Kinetic Model under Light-Limited Condition for Photoinitiated Thiol-Ene Coupling Reactions.

Thiol-ene click chemistry has become a powerful paradigm in synthesis, materials science, and surface modification in the past decade. In the photoinitiated thiol-ene reaction, an induction period is often observed before the major change in its kinetic curve, for which a possible mechanism is proposed in this report. Briefly, light soaking generates radicals following the zeroth-order reaction kinetics. The radical is the reactant that initializes the chain reaction of thiol-ene coupling, which is a first-order reaction. Combining both and under the light-limited conditions, a surprising kinetics represented by a Gaussian-like model evolves that is different from the exponential model used to describe the first-order reaction of the final product. The experimental data are fitted well with the new model, and the reaction kinetic constants can be pulled out from the fitting.

Introduction

Click reactions have been one of the widely used methods for chemical synthesis because of their properties, as defined by Sharpless in 2001, which are as follows: (a) they produce high yields under mild conditions, (b) they are not affected by the presence of oxygen and water, and (c) they are compatible with a wide variety of starting materials, which selectively react only with each other (i.e., the reaction is orthogonal to other organic reactions).[1] The reaction between an alkene, or an alkyne, and a thiol is widely accepted as having the properties of a click reaction.[2] It has become a powerful tool in various synthetic methodologies, biofunctionalization, surface and polymer modification, and polymerization.[3−15] This thiol–ene click reaction can follow two different mechanisms depending on the type of initiator that is used.[8] The base-catalyzed thiol–ene (also called thiol-Michael addition) reaction is initiated by adding a nucleophile, usually phosphines.[16,17] Or the reaction is initiated by exposing the thiols to heat or UV light, which generate thiyl radicals that will react with the alkene, or alkyne (Figure A).[18−22]

Figure 1

(A) Scheme for the simplified mechanism of the photoinitiated thiol–ene coupling reaction catalyzed by the addition of a chemical photoinitiator (PI).[22] (B) Schemes of the two reactions in this report.

In a series of recent studies, the most important factor governing the overall kinetics of thiol–ene polymerization was found by Bowman and co-workers to be the ratio of the propagation rate (kP) to the chain-transfer rate (kCT).[8,16,17,20,23−29] The reaction kinetics is simplified into first-order reactions with respect to either thiol or alkene, or both, depending on the kP/kCT ratio.[24] This explains the overall exponential growth curve of the product but ignores the short induction period that is often observed before the steep portion of the growth curve, making it difficult to extract the rate constants from the curve fitting. Here, we propose an expanded mechanism to explain the induction period using the same kinetic model from the literature but taking the photoinitiator into account.

In this report, we measured the reaction kinetics of vinyltrimethoxysilane (VTMS) with thioglycolic acid (TGA) photoinitiated by 2,2-dimethoxy-2-phenylacetophenone (DMPA), and VTMS with cysteamine hydrochloride (CAH) also photoinitiated by DMPA (Figure B) using 1H NMR spectroscopy. The mixture was exposed to a 368–400 nm light source that is 3 cm away from the sample. Silanes were used in this study because many applications require the immobilization of a molecule tailored for a specific property via the silane group. The ability to couple a molecule to a silane through thiol–ene click reactions widens the range of molecules that can be grafted onto oxide surfaces.

Experimental Section

All reagents were purchased from Sigma-Aldrich, and all spectra were collected using a Bruker Ascend 500 MHz NMR spectrometer under the same instrument options. Reactants TGA (or CAH), VTMS, and photoinitiator DMPA were added into an NMR tube, with solvent CDCl3 for the VTMS (0.427 M), TGA (0.857 M), and DMPA (0.03 M) mixture, and (CD3)2CO–D2O (1:4 v/v) co-solvent for the VTMS (0.058 M), CAH (0.112 M), and DMPA (0.012 M) mixture to dissolve both the CAH salt and the nonpolar aromatic DMPA. Both solutions were split into two equal portions. One portion was exposed to a 368–400 nm light source with a visible light-blocking filter (Sunlite SL20 lamp, 20 W). The lamp was placed ∼3 cm away from the sample, where the power density was ∼4 mW/cm2 measured using a Thor Labs PM100-S121C photodiode sensor. UV exposure was done at specific time intervals under room temperature prior to 1H NMR measurements. The remaining portions of both solutions were left unexposed and kept as the experimental controls. Baseline correction of all 1H NMR spectra, peak area calculations, and curve fitting (using least-squares regression analysis) were performed using MATLAB.

Results and Discussion

The 1H NMR spectra in Figure confirm the formation of the click reaction products from the thiol–ene coupling of the VTMS with TGA (Figure A), or with CAH (Figure B). The coupling reaction in the mixture of VTMS, TGA, and DMPA was complete in less than 5 min of exposure to UV light (Figure A). The NMR spectra before and after UV exposure in Figure A clearly show the disappearance of the vinyl peaks in VTMS (at 5.8, 5.9, and 6.1 ppm) and the appearance of new peaks at 1.0 and 2.7 ppm that indicate the presence of the thioether product.[30] The intensity of the TGA thiol peak at around 2.0 ppm was also reduced to half, which is consistent with the initial thiol–vinyl concentration ratio of 2:1 and a 1:1 thiol–vinyl reaction stoichiometry. These initial ratios have been reported to produce the highest yield in the shortest time.[9] The peaks around 3.6 ppm in Figure A come from the hydrolysis of the Si–OCH3 groups. Control experiments showed that indoor room lighting has no effect and that UV light is necessary for initiating the thiol–ene reaction. Without UV exposure, the NMR spectra in the control experiment (Figure C) indicate that no reaction occurs after 1 h of incubation at room temperature and under room light.

Figure 2

(A) 1H NMR spectra of the reactant mixture of VTMS, TGA, and DMPA in CDCl3, before (red trace) and after (blue trace) exposure to the solar simulator for 300 s. (B) 1H NMR spectra of the reactant mixture of VTMS, CAH, and DMPA in (CD3)2CO–D2O (1:4 v/v) co-solvent, before (red trace) and after (blue trace) exposure to the solar simulator for 150 s. (C, D) 1H NMR spectra of control experiments of the unexposed reaction mixtures of (C) VTMS with TGA showing no change in the vinyl and thiol peaks, and (D) VTMS with CAH also showing no change in the vinyl peak without UV light exposure.

The absence of VTMS-homopolymerization products is consistent with the click property of the thiol–ene coupling reported in the literature,[8] whose products would have shown NMR peaks between 0.5 and 0.9 ppm for the methyl protons in the −Si–CH2– group,[31−33] which were negligible in the 1H NMR spectra that were collected.

The 1H NMR spectra of the mixture of VTMS, CAH, and DMPA before and after exposure to UV light are shown in Figure B. The hydrogen peak of HOD at ∼4.4 ppm is broadened due to its hydrogen bonding to the other molecules; peaks from the −SH and NH3+ exchanges with D2O and are not observed; hydrogen in Si–OCH3 is observed at 3.2 ppm; hydrogens in N–CH2–CH2 are observed around 3.8 and 3.4 ppm, respectively.[34] The progress of the reaction can be most clearly tracked through the disappearance of the signal from the vinyl group at 5.9 ppm and the appearance of the thioether-specific peaks at 2.7 and 0.9 ppm.[30] The reason for seeing only one peak near the center of the chemical shifts for the vinyl group is due to the peak broadening caused by self-aggregation.[35] The methoxy groups of VTMS are partially hydrolyzed in D2O and the silanes are cross-linked, which may cause the peaks between 3.2 and 3 ppm. The reaction between VTMS and CAH was complete after 150 s of UV exposure (Figure B), with the thioether as the only product. The spectra from the control experiment without UV light is shown in Figure D, again confirming that UV light is required as indicated by the absence of the thioether product peaks (2.7 and 0.9 ppm) and the retention of the vinyl peak (5.9 ppm).

The kinetics of both reactions were monitored with a series of 1H NMR spectra collected after every exposure interval (Figure ), with the decay curves shown in Figure . In the VTMS–TGA coupling (Figure A), the vinyl (red traces) peak gradually disappears, while the thiol (black traces) peak was gradually reduced to roughly half after exposure to UV light over time. At the same time, there was a gradual increase in the peak intensities centered at 2.73 ppm (green traces) and 0.98 ppm (blue traces), both specific to the thioether product. Figure B shows a plot of the integrated peak areas corresponding to each moiety in Figure A as a function of exposure time. The data indicate that the photoinitiated reaction between VTMS and TGA has a short induction time followed by an exponential decay of the reactants, with a corresponding exponential growth of the thioether product. The yield for TGA and VTMS is >90%. The yield for CAH and VTMS is ∼50%. The relatively lower yield of the VTMS and CAH reaction is due to the recently observed inhibition of thiol–ene reactions by amine groups.[36] Induction times were also observed in other photocatalytic reactions.[19,28,29,37] The length of the induction time is dependent on the reaction conditions, such as light intensity.[25,38]

Figure 3

(A) Reaction kinetics in bulk solution containing VTMS, TGA, and DMPA in CDCl3 measured by 1H NMR. The sample was exposed to 368–400 nm light for specific time intervals (0, 9, 14, 30, 35, 40, 45, 50, 55, 60, 70, 80, 90, 120, 180, and 300 s) prior to each 1H NMR measurement. (B) Similar experiments performed on a reaction mixture containing VTMS, CAH, and DMPA in (CD3)2CO–D2O co-solvent (1:4 v/v) and at time intervals of 0, 5, 10, 15, 20, 25, 30, 40, 50, 60, 75, 90, and 150 s.

Figure 4

(A) Selected NMR spectra of the reaction mixture of TGA, VTMS, and DMPA in CDCl3 focusing on the vinyl (red), thiol (black), and product (green and blue) regions from Figure . (B) Plots of integrated peak areas of the color-coded regions shown in (A). The scatter plots are integrated peak areas after each exposure time for each of the corresponding peaks in (A), which are fitted with the models described in the text. (C) Selected NMR spectra of the reaction mixture of CAH, VTMS, and DMPA in (CD3)2CO–D2O (1:4 v/v) co-solvent focusing on the vinyl (red) and product (green and blue) regions. (D) Plots of integrated peak areas of the color-coded regions shown in (C), where the scatter plots and solid traces represent integrated peak areas and fitted curves, respectively. These curves were fitted using eqs –13.

The kinetics of the photoinitiated thiol–ene reactions is best described by a half-Gaussian function, as shown by the solid curves in Figure B,D. The photocatalytic reaction cycle shown in Figure has been assumed to have two competing steps in the literature, chain transfer and chain propagation.[8,27] The relative rate constants of these processes and the chain initialization and termination processes will determine the kinetic trajectory of the reaction. These processes can be specified to have the following elementary reactions expanded from the model in the literature[27]where PI is the photoinitiator that has not been considered in the literature model;[24,27] R is an unknown molecule (e.g., another PI, oxygen, or other molecules); S, V, and SV are the thiol, vinyl, and thioether products, respectively; and “•” indicates if the species is a radical. These reactions can be grouped into three major processes. Reactions and 2 are the initiation steps of the photochemical reaction. Reactions and 4 are the chain-transfer and chain-propagation steps, respectively. Reactions –7 are the termination steps of the photoinitiated thiol–ene reaction. Varying the rates of reactions –7 has resulted in various exponential curves with no obvious induction period.[24] Thus, we assume the photoactivation steps to be the rate-limiting steps to check if they contribute to the induction period. This assumption is reasonable because of the low light intensity used in this study (4 mW/cm2, ∼4% sunlight intensity at air mass 1.5).

Solving these reactions is difficult for us, but useful equations can be drawn with approximations. The reaction rate can be expressed as a function of exposure time, t, and reactant concentrations, from step 4, which produces the thioether productWe can assume that the concentration of the activated [SV•] is small and rate-limited by the light exposure because the concentration of the photoinitiator [PI] is much smaller than the concentration of vinyl [V] and thiol [S] and because the light source is relatively weak, thus making [PI•] and [R•] small. Assuming typical light-limited photoreactions as zeroth-order reactions and assuming a relatively small termination rate compared to the chain-transfer rate,[27] the concentration of the photoactivated [SV•] radical at any time, t, can be approximated aswhere k8 is a new constant, which under light-limited conditions is directly proportional to k1, e.g., k8 = k1, when the radical propagation rate is much higher than the termination rate. By combining eqs and 9, we getwhere k9 = k4 × k8 is another new constant, [S]0 is the initial concentration of the thiol, which is consumed to form the equal amount of [SV]. After integration with the boundary condition that at time = 0, [SV]0 = 0, we getThese three Gaussian equations are different from the exponential decay functions used in the literature,[24] but explain our light-limited induction period better (Figure B,D, fitted curves). A short induction period has been commonly observed in the literature for thiol–ene and thiol–yne couplings.[22,27,39] This bulk kinetic analysis allowed us to estimate the value of k9 under our experimental conditions (Table ) and thus select a relatively longer exposure time than the half-life, t1/2, for the surface-grafting experiments, where from eq We note that at a higher light intensity, the photoactivated photoinitiator reaches a steady state quickly (when eq becomes [SV•] = k8), and the kinetic equations (eqs and 12) converge into the single-exponential decay functions found in the literature.[22,27,39] The treatment of the rate-limiting step is different in this report compared to that in the literature, which allows us to predict the induction period observed under our reaction conditions.

Table 1

Rate Constants and Half-Lives of Thiols (from Equation ) in the Photoinitiated Thiol–Ene Reactions Calculated from the 1H NMR Data (Figure )

reaction	k9 (s–2)	t1/2 (s)
VTMS + TGA	0.0011	36
VTMS + CAH	0.0014	31

In summary, we have extended the existing kinetic model and demonstrated proof-of-concept experimental evaluations to provide a more complete reaction mechanism for thiol–ene click reactions. The observed consumption rates of the reactants and formation rates of products were best fit by half-Gaussian functions (eqs –13) that contain induction periods. These functions look empirical but are supported by the proposed light-limited reaction mechanism. When the photoinitiating step of the chain-reaction model is the rate-limiting step, a short induction period emerges prior to the exponential-like growth of the product (or decay of the reactant) concentration. This mechanism may help the researchers to elucidate and understand the factors that underlie the confusing kinetic curves observed before.