Infrared (IR) spectroscopy is used in various scientific and industrial fields. In particular, attenuated total reflectance Fourier-transform infrared spectroscopy (ATR-FTIR) is a powerful tool to study liquid or solid samples [1]. In a typical ATR-FTIR, an IR beam passes through a total reflectance prism in such a way that it reflects at least once off the internal surface in contact with the sample. This reflection forms the evanescent wave which penetrates into the sample. The penetration depth into the sample is typically between 0.5 and 2 μm. The beam is then introduced into a Fourier-transform infrared spectrometer. In this method, little or no sample preparation is required and samples can be set under various conditions, for example, high pressure, low temperature, etc.

ATR-FTIR has been used to obtain fruitful information for understanding biological functions, such as energy conversion and signal transduction [2, 3]. Recently, a new time-resolved method for ATR-FTIR spectroscopy has been developed, which can be utilized for many biological systems just by exchanging a solution w/o a substrate perfused on the sample to initiate the reaction [4]. However, the time resolution of the Fourier-transform infrared spectroscopy (FTIR) with the rapid scan mode is limited by the speed of the FTIR device which needs time (at least several milliseconds) for scanning the delay of the interferometer. The step scan method could be possible to improve the time resolution to the response time of the detector (∼10 ns for a typical photovoltaic MCT detector), but it is only applicable for a system which can be triggered repeatedly (more than several hundreds).

Chirped-pulse upconversion is a noteworthy method to improve the time resolution of IR spectroscopy. In this method, an infrared beam is upconverted into a visible beam and the spectrum is measured with a high performance visible dispersive spectrometer. Nowadays, optical parametric amplifiers which produce IR pulses with a duration of femtosecond regime [5–7] are commercially available. Such intense IR pulses are easily converted into visible pulses. So far, several varieties of IR spectroscopy with chirped-pulse upconversion were reported [8–11]. Recently, it becomes possible to obtain IR spectra which spread from 200 to 5500 cm⁻¹ within 1 ms [12].

In this paper, we report attenuated total reflectance (ATR) spectroscopy connected with chirped-pulse upconversion. Single-shot IR absorption spectrum measurement of liquids from 200 to 5500 cm⁻¹ has been realized. To demonstrate the benefit of the high-speed IR spectrum acquisition, we have observed the dynamics of acetone-water liquid exchange process with a rapid solution exchange system by measuring the change of the IR spectrum of the mixture. We observed the dynamic change of the infrared absorption bands in the millisecond time scale and estimated that the exchange was complete within ∼10 ms.

The schematic of the experimental setup is shown in Fig. 1(a). The light source is a Ti:sapphire multi-pass amplifier (790 nm, 30 fs, 0.85 mJ at 946 Hz, Femtopower compact-Pro, FEMTOLASERS). The beam was split into two with a beam splitter (15% reflection). The transmitted beam was used to generate an ultrabroadband IR pulse and the reflected beam was used as a chirped-pulse.

 

Fig. 1 (a) Schematic of the ATR system with chirped-pulse upconversion. BBO: β-BaB₂O₄ crystal (Type 1, θ = 29°, t = 100 μm), DP: delay plate (calcite crystal, t = 1.7 mm), DWP: dual wave plate (λ at 400 nm, λ/2 at 800 nm), CM1: r = 1 m concave dielectric mirror, CM2: r = 0.5 m concave mirror with a hole (ϕ = 7 mm), CM3, CM4: r = 1 m concave silver mirror, MH: aluminium mirror with a hole (ϕ = 7 mm), ITO: indium tin oxide coated plate, ATR: ATR top-plate (Golden Gate ATR, Specac), OP: off-axis parabola (f = 50 mm), BF: blue filter. (b) Side and (c) top views of the ATR system with the solution-exchange system.

Download Full Size | PDF

The reflected pulse was stretched by passing through two zinc selenide (ZnSe) plates with the thickness of 10 mm with a Brewster angle, which corresponds to the group delay dispersion (GDD) of ∼22,000 fs². To avoid nonlinear effects in the crystal, we expanded the beam to the diameter of 15 mm. The pulse duration of the chirped-pulse was estimated as 4.9 ps by measuring the spectrally resolved cross-correlation between the chirped-pulse and the IR pulse [12]. The value of the instantaneous frequency was used for retrieving IR spectra from the upconverted spectra. The frequency resolution of the system is estimated as 4.2 cm⁻¹ from the duration of the chirped-pulse.

The IR pulse was generated by using four-wave mixing through two-color filamentation. The detail of the generation scheme is shown in the references [13, 14]. The generated pulse was reflected by two indium tin oxide coated substrates, which reflect the IR pulse and transmit the other visible beams. The IR pulse was gently focused into a diamond prism attached in the ‘ATR top-plate’ (Golden Gate ATR, Specac) by using a concave mirror (r=1.0 m). The prism is a right angle prism with the reflection surface size of 2 mm×2 mm. The spot size on the back surface of the prism is estimated as ∼ 0.5 mm. After the total reflection, the IR pulse was combined with the chirped-pulse through a mirror with a hole.

The combined beam was tightly focused into krypton gas with a parabolic mirror (f =50 mm). The generated four-wave mixing signal was introduced into a spectrometer which consisted of an imaging spectrograph and an EMCCD camera (SP-2358 and ProEM+1600, Princeton Instruments). The whole path for the IR pulse was purged with nitrogen gas.

One of the largest advantages of ATR spectroscopy over transmission IR spectroscopy in terms of the combination with chirped-pulse upconversion is that neither delay shift nor spatial displacement between the IR and chirped pulses exists by exchanging the samples. With transmission IR spectroscopy, it is not possible to avoid the delay shift between the IR and chirped pulses when the sample is exchanged. It causes systematic error of the frequency values. In addition, some chirp might be added to the IR pulse due to the dispersion of the sample at transmission IR spectroscopy.

Here we discuss the effect of the absorption and dispersion of the diamond prism. In principle, as long as the dispersion of the crystal is small enough that the duration of the IR pulse is kept much shorter than the chirped-pulse, a simple algorithm can be used to retrieve IR spectra from the upconverted spectra [15]. We measured the duration of the IR pulse after the total reflection by using four-wave mixing XFROG [13, 14, 16]. The retrieved time and frequency domain pictures are shown in Fig. 2. The duration of the pulse was 25.7 fs, which was more than three times longer than that of the incident pulse, 7 fs. However, it is much shorter than that of the chirped-pulse (4.9 ps). The GDD of the diamond crystal was estimated as +51 fs² at 3000 cm⁻¹ for 1.4 mm of the propagation length.

 

Fig. 2 The intensity (shaded curve) and phase (open squares) of the IR pulse after its reflection by the ATR diamond prism in (a) time and (b) frequency domain.

Download Full Size | PDF

At first we measured an upconverted spectrum without any samples. The measured spectrum is shown in Fig. 3. Absorption of the diamond prism at ∼2000 cm⁻¹ is clearly visible. In spite of the absorption, diamond prisms are commonly used for ATR-FTIR to take advantage of the transparency in the low frequency region, <700 cm⁻¹. In particular, the transparency would be very important for our chirped-pulse upconversion scheme because one of the largest benefit of the chirped-pulse upconversion with a gas is the sensitivity in the low frequency region.

 

Fig. 3 Measured IR spectrum with chirped-pulse upconversion (red) and conventional FTIR spectrometer (blue). The light source of the FTIR spectrometer is a halogen lamp.

Download Full Size | PDF

Next, we put some samples on the ATR prism and measured the IR absorption spectra. Figure 4 shows absorption spectra of water (H₂O) and acetone (C₃H₆O). The frequency resolution and measurement time were 4.2 cm⁻¹ and 1 s, respectively. We have compared the results with a conventional FTIR spectrometer (FT/IR-6100, JASCO). The frequency resolution and measurement time were 1.3 cm⁻¹ and 10 s, respectively. The frequencies of the measured absorption lines with the two different methods are identical to each other, whereas the intensities of the absorption lines are slightly different. The reason would be the difference in the penetration depth of the evanescent wave between the two different ATR systems.

 

Fig. 4 Absorption spectra of water, acetone and mixture of water and acetone (ratio 1:1). measured with (a), (b) the chirped-pulse upconversion system and (c), (d) a conventional FTIR spectrometer.

Download Full Size | PDF

To demonstrate rapid acquisition of IR spectra with the chirped-pulse upconversion technique, we have implemented a rapid solution exchanging system to our current system. We performed a proof-of-principle experiment by continuously recording IR spectra while exchanging the liquid on the ATR prism from acetone to water and vice versa.

The rapid solution exchanging system is a useful tool to study some biological samples at advanced time-resolved ATR-FTIR spectroscopy [4]. It is possible to monitor the IR spectrum in real-time with initiating some chemical reactions by exchanging the solutions on the ATR crystal by using two pneumatic drive pump systems, which are generally used in a stopped flow system. The detail of the rapid solution-exchange system is described in [4]. It was newly designed for rapidly exchanging a solution on the ATR crystal within several milliseconds by alternately pushing two syringes, which were filled with different solutions, with two pneumatic drive systems (nitrogen gas at 0.18 MPa). A Teflon unit (4 mm in thickness) with a flow channel (1 mm in depth and 2 mm in width) was mounted onto the ATR crystal (2 mm×2 mm) and tightened by the stainless parts holding the flow channels to supply and eject the buffers. The Teflon-flow channel covered the entire area of the ATR crystal. The conceptual illustration of the buffer-exchanging system is shown in Figs. 1(b) and 1(c).

The measurement was performed as follows. We used the laser pulse train signal (946 Hz) as a master clock. Each single-shot measurement of an upconverted spectrum was triggered with the master clock. After 100 spectra were measured, a trigger signal was sent to the rapid solution exchange system controller, and one of the two liquids was supplied to the sample chamber. The delay time of the solution-exchange from the trigger signal was measured as 79 ms, corresponding to 75 shots of the measurement. The spectrum was continuously recorded, and the total number of spectra was 500. Next, another 100 spectra were measured, and another trigger signal was sent to the controller. The alternate liquid was supplied to the chamber, and did the same measurement. We repeated the measurements 30 times and averaged the results.

Here we explain the result of the sample exchange from acetone to water. Three spectra recorded at the timings of 169, 182, 422 ms are shown in Figs. 5(a) and 5(b). At 169 and 422 ms only water and acetone, respectively, should be on the ATR prism. 182 ms is the timing at which the exchange occurs. We also show the absorption change in the time domain in Fig. 5(c). It is clear that the exchange is finished within ∼10 ms. The full width at half maximum of the instrumental function is estimated as 5.6 ms.

 

Fig. 5 Measured dynamics of exchange of liquids with the rapid solution-exchange ATRCPU system. (a),(b)Absorption spectra at each timing. (c) Absorption change at each frequency.

Download Full Size | PDF

At the timing of the solution exchange, we have observed that the absorption spectra slightly shifted. We believe that it comes from the mixture of the two solutions. The absorption spectra of the mixture of water and acetone with the ratio of 1:1 are shown as blue curves in Figs. 4 (c) and 4(d). The shifts of the infrared absorption bands are similar to those at the sample exchange experiment. For example, the C=O stretching mode (1712 cm⁻¹) is redshifted due to hydrogen bonding between water and acetone molecules. The C-C stretching and CH₃ deformation modes (1223 cm⁻¹ and 1362 cm⁻¹, respectively) are blueshifted due to strengthening of the C-C bond as a consequence of the weakening of the C=O bond. The report of these phenomena and the detailed analyses has already been published in [17, 18].

The fact which we observed the shift of the absorption bands during the solution exchange means that the significant amount of mixture of water and acetone exists at the boundary of the two solutions. More concretely, at least 0.5 mm of length of the flow channel (the spot size of the IR beam at the prism) at the boundary of the two solutions was filled with the mixture. This information would be important for buffer-exchange experiments for biological tissues, and would initiate more advanced design of the solution-exchange system.

In summary, ATR spectroscopy with chirped-pulse upconversion has been demonstrated. Absorption lines of some liquid samples were clearly observed and the frequencies of the lines were accurate. We have implemented a rapid solution-exchange system and observed the dynamic change of the infrared absorption bands of the solution with millisecond time resolution. We believe that this unique system is useful for advanced studies of biological tissues.