150
C. Li and J.-X. Cheng
The first problem was addressed by recent advances in mid-IR laser sources in
the last two decades. In particular, modern quantum cascade laser (QCL) offers
the ability to perform broadband wavelength scan with ultra-high spectral resolution
(~0.1 cm
−1 ) discretely [111–113], in which specific vibrational bands of interests are
pinpointed at high speed while providing enough chemical information to determine
the composition of samples through spectra analysis [30]. Therefore, the majority of
mid-IR PT microscopes use QCL as IR pump beam [59, 84, 110]. Besides QCL, difference frequency generation (DFG) is another effective approach to produce tunable
mid-IR sources as the pump beam of PT microscope [114–116]. The fundamental
beams usually consist of a tunable beam and a monochromatic beam. Both beams
are focused onto nonlinear crystals, such as periodically poled LiNbO 3 , to produce
tunable nanosecond mid-IR pulses whose average power can reach up to 200 mW
with less than 10 cm
−1 linewidth, which is sufficient to induce strong PT effect in
most absorption bands and measure liquid phase IR spectra. Despite the additional
complexity of DFG, the produced mid-IR pump beam can be tuned to various wavelengths with better coverage of high wavenumber mid-IR regions, including the “cell
silent region”, where most intrinsic biomolecules show no absorption, and deuterated
molecules are often used to investigate cell metabolisms [116, 117]. Therefore, QCL
and DFG collectively provide more options to the pump beam selection of mid-IR
excited PTM and enable more important applications in biological studies.
The second problem found a solution by the deployment of reflective optics,
including the Cassegrain objectives and off-axis parabolic mirrors, since these optics
are immune to chromatic aberration across the whole spectral window. Like visible
excited PTM, a typical mid-IR excited PT microscope deploys coaxial configuration,
as illustrated in Fig. 6.8a, where the mid-IR pump beam and visible probe beam are
combined at the dichroic mirror and collinearly sent to the Cassegrain objective [59].
To avoid mid-IR beam power losses, the reflective optics are usually coated with gold
or silver, and the samples are sandwiched by ultra-broadband IR transparent materials
such as CaF 2 or MgF 2 . Since the water vapor and some other trace organic vapors in
air all absorb mid-IR, the beam path of the mid-IR pump beam is usually designed
with minimum distance, to further maintain the pump beam power at the sample.
In some cases, to diminish the power loss and push the detection limit, methods
used in conventional mid-IR measurements, including purging dry nitrogen into
the system to reduce water vapor and CO 2 absorption, are needed [118]. Note that
the reflected mid-IR residue (power <1 mW) from the dichroic mirror is collected
by a mercury–cadmium–telluride (MCT) detector to record the IR power at each
wavelength in real time. This power spectrum is used to normalize the PT signal
acquired by the photodiode to produce the IR absorption spectra of samples.
There have been several reports aiming to improve the SNR in mid-IR excited
PTM through various optimization approaches. For instance, Zhang et al. investigated the frequency dependence of the mid-IR PT signal, laser noise, and the system
SNR (Fig. 6.8b) [59]. By considering the trade-off between the SNR and data acquisition speed, the pump beam modulation frequency was set at 100 kHz for optimal
performance. As shown in Fig. 6.8c, a high-Q tunable resonant amplifier with a center
frequency of 102.5 kHz was installed after the photodiode such that the modulated
C. Li and J.-X. Cheng
The first problem was addressed by recent advances in mid-IR laser sources in
the last two decades. In particular, modern quantum cascade laser (QCL) offers
the ability to perform broadband wavelength scan with ultra-high spectral resolution
(~0.1 cm
−1 ) discretely [111–113], in which specific vibrational bands of interests are
pinpointed at high speed while providing enough chemical information to determine
the composition of samples through spectra analysis [30]. Therefore, the majority of
mid-IR PT microscopes use QCL as IR pump beam [59, 84, 110]. Besides QCL, difference frequency generation (DFG) is another effective approach to produce tunable
mid-IR sources as the pump beam of PT microscope [114–116]. The fundamental
beams usually consist of a tunable beam and a monochromatic beam. Both beams
are focused onto nonlinear crystals, such as periodically poled LiNbO 3 , to produce
tunable nanosecond mid-IR pulses whose average power can reach up to 200 mW
with less than 10 cm
−1 linewidth, which is sufficient to induce strong PT effect in
most absorption bands and measure liquid phase IR spectra. Despite the additional
complexity of DFG, the produced mid-IR pump beam can be tuned to various wavelengths with better coverage of high wavenumber mid-IR regions, including the “cell
silent region”, where most intrinsic biomolecules show no absorption, and deuterated
molecules are often used to investigate cell metabolisms [116, 117]. Therefore, QCL
and DFG collectively provide more options to the pump beam selection of mid-IR
excited PTM and enable more important applications in biological studies.
The second problem found a solution by the deployment of reflective optics,
including the Cassegrain objectives and off-axis parabolic mirrors, since these optics
are immune to chromatic aberration across the whole spectral window. Like visible
excited PTM, a typical mid-IR excited PT microscope deploys coaxial configuration,
as illustrated in Fig. 6.8a, where the mid-IR pump beam and visible probe beam are
combined at the dichroic mirror and collinearly sent to the Cassegrain objective [59].
To avoid mid-IR beam power losses, the reflective optics are usually coated with gold
or silver, and the samples are sandwiched by ultra-broadband IR transparent materials
such as CaF 2 or MgF 2 . Since the water vapor and some other trace organic vapors in
air all absorb mid-IR, the beam path of the mid-IR pump beam is usually designed
with minimum distance, to further maintain the pump beam power at the sample.
In some cases, to diminish the power loss and push the detection limit, methods
used in conventional mid-IR measurements, including purging dry nitrogen into
the system to reduce water vapor and CO 2 absorption, are needed [118]. Note that
the reflected mid-IR residue (power <1 mW) from the dichroic mirror is collected
by a mercury–cadmium–telluride (MCT) detector to record the IR power at each
wavelength in real time. This power spectrum is used to normalize the PT signal
acquired by the photodiode to produce the IR absorption spectra of samples.
There have been several reports aiming to improve the SNR in mid-IR excited
PTM through various optimization approaches. For instance, Zhang et al. investigated the frequency dependence of the mid-IR PT signal, laser noise, and the system
SNR (Fig. 6.8b) [59]. By considering the trade-off between the SNR and data acquisition speed, the pump beam modulation frequency was set at 100 kHz for optimal
performance. As shown in Fig. 6.8c, a high-Q tunable resonant amplifier with a center
frequency of 102.5 kHz was installed after the photodiode such that the modulated