5.2 Microscopy and Spectroscopy
143
5.2.4 Time-Resolved Measurements
The dynamics in electronic systems can be studied by time-resolved spectroscopy.
For this purpose, different techniques have been established over the last decades,
whereas the main difference originates from the time scales accessible and the equipment used to resolve fast to ultrafast processes. With time-resolved (TR) measurements (cf. Fig. 5.8), one can determine carrier lifetimes, as well as characterise the
recombination mechanisms in the investigated structure.
One common technique to slice the time in order to have temporal resolution in
signal acquisition is given by the employment of gated detectors such as intensified
CCDs. A trigger pulse coinciding with the excitation pulse is used to define an electronic delay for the detection time window. Thereby, an electronic shutter determines
from which time frame the detector obtains signal. Then, the time trace is established
by scanning through the delay time electronically. Typically, this technique is useful
for processes which are longer than nanoseconds, such as excitonic decay in colloidal quantum dots [15]. The repetition rate of the laser must be correspondingly
low in order to allow decay processes lasting up to the micro- or milliseconds to be
investigated.
A much better temporal resolution down to below 1 ps for optimized devices is
provided by the streak-camera technique, which works the same way as a stroboscope. Spectrally-resolved incident photons impinge on a photocathode from which
they eject electrons. These electrons accelerated in a field experience a controlled
deflection periodic in time based on a sweep circuit. A trigger signal fed into the
circuit sets the rising flank of a sinusoidal AC bias with its most linear part around
the zero delay time with regard to the optical excitation pulse. Based on their time
of ejection, the deflected photoelectrons end up on different parts of a phosphorous
screen behind a multi-channel plate, resulting in temporal resolution of spectroscopic
data. The signal on the phosphorous screen is then imaged by a CCD-type camera.
From the acquired 2D image (an example is given in Fig. 5.9), transients (see right
panel) and spectra (see upper panel) can be retrieved by integration over the desired
energy or time range, respectively. In case of a mono-exponential decay,
3 a typical
fit of the form
I (t) = I 0 exp
−
t − t 0
τ
(5.1)
3 Note that, while an exponential decay is typically the case for excitons (after being initially prepared
by an ultrashort excitation to a starting population N 0 ), for which a simple rate equation applies,
the transient of uncorrelated electron–hole pairs in a plasma follows a different trend. In some
cases, where density-dependent Auger-like processes such as exciton–exciton annihilation occur,
the mono-exponential fit also cannot hold. Some systems may feature short and long decay time
constants, which might require a biexponential fitting. In the case of a reservoir feeding a lower-lying
state, such as for cavity–polaritons, more elaborate rate equations which take into account various
scattering and decay channels may reproduce the dynamics of a system appropriately (see e.g.,
[16]). Similarly, the transfer between free and bound excitons could require a specific rate equation
(described e.g., for GaSe and GaTe in [17]). For further insights on charge-carrier dynamics in
semiconductors, the interested reader is referred to semiconductor theory textbooks such as [18].
143
5.2.4 Time-Resolved Measurements
The dynamics in electronic systems can be studied by time-resolved spectroscopy.
For this purpose, different techniques have been established over the last decades,
whereas the main difference originates from the time scales accessible and the equipment used to resolve fast to ultrafast processes. With time-resolved (TR) measurements (cf. Fig. 5.8), one can determine carrier lifetimes, as well as characterise the
recombination mechanisms in the investigated structure.
One common technique to slice the time in order to have temporal resolution in
signal acquisition is given by the employment of gated detectors such as intensified
CCDs. A trigger pulse coinciding with the excitation pulse is used to define an electronic delay for the detection time window. Thereby, an electronic shutter determines
from which time frame the detector obtains signal. Then, the time trace is established
by scanning through the delay time electronically. Typically, this technique is useful
for processes which are longer than nanoseconds, such as excitonic decay in colloidal quantum dots [15]. The repetition rate of the laser must be correspondingly
low in order to allow decay processes lasting up to the micro- or milliseconds to be
investigated.
A much better temporal resolution down to below 1 ps for optimized devices is
provided by the streak-camera technique, which works the same way as a stroboscope. Spectrally-resolved incident photons impinge on a photocathode from which
they eject electrons. These electrons accelerated in a field experience a controlled
deflection periodic in time based on a sweep circuit. A trigger signal fed into the
circuit sets the rising flank of a sinusoidal AC bias with its most linear part around
the zero delay time with regard to the optical excitation pulse. Based on their time
of ejection, the deflected photoelectrons end up on different parts of a phosphorous
screen behind a multi-channel plate, resulting in temporal resolution of spectroscopic
data. The signal on the phosphorous screen is then imaged by a CCD-type camera.
From the acquired 2D image (an example is given in Fig. 5.9), transients (see right
panel) and spectra (see upper panel) can be retrieved by integration over the desired
energy or time range, respectively. In case of a mono-exponential decay,
3 a typical
fit of the form
I (t) = I 0 exp
−
t − t 0
τ
(5.1)
3 Note that, while an exponential decay is typically the case for excitons (after being initially prepared
by an ultrashort excitation to a starting population N 0 ), for which a simple rate equation applies,
the transient of uncorrelated electron–hole pairs in a plasma follows a different trend. In some
cases, where density-dependent Auger-like processes such as exciton–exciton annihilation occur,
the mono-exponential fit also cannot hold. Some systems may feature short and long decay time
constants, which might require a biexponential fitting. In the case of a reservoir feeding a lower-lying
state, such as for cavity–polaritons, more elaborate rate equations which take into account various
scattering and decay channels may reproduce the dynamics of a system appropriately (see e.g.,
[16]). Similarly, the transfer between free and bound excitons could require a specific rate equation
(described e.g., for GaSe and GaTe in [17]). For further insights on charge-carrier dynamics in
semiconductors, the interested reader is referred to semiconductor theory textbooks such as [18].