Full Papers
doi.org/10.1002/cplu.202000767
ChemPlusChem
the sub-ps transient absorption setup 10 mL of the sample were
cycled through an in-house build cell with a pathlength of ca.
100 μm for pump and probe beams. In the sub-ns transient
absorption setup 10 mL of the sample were cycled through a flow
cell (Starna) with a pathlength of 2 mm for pump and probe beams.
In all cases all scans resulted in reproducible data sets. Additionally,
the integrity of the sample was checked by recording stationary
absorption spectra before and after each measurement. No photo-
catalyst degradation was observed under the used conditions. The
shown data correspond to one representative measurement. No
smoothing or filtering procedures were applied to the data.
Time-resolved UV/Vis absorption spectroscopy
1
2
3
4
5
6
7
8
9
The ns to ms transient absorption spectroscopy was recorded by
a streak camera setup as described previously.[16h,21a,30–31] In brief, the
third harmonic of a Nd:YAG laser (10 Hz, Surelite II, Continuum)
pumping an Optical Parametric Oscillator (OPO, Continuum) tuned
to 355/460 nm (10 mJ, ca. 10 ns) was used for sample excitation. As
a probe light a pulsed 150 W Xe-flash lamp (Applied Photophysics)
was used which was focused three times via toric mirror optics: i)
before probe shutter, ii) into sample, iii) into spectrograph. The
entire white light probe pulse was analyzed by a combination of a
spectrograph (200is, Bruker) and a streak camera (C7700, Hamamat-
su Photonics). The use of mechanical shutters enabled the record-
ing of a sequence of three individual data sets: i) an image (DFL)
with both flash lamp and laser, ii) an image (D0) without any
incoming light, and iii) an image (DF) only with the flash lamp. 100
of such sequences were recorded and corresponding data sets
were averaged. Then, the TA was calculated as [Eq. (1)]:
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Transient absorption data analysis and modelling
SVD-based rank analysis and global fitting were performed using an
in-house written program described previously.[21b,30–31] In brief, the
linear least squares problem in Equation (2)
�
�
2
c2 ¼kDAÀFBk¼Min
DF À D0
DFL À D0
(2)
DOD ¼log
(1)
~
is solved, where A is the time-resolved absorption data matrix, F
is the matrix containing the analytical functions accounting for the
temporal changes in the data, i.e. exponential decays (convoluted
with the instrument response, typically a Gaussian function), and B
is the matrix with the to be determined spectra. Further
optimization of c2 is achieved by optimizing the rate constants in F
by a nonlinear least squares algorithm. As a result of such fits so-
called decay associated difference spectra (DADS in matrix B) and
their associated optimized rate constants are obtained. These are
the unique result of the global fit and this treatment does not
require any model for the kinetics involved in the transient
processes. The number of exponentials in the global fit is
determined by the SVD-based rank analysis, which is described
elsewhere.[33] The model that relates the actual species kinetics to
the elementary function is applied afterwards resulting in species
associated spectra (SAS). The shape of the SAS in terms of identity
with well-known spectra or following physical laws decides about
the appropriateness of the model. This step does not change the c2
value found in the global fit and, therefore, this procedure has the
advantage that all interpretation is performed with the same quality
of fit.
A 10 mL sample was stepwise cycled by a peristaltic pump (ecoline,
ISMATEC) through a flow cell with a pathlength of 2 mm for pump
and 10 mm for probe beams (dimensions: 2 mm×10 mm×30 mm,
Starna) ensuring a total replacement of the sample prior to each
individual measurement. No photocatalyst degradation was ob-
served under the used conditions.
The Sub-ps Pump/Supercontinuum-Probe Spectroscopy were
carried out using UV/Vis pump-supercontinuum probe spectrom-
eter at 1 kHz repetition frequency as described in ref.[16h] In brief, a
Ti-sapphire amplifier system (Coherent Libra) was used to generate
800 nm with 1.2 mJ pulses at 1 kHz. The output was split into three
parts of which only two were used: 1) Ca. 50% of the 800 nm
pulses were used to pump a colinear Optical Parametric Amplifier
(OPA, TOPAS-800-fs, Light Conversion) tuned to pump pulses
centered at ca. 450 nm (ca. 100 fs, ca. 400 nJ at the sample position)
for sample excitation. 2) Ca. 10% were used to pump a non-
colinear Optical Parametric Amplifier (NOPA, In-house build) tuned
to pulses centered at ca. 530 nm (ca. 100 fs, ca. 5 μJ at the CaF2
position) for generation of supercontinuum white light probe
pulses by focusing into a moving CaF2 disc of 1 mm thickness
giving a probe spectrum ranging from 310 to 700 nm. Pump pulses
were delayed via a motorized delay line equipped with an open
corner cube mirror up to 2 ns. Two complementary high-speed
spectrographs (Entwicklungsbüro EB Stresing) for signal and
reference recording were used. The pump and probe pulses were
focused colinearly into the sample to spot sizes of ca. 80 μm and
60 μm full width at half maximum (FWHM), respectively. For longer
As an alternative analysis, known species spectra, taken either from
literature or recorded in this work, were taken in order to
decompose the recorded time-resolved data matrix using the
transpose of the data matrix in [Eq. (2)] and using the basis spectra
instead of analytical functions. The resulting concentration-time-
profiles inform about the appropriateness of the basis spectra and
the physical reasonability, i.e. total sum of species being constant
to 1.
delays reaching out from ns to μs time ranges
a similar
spectrometer was used in which the pump laser was electronically
delayed relative to the probe laser. A detailed description can be
found in ref.[32] The relative polarizations between the pump and
probe were set by a half-wave plate in the pump-beam path to
Quantum chemical calculations
Quantum-chemical calculations on the excited singlet and triplet
°
magic angle (54.71 ) for observations of pure population changes.
states of 2b as well as on the radical species of 2b that potentially
*
contribute to the time-resolved absorption signals, i.e. 2b and
*
The averaged pre-t0 laser scatter signal was subtracted from the
data and the ca. 1 ps chirp of the white light was corrected for prior
to data analysis using the coherent artefact as an indicator for time
zero at each wavelength. Throughout the probe range, the spectral
resolution was better than 4 nm and the temporal resolution was
better than 150 fs. 10 individual scans with averaging 100 spectra
per time point were typically recorded. The time axis – within total
500 points – was linear between À1 and 2 ps and logarithmic from
2 ps to the maximum time delay ensuring that the dynamics on
every timescale will have equal weighting in the fitting analysis. In
2b +, were performed using the Firefly QC package,[34] which is
partially based on the GAMESS (US)[35] source code. All ground state
structures were optimized on the level of restricted open shell
density functional theory (ROHF-DFT) using the B3LYP functional
and the split-valence 6–31G(p,d)+ + basis set. In terms of the radical
spectrum calculation, complete active space self-consistency field
(CASSCF) theory was used in order to calculate the static correlation
energy. The most intense electronic transitions of 2b radicals, i.e.
DX D0 are of π–π* and n–π* type. Thus, 13 contributing π
ChemPlusChem 2021, 86, 373–386
384
© 2020 The Authors. ChemPlusChem published by Wiley-VCH GmbH