensity of the observed light, and converts the time domain interferogram into a frequency domain to acquire a spectrum; an optical element that disperses the second beam into two beams having different optical frequencies; a second photodetector and a third photodetector for detecting the respective intensities of the two beams dispersed from the second beam; Equipped with the light source unit modulates a frequency difference between the first repetition frequency and the second repetition frequency so that the group delay is equal to or less than a coherence lifetime of vibrations of the sample excited by irradiation with the pump pulse; The computer calculating a phase representing a periodic variation in amplitude for each of the two beams dispersed from the second beam from a two-color interference signal representing the intensities of the two beams dispersed from the second beam; calculating the group delay from the frequencies and the phases of the two beams dispersed from the second beam; calibrating the time domain interferogram using the calculated group delay to obtain a calibrated time domain interferogram; A spectroscopic measurement device that converts the calibrated time domain interferogram into a frequency domain to obtain the spectrum.

2. The spectroscopic measurement device according to claim 1 , wherein the light source unit fixes one of the first repetition frequency and the second repetition frequency, and modulates the other repetition frequency around the one repetition frequency.

3. The computer obtaining a reconstructed two-color interference signal by removing unwanted frequency components from the two-color interference signal; calculating the phase from the reconstructed two-color interference signal; The spectroscopic measurement device according to claim 1 .

4. outputting a first optical pulse at a first repetition frequency and a second optical pulse at a second repetition frequency different from the first repetition frequency by a light source unit; a beam splitter as a superposition unit superimposing the first optical pulse and the second optical pulse to form a pulse pair consisting of a pump pulse and a probe pulse, generating an optical pulse train in which the group delay between the pump pulse and the probe pulse is swept, and splitting the optical pulse train into a first beam for irradiating a sample and a second beam for measuring the group delay; detecting, by a photodetector, the intensity of observation light emitted from the sample as a result of irradiation of the first beam onto the sample; a computer to obtain a time domain interferogram based on the intensity of the observed light, and convert the time domain interferogram into a frequency domain to obtain a spectrum; dispersing the second beam into two beams of different optical frequencies by an optical element; detecting the intensities of the two beams dispersed from the second beam by a second photodetector and a third photodetector; modulating a frequency difference between the first repetition frequency and the second repetition frequency so that the group delay becomes equal to or less than a coherence lifetime of vibrations of the sample excited by irradiation with the pump pulse when the first light pulse and the second light pulse are output by the light source unit; by the computer calculating a phase representing a periodic variation in amplitude for each of the two beams dispersed from the second beam from a two-color interference signal representing the intensities of the two beams dispersed from the second beam; calculating the group delay from the frequencies and the phases of the two beams dispersed from the second beam; calibrating the time domain interferogram using the calculated group delay to obtain a calibrated time domain interferogram; A method of spectroscopy, wherein the calibrated time domain interferogram is transformed into a frequency domain to obtain the spectrum.

Images