5.2 Microscopy and Spectroscopy
149
excitons in the valley landscape of monolayer TMDCs (cf. [46] and references
therein), promising a better understanding of valley depolarisation in TMDCs through
valley dynamics studies as a function of the quasi-particles’ centre-of-mass (in-plane)
momentum.
Using another time-resolved investigation method without spectral resolution,
which utilized four-wave mixing (FWM), Rabi-flopping signatures of quantum-well
excitons under pulsed THz illumination in the time domain were demonstrated to
become irreversible for strong “ionising” THz fields. Whereas, weak THz radiation
quasi-resonantly drives intraexcitonic transitions between the 1s and higher-order
states [47]. Here, typically large laser spots on the sample from an amplifier laser
are used owing to the THz exposed areas being well above 300 µm in diameter. To
employ similar techniques in combination with TMDCs successfully, the ability to
focus strongly onto micro-sized mono-/few-layer flakes or the availability of largearea mono-/few-layers is needed, e.g., for two-pulse or three-pulse (laser-induced
dynamic gratings) FWM experiments. In such experiments, the laser pulses induce
polarisations in the crystal which interfere and form lattices (gratings). The refracted
signal for different time delays between the pulses is acquired for a specific diffraction order with a photodiode, from which a FWM transient can be obtained that is
analysed with respect to dephasing processes. In addition, to address intraexcitonic
transitions, the high binding energies of TMDC excitons shifts the manipulationpulse’s wavelength from the THz to the IR (cf. optical-pump–IR-probe in [41]).
First FWM attempts by the author’s team on CVD samples showed the challenges
in recording monolayer signal. The interested reader is referred to a shortly later
published comprehensive article on FWM microscopy with monolayer MoSe 2 in the
literature [48], nicely showing how FWM can give access both to dynamics (radiative lifetime T 1 ) and the coherence (dephasing times T 2 , homogeneous linewidth
γ = 2/T 2 ) measurements, with radiatively-limited dephasing T 2 = 2T 1 demonstrated at a temperature of 6–10 K (γ <10 K ≈ 1 meV).
A different example of time-resolved experiments is given by the measurement of a
laser’s output characteristics, which can be analysed using fast photodetectors. Intensity as a function of time is recorded and the time trace Fourier-transformed to obtain
a noise spectrum. The frequency bandwidth of the detector limits the maximummeasurable frequency components present in relative intensity noise (RIN). Such a
RIN characterisation of dual-wavelength laser emission from a VECSEL revealed
power-dependend antiphase-noise dynamics between the two wavelengths simultaneously oscillating in the resonator [49].
Another interesting time-resolving experiment reveals the photon statistics of an
emitter and is referred to as Hanbury–Brown-and-Twiss (HBT) experiment [51] (see
Fig. 5.10). With this method, photon bunching, antibunching and random emission
(coherent state) can be distinguished via temporal photon statistics, by obtaining
the second-order temporal autocorrelation function g
(2)
(τ = 0) =
: ˆ
n
2
:
/
ˆ
n
2 at
zero time delay (τ = 0) where the correlation between photons is maximum. For
increased time intervals between photons, their correlation decays. ˆ
n is the photon
counting operator. Colons indicate normal ordering of the photon field operators.
149
excitons in the valley landscape of monolayer TMDCs (cf. [46] and references
therein), promising a better understanding of valley depolarisation in TMDCs through
valley dynamics studies as a function of the quasi-particles’ centre-of-mass (in-plane)
momentum.
Using another time-resolved investigation method without spectral resolution,
which utilized four-wave mixing (FWM), Rabi-flopping signatures of quantum-well
excitons under pulsed THz illumination in the time domain were demonstrated to
become irreversible for strong “ionising” THz fields. Whereas, weak THz radiation
quasi-resonantly drives intraexcitonic transitions between the 1s and higher-order
states [47]. Here, typically large laser spots on the sample from an amplifier laser
are used owing to the THz exposed areas being well above 300 µm in diameter. To
employ similar techniques in combination with TMDCs successfully, the ability to
focus strongly onto micro-sized mono-/few-layer flakes or the availability of largearea mono-/few-layers is needed, e.g., for two-pulse or three-pulse (laser-induced
dynamic gratings) FWM experiments. In such experiments, the laser pulses induce
polarisations in the crystal which interfere and form lattices (gratings). The refracted
signal for different time delays between the pulses is acquired for a specific diffraction order with a photodiode, from which a FWM transient can be obtained that is
analysed with respect to dephasing processes. In addition, to address intraexcitonic
transitions, the high binding energies of TMDC excitons shifts the manipulationpulse’s wavelength from the THz to the IR (cf. optical-pump–IR-probe in [41]).
First FWM attempts by the author’s team on CVD samples showed the challenges
in recording monolayer signal. The interested reader is referred to a shortly later
published comprehensive article on FWM microscopy with monolayer MoSe 2 in the
literature [48], nicely showing how FWM can give access both to dynamics (radiative lifetime T 1 ) and the coherence (dephasing times T 2 , homogeneous linewidth
γ = 2/T 2 ) measurements, with radiatively-limited dephasing T 2 = 2T 1 demonstrated at a temperature of 6–10 K (γ <10 K ≈ 1 meV).
A different example of time-resolved experiments is given by the measurement of a
laser’s output characteristics, which can be analysed using fast photodetectors. Intensity as a function of time is recorded and the time trace Fourier-transformed to obtain
a noise spectrum. The frequency bandwidth of the detector limits the maximummeasurable frequency components present in relative intensity noise (RIN). Such a
RIN characterisation of dual-wavelength laser emission from a VECSEL revealed
power-dependend antiphase-noise dynamics between the two wavelengths simultaneously oscillating in the resonator [49].
Another interesting time-resolving experiment reveals the photon statistics of an
emitter and is referred to as Hanbury–Brown-and-Twiss (HBT) experiment [51] (see
Fig. 5.10). With this method, photon bunching, antibunching and random emission
(coherent state) can be distinguished via temporal photon statistics, by obtaining
the second-order temporal autocorrelation function g
(2)
(τ = 0) =
: ˆ
n
2
:
/
ˆ
n
2 at
zero time delay (τ = 0) where the correlation between photons is maximum. For
increased time intervals between photons, their correlation decays. ˆ
n is the photon
counting operator. Colons indicate normal ordering of the photon field operators.