Visible light-induced polymerization initiated by borate salts of bicationic monochromophoric benzothiazolestyrylium dyes

The structures of novel bicationic monochromophoric benzothiazolestyrylium dyes are presented in Scheme 1. Details of the synthesis are given in the “Experimentalˮ section. An inspection of the experimental data shows that the melting points for all sensitizers are sharp, indicating high purity and crystalline phase of the resulting dyes. Figure 1 shows the ¹H NMR spectra of selected hemicyanine dye (II) in the region of δ = 1.0–9 ppm. In the region, the signals characteristic for the protons in benzene ring, pyridinyl ring, and the central double bond usually are present. All the peaks are assigned to the corresponding hydrogen atoms in molecule, as presented in “Experimentalˮ section. It is noteworthy that the ¹H NMR spectra display two characteristic doublets localized at chemical shifts in the range of 6.8–8.3 ppm. They are attributed to both vinyl hydrogen atoms.

The bands that appeared in the range from about 6.5 to 9.1 ppm are the signal for the aromatic protons. According to thin-layer chromatography analysis (silica gel 60, F-254), with the use of methanol-acetone mixture (2:1) as the eluent, the hemicyanine dyes are of high purity. Taking into account the ¹H NMR spectra, the elemental analysis, and thin-layer chromatography results, one can conclude that purity of the dyes was as it is required for spectroscopic studies (>99 %).

The tested compounds belong to the group of asymmetric dyes, so-called hemicyanine dyes. These dyes have a D-π-A⁺X⁻ structure and differ from each other the type of substituent in the vinyl group. An asymmetric structure of compounds has an influence on the nature of both the electron absorption spectra and fluorescence spectra. The examples of electron absorption spectra of the bicationic monochromophoric hemicyanine dyes studied are shown in Figs. 2 and 3. Table 1 collects the values of the absorption maximum positions, the fluorescence maximum positions, energy of 0 → 0 transition, Stokes shifts, reduction potentials, and oxidation potentials.

The bicationic monochromophoric hemicyanine dyes show two characteristic absorption bands. The absorption band localized in the longer wavelengths spectrum region is the charge-transfer band, attributed to the intermolecular charge transfer from free electron pair of nitrogen atom on the quaternary nitrogen atom in benzothiazole moiety. The absorption band in the range from 360 to 400 nm is attributed to the π → π^* transition. The position of absorption bands depends on the dye structure. The dyes studied differ in the type of N-alkylamino group in styryl moiety. Dye I possessing N,N-dimethylamino group in styryl moiety has the maximum absorption band at 534 nm. The introduction of N,N-diethylamino group in dye structure causes in the bathochromic shift of the absorption band about 10 nm. The presence of N-formylcarbazole moiety (dye V) provides the highest absorption band shift toward the longer wavelengths (about 20 nm). The molar absorption coefficient of dyes under study ranges from 20,000 to 60,000 dm³/mol/cm and depends on the dye structure and solvent polarity. Based on the results presented in Table 1, it is seen that the solvent polarity only weakly affects the position of the maximum absorption band. The increase in solvent polarity causes a slight bathochromic shift of long-wave absorption band (Fig. 2, Table 1). The polarity of the solvent has an insignificant impact on the solvatochromic properties of the new dyes.

The benzothiazolestyrylium dyes posses an intense fluorescence in the range from 550 to 750 nm (Figs. 4 and 5).

From the study of Strehmel et al. [12] on the spectroscopic properties of monocationic styryl pyridinium dyes, it is known that dyes possessing D⋅⋅⋅π⋅⋅⋅A structure may be characterized by the presence of three fluorescence states corresponding to different geometric conformations of dye in excited state. Scheme 2 presents the energy levels of three different emission states and the corresponding conformations of the benzothiazolestyrylium dyes.

The fluorescence spectra of dyes tested show the presence of long-wave band with a maximum in the range from 580 to 670 nm and the second emission band in the blue region of the spectrum at about 580 nm. Based on the energy levels shown in Scheme 2, it may be concluded that the long-wave band is an evidence of the presence of the lowest energy conformers, e.g., E or A₃ (planar conformation or twisted single bond). By contrast, the presence of a second emission band is a result of the rotation around the C-N single bond in styryl moiety. The position of maximum of emission depends on the dye structure. Dye I possessing the N,N-dimethylamino group in styryl moiety has the maximum of fluorescence at 605, 658, and 622 nm. The introduction of N,N-diethylamino group in dye structure causes in the bathochromic shift of the fluorescence band about 10–30 nm. The type of solvent affects on the position of the fluorescence band.

In order to determine the energy transition (singlet-state energy E ₀₀) between the lowest ground state and the lowest excited state, it was necessary to overlap both normalized absorption and fluorescence spectra of new sensitizers. Figure 6 presents the normalized absorption and fluorescence spectra of selected dye in acetonitrile as a solvent. The fluorescence spectra of the dyes tested are not a mirror image of the absorption spectra. Based on this, one can conclude that the geometry of sensitizer in both ground and excited states is different.

The values of energy E _0 → 0 are presented in Table 1. The difference in energy between the absorbed and emitted radiation is known as the Stokes shift [13]. The Stokes shift values obtained for novel bicationic monochromophoric benzothiazolestyrylium dyes are shown in Table 1. The Stokes shift of most of dyes tested is in the range from 2,000 to 3,600/cm. The highest value of Stokes shift (>3,000/cm) is observed in acetone as a solvent. The high values of Stokes shift suggest the difference in the dye geometrical structure in ground and excited states.

The electrochemical reduction of the dyes is reversible, as it is shown in Fig. 7.

In such dye-based photoinitiator systems, the initiating radicals are generated via photoinduced electron-transfer process [14]. Therefore, intermolecular electron transfer between the excited dye and organoborate salt must be thermodynamically allowed. Taking this into account, the values of free energy changes for electron-transfer process (ΔG _el), calculated from the Rehm-Weller equation (Eq. 1), must be negative [15].

The value of ΔG _el was calculated using the oxidation potential of n-butyltriphenylborate salt (E _ox = 1.16 eV), the singlet excited state energy (E ₀₀(S)) of the dye (Table 1) and the reduction potential E _red(Dye/Dye^•) of the sensitizers (see data in Table 1). The Coulombic energy coefficient (Z ₁ Z ₂/εr₁₂) was omitted in these calculations because the neutral radicals of borate salt, as well as sensitizers, were formed in the electron-transfer process. The E _ox and E _red of both photoredox pair components were determined from cyclic voltammetry measurements (Table 1).

The oxidation of tetramethylammonium n-butyltriphenylborate is dissociative, and this in turn causes the electrochemical process irreversible. The thermodynamically meaningful oxidation potential can be established using an indirect, kinetic method described by Murphy and Schuster [15]. Oxidation potentials measured electrochemically and kinetically differ by about 0.3 V. Therefore, the obtained electrochemical values for both components of photoredox pairs may have only approximate thermodynamic meaning, yet allow us to estimate roughly the free energy of activation (ΔG _el) for the photoinduced electron-transfer process.

The measured values of the reduction potential of sensitizers, the oxidation potential of the electron donor (E _ox = 1.16 eV), and the singlet-state energy of the dyes allow us to calculate the free energy change for the photoinduced intermolecular electron-transfer process. The estimated data are summarized in Table 2.

The values of ΔG _el for tested photoinitiating systems oscillate in the range from −0.248 to 0.152 eV. Taking this into account, the thermodynamic parameters presented in Table 2, it is seen that all combinations of benzothiazolestyrylium dyes/borate salt photoinitiating systems were found to possess small driving forces of electron-transfer process (ΔG _el > −0.248 eV) upon exposure to light. The estimated, according the Rehm-Weller equation, values of free energy activation for the electron-transfer process from borate anion to the excited hemicyanine dyes show that for tested photoredox pairs, the electron-transfer process is thermodynamically allowed (negative values of ΔG _el) only in the case of the following photoinitiators: IIIB and IVB. Basing on this, only these photoinitiators should initiate free-radical polymerization. But, the kinetic results obtained and presented in Table 2 do not confirm the thermodynamical requirement. Therefore, one should conclude that there are other factors which may effect on the overall rate of free-radical polymerization.

If the electron transfer is not diffusion controlled, it could be the rate-determining step for the polymerization. In such a case, the polymerization rate would increase with the increase in −ΔG _el values, at least up to the diffusion-controlled limit. The rate of electron transfer cannot be the rate-determining step if it approaches the diffusion-controlled limit (then it becomes independent of ΔG _el). In such a situation, other factors (for example, the reactivity of the primary radicals) would control the polymerization rate. Thus, the efficiency of any PET mechanism decreases as the viscosity of the polymerizing medium increases, and the formation of reactive radicals by the PET mechanism may be effective only at the initial time of polymerization.

Five photoredox pairs, consisting of benzothiazolestyrylium dyes (I)–(V) (acting as electron acceptors) and n-butyltriphenylborate salt (acting as electron donor), were tested as photoinitiating systems for free-radical polymerization of TMPTA.

To transform the obtained hemicyanine dyes into efficient free-radical polymerization initiating systems, an anion exchange from halide to borate was necessary. The ion exchange reaction was performed using the procedure given by Damico [9]. The final products were identified by ¹H NMR. The spectra and results obtained were the evidence that the tested salts were of desired structures.

Since the maximum of absorption band of tested dyes is located at about 530 nm for the photoinitiation of polymerization, the emission of an argon-ion laser emitting at 488 and 514 nm (Meles Griot) was used. Preliminary experiments show that TMPTA does not polymerize at 488 nm if any one component of the photoinitiating system was missing. The kinetic curves obtained during the polymerization of TMPTA/MP (9/1) mixture employing benzothiazolestyrylium dyes borate salts, under irradiation with a visible light, are shown in Fig. 8.

The rates of photoinitiated polymerization measured for all tested photoredox pairs are collected in Table 2. The initiation of polymerization via photoinduced intermolecular electron-transfer process involves many steps. Those include PET from an electron donor to an excited state of a dye or from an excited electron donor to the ground state of an electron acceptor followed by secondary reactions yielding a neutral radical initiating polymerization.

The step determining the reaction rate of the free-radical-initiated polymerization via intermolecular electron-transfer process (PET) is dependent on the nature of the dye and the electron donor (or acceptor). In the case of alkyltriphenylborate anion acting as an electron donor, the alkyl radical is formed as a result of rapid cleavage of the alkylboron bond within the boranyl radical produced after an electron transfer [5, 6]. According to the study Chatterjee’s et al studies on symmetrical cyanine borate initiators in nonpolar or medium polarity solvents, one can treat cyanine cation and borate anion as an ion pair. However, our studies on the influence of co-initiator concentration on the rate of photoinitiated polymerization show a distinct increase in the rate of polymerization as the concentration of borate anion increases for an identical monomer-dye formulation [14, 17, 18]. This finding suggests that at the concentration of borate anion equal to the concentration of dye cation, only a part of photoredox pairs exists as ion pairs. Since the electron transfer in cyanine dyes occurs in their singlet state, the existence of a cyanine cation and borate anion as ion pair is the basic prerequisite for the effective electron transfer. It is obvious that the additional amount of borate anion in the polymerizing composition should shift the equilibrium between free ion-ion pairs to a higher ion pair concentration and cause an increase in the photoinitiation efficiency of the dye-borate salt. It seems to be obvious that an artificial increase of the number of electron donor moieties within one molecule achieved by the coupling of the second borate anion to the dye molecule should improve the photoinitiating ability. Our experiments confirm this notion.

The data presented in Table 2 and Fig. 8 clearly indicate that the efficiency of TMPTA polymerization depends on the sensitizer structure. The highest efficiency for TMPTA free-radical polymerization was obtained for sensitizer possessing p-(N,N-dimethylamino)phenyl group, whereas the lowest efficiency was observed in the case of p-(pirydyno)phenyl substituent. These results are related to different reduction potentials (E _red) of hemicyanine dyes (see data presented in Table 1). Specifically, the dyes (I) and (V) can be readily reduced in the presence of co-initiator used, in comparison with dyes (III) and (IV) due to their lower reduction potentials. Therefore, the photoinitiating systems that contain IB have the highest values for photoinduced electron transfer ΔG _el (Table 2). The degree of conversion of the monomer into polymer also depends on the sensitizer structure (Fig. 9).

The values of degree of TMPTA conversion determined after a specific irradiation time are presented in Table 2. From the data obtained during the measurements, it was deducted that the conversion of monomer ranges from 4.75 to 16.7 %.

Another interesting feature of the tested new heterobicationic benzothiazolestyrylium dyes tested is worthy of attention. Analyzing the data summarized in Table 2, one can find that the type of substituent present in the styrylium moiety attached to the benzothiazole ring causes a significant variation in the photoinitiating abilities of the bicationic hemicyanine borates. The sensitizers possess asymmetrical structure that consists of an electron-donating moiety and a heterocyclic moiety, which can act as an electron-deficient acceptor. Furthermore, donor and acceptor are in direct electronic conjugation via a vinyl π-system, and the acceptor moiety is charged [13]. We tested asymmetrical cyanine dyes with electron donor-acceptor moieties on the opposite sides of the vinyl group with additionally bonded pyridinium ring. The electron-donating moiety possesses various substituents in para position of phenyl ring (except dye (V)). On the basis of our experiments, it appears that the initiators possessing the dimethylamino, N-formylcarbazole and diethylamino groups in para position of phenyl ring (IB, VB, IIB), initiate the free-radical polymerization with the highest rates. For the cationic photoinitiating systems, the photoinitiating ability decreases in the following order:

This suggest that the presence of –N(CH₃)₂, –N-formylcarbazole, or /–N(C₂H₅)₂ in styrylium moiety causes an enhancement of the electron donor concentration (borate anion) in the proximity to the absorbing dye. In summary, their photoinitiating ability is comparable or lower than the photoinitiating ability of well-known commonly used photoinitiators operating in the visible-light region [1, 8, 13, 16‐18].

The degree of monomer conversion and also the rate of polymerization depend on both the structure and reduction potential of the dye (Figs. 10 and 11).

The concentration of the photoinitiating system itself also plays a key role in photopolymerization. In a conventional UV-Vis photopolymerization, R _p increases with the increase of concentration of the initiator up to a certain value, than it decreases. This behavior is attributed to the “internal filter effect.” This becomes more significant for photoinitiators with high molar absorption coefficient [19]. The kinetic curves recorded for photoinitiated polymerization at different photoinitiator concentrations are presented in Fig. 12.

Data in Fig. 12 shows that the time needed to reach the curve maximum decreases when the initiator (IB) concentration increases. The observed property is related to the rate of heat evolution in the initial time of polymerization. Figure 13 presents the relationship between the rate of polymerization and the photoinitiator concentration.

It is evident that as photoinitiating system concentration increases, the rate of polymerization increases and reaches a maximum. The highest rates of polymerization (for 3-mm thick sample) were achieved at an initiator concentration of about 5 × 10⁻³ M for the tested photoinitiating systems with two electron donors in one molecule.