Transient absorption spectroscopy of the electron transfer step in the photochemically activated polymerizations of N-ethylcarbazole and 9-phenylcarbazole

The polymerization of photoexcited N-ethylcarbazole (N-EC) in the presence of an electron acceptor begins with an electron transfer (ET) step to generate a radical cation of N-EC (N-EC˙+). Here, the production of N-EC˙+ is studied on picosecond to nanosecond timescales after N-EC photoexcitation at a wavelength λex = 345 nm using transient electronic and vibrational absorption spectroscopy. The kinetics and mechanisms of ET to diphenyliodonium hexafluorophosphate (Ph2I+PF6−) or para-alkylated variants are examined in dichloromethane (DCM) and acetonitrile (ACN) solutions. The generation of N-EC˙+ is well described by a diffusional kinetic model based on Smoluchowski theory: with Ph2I+PF6−, the derived bimolecular rate coefficient for ET is kET = (1.8 ± 0.5) × 1010 M−1 s−1 in DCM, which is consistent with diffusion-limited kinetics. This ET occurs from the first excited singlet (S1) state of N-EC, in competition with intersystem crossing to populate the triplet (T1) state, from which ET may also arise. A faster component of the ET reaction suggests pre-formation of a ground-state complex between N-EC and the electron acceptor. In ACN, the contribution from pre-reaction complexes is smaller, and the derived ET rate coefficient is kET = (1.0 ± 0.3) × 1010 M−1 s−1. Corresponding measurements for solutions of photoexcited 9-phenylcarbazole (9-PC) and Ph2I+PF6− give kET = (5 ± 1) × 109 M−1 s−1 in DCM. Structural modifications of the electron acceptor to increase its steric bulk reduce the magnitude of kET: methyl and t-butyl additions to the para positions of the phenyl rings (para Me2Ph2I+PF6− and t-butyl-Ph2I+PF6−) respectively give kET = (1.2 ± 0.3) × 1010 M−1 s−1 and kET = (5.4 ± 1.5) × 109 M−1 s−1 for reaction with photoexcited N-EC in DCM. These reductions in kET are attributed to slower rates of diffusion or to steric constraints in the ET reaction.


S2. Steady State UV-Vis Spectra
On addition of N-EC and Ph 2 I + PF 6to DCM and with UV irradiation, the solution turns from pale yellow to dark blue. This colour only forms when DCM is used as the solvent of choice. When ACN is used, S4 the colour change is not observed. This colour change indicates a reaction between DCM and photoexcited reagents. Sari et al. reported that the blue colour is removed when PF 6 -ions are replaced by hydroxide anions. 1 However, the origins of this colour change remain uncertain.

S5. Analysis of Transient Absorption Spectra
Initial processing of both transient electronic absorption spectroscopy (TEAS) and transient vibrational absorption spectroscopy (TVAS) data used the KOALA program. 2 Flat-shift baseline corrections were applied to TEAS data, and quadratic baseline corrections to TVAS data. The TEAS spectra obtained at short time delays were further corrected to remove the effects of chirp in the white-light continuum probe pulse. A pre time-zero (negative time) spectrum was subtracted from all transient spectra.
TVAS data were analysed by fitting Gaussian peaks of adjustable centre and width to transient absorption and GS bleach features, and the resulting data sets of time-dependent integrated band intensities were exported as kinetic traces, as seen in figure S14. S10 Figure S14. Example decomposition of TVAS spectra obtained at time delays of 0.7 -1189 ps for a solution of N-EC (7 mM) and Ph 2 I + PF 6 -(84 mM) in DCM. The decomposition was carried out using Koala software. 2 The various panels show the experimental spectrum (black), and the overall fit (red).
Broad and potentially overlapping bands in TEAS data were analysed using a basis spectrum decomposition method. Each system was analysed in a different way, as explained in the following sub-sections. S11

S5.1. N-EC / 9-PC in DCM or ACN
The analysis of all TEAS spectra relied on deriving reliable basis spectra representative of ESA from the S 1 and the T 1 states of the carbazole. These were obtained using the procedure summarized below, for each carbazole and for the two solvents (DCM and ACN) by analysis of TEAS data obtained in the absence of an electron acceptor.
The analysis windows for all spectra decompositions in the absence of EA were set to cover wavelengths typically between 380 and 660 nm. This meant that the applied basis spectra were only fitted to data within the selected analysis window. Hence spectral data outside this window did not contribute to the kinetic analysis. For the spectral decomposition, a first basis spectrum was chosen to be an experimentally measured early time TEAS spectrum, generally corresponding to a time delay of 0.3 ps. This spectrum was selected to define the carbazole S 1 -state ESA signature in the absence of other species. The spectrum acquired at the largest experimental time delay was then used to prepare a basis spectrum corresponding to T 1 absorption. This late-time spectrum contained ESA contributions from both the S 1 and T 1 states because ISC is incomplete on timescales of up to 1.3 ns available in our measurements. The S 1 contribution was therefore removed using the following procedure to generate a T 1 ESA basis spectrum. First, an integration window was placed over the wavelength region from

S5.2. N-EC and Ph 2 I + PF 6in DCM
The analysis of all TEAS spectra collected upon the addition of an electron acceptor relied on deriving reliable basis spectra representative of ESA from the S 1 and T 1 states of the carbazole as well as the ground-state absorption of the radical cation. These basis spectra were obtained using the procedure summarized above for the S 1 ESA bands, and below for the radical cation absorption. T 1 ESA bands were represented by the basis spectra prepared using the procedure described in S5.1.
The basis spectrum for N-EC +. GS absorption in DCM was obtained using the TEAS spectrum acquired at the largest time delay, and with the highest concentration (84 mM) of EA on the assumptions that S13 under these conditions complete ET from S 1 had occurred, leaving no residual S 1 ESA, and that the electron transfer reaction outcompeted ISC so the late-time spectrum contained no T 1 ESA component. These basis functions were used in the decomposition of TEAS spectra obtained at lower EA concentrations, with the analysis window set to 380 -500 nm. The triplet basis spectrum was applied to low concentration of EA data and omitted from analysis of spectra obtained with 84 or 42 mM of EA because ET dominates the ISC.

S5.3. N-EC and Ph 2 I + PF 6in ACN
The same method to obtain spectral basis functions for decomposition of TEAS data for experiments in ACN was used as described in S5.2. In the preparation of the basis functions, the analysis windows were set to cover wavelengths between 380 to 660 nm. An example of the decomposition of TEAS data for N-EC and Ph 2 I + PF 6in ACN is shown in figure S16.

S5.4. N-EC and Me 2 Ph 2 I + PF 6in DCM
TEAS data sets for N-EC and Me 2 Ph 2 I + PF 6solutions in DCM were analysed differently to the procedures described in S5.2 and S5.3 because the highest EA concentration data set available was 28 mM. As a result, ET could not be assumed to be complete at the longest measured time delays.
The following procedure was therefore adopted. The analysis windows were set to cover wavelengths between 410-500 and 380-500 nm. A basis spectrum was made from an early time spectrum in each data set and was used to describe the N-EC S 1 ESA bands. The basis spectrum used to represent N-EC +. was the same one as made in S5.2. However, the TEAS data had its wavelength scale recalibrated to match the absorption of the N-EC +. feature in S5.2. The absorption of the N-EC +. feature should be identical in all spectra for N-EC in DCM however, as the spectra were taken on different days the calibration of the wavelength axis may differ slightly. The EA does not absorb at 345 nm and therefore does not contribute any signatures to the TEAS spectra. A T 1 ESA basis spectrum was obtained as described in Section S5.1 and included in the analysis of to all data sets.

S5.5. N-EC and t-butyl-Ph 2 I + PF 6in DCM
The limited solubility of t-butyl-Ph 2 I + PF 6in DCM meant the same method as described in S5.4 was applied to analysis of TEAS data for this system, with the analysis window set to cover wavelengths from 390-500 nm.

S5.6. 9-PC and Ph 2 I + PF 6in DCM
Data for 9-PC and Ph 2 I + PF 6in DCM were analysed using a modified procedure because the ET reaction was incomplete even for 84 mM EA solutions at the largest time delay, leaving a significant ESA feature S14 from the S 1 state of the carbazole. Using an analysis window spanning wavelengths from 400-610 nm, the following method was adopted.
A basis spectrum was made from an early time spectrum in each data set and was used to describe the 9-PC S 1 ESA bands. A basis spectrum was also generated to represent 9-PC +. formation with the same method used to generate the T 1 -state ESA basis spectrum in S5.1 but now in the presence of 84 mM of EA. This basis spectrum was then reused for all data sets with varying EA concentrations. A T 1 ESA basis spectrum was also applied to all data sets, except for those obtained with the addition of 84 mM EA.
Time-dependent band intensity data were fitted to a version of the Smoluchowski kinetic model     S20

S7.2. TVAS Fitting Results
For the TVAS experiment with 7 mM N-EC solutions in DCM in the absence of EA, the kinetics were modelled with the values obtained in TEAS data for the same system, as reported in table S2 and shown in figure S20. Extracted kinetic data for 7 mM N-EC with 84 mM Ph 2 I + PF 6in DCM were fit to equations (2) and (3), as seen in section 3.4 in the main paper, without the additional fast exponential term for equation 3. This difference was introduced because the N-EC +. absorption band was evident at the earliest time delays used to acquire TVAS data, and the kinetic fit was not improved by inclusion of a fast exponential growth term for product formation. Instead, vertical offsets were applied in fits to all kinetic traces except the S 1 population decay to account for this initial absorption by products at 0.7 ps. In the kinetic analysis of TVAS data, values of the C,  0 and  1 parameters were fixed to the values obtained from the corresponding TEAS data analysis, as shown in  The TVAS kinetics were modelled with values from TEAS data for the same system.