Photoinitiated Polymerization - ACS Publications - American Chemical

Chapter 2

Spectroscopic Investigation of Three-Component Initiator Systems 1,2

1

Downloaded by FUDAN UNIV on February 20, 2017 | http://pubs.acs.org Publication Date: March 3, 2003 | doi: 10.1021/bk-2003-0847.ch002

1

Kathryn Sirovatka Padon , Dongkwan Kim , Mohamed El-Maazawi , and Alec B. Scranton

1

1

Department of Chemical and Biochemical Engineering, University of Iowa, 125 Chemistry Building, Iowa City, IA 52242 Current address: Kimberly-Clark Corporation, 2100 Winchester Road, Neenah, WI 54956 2

In this contribution, photodiffrential scanning calorimetry (DSC) and in situ time-resolved steady state fluorescence spectroscopy are used to study the initiation mechanism of the three-component systems: 1) methylene blue/N-methyldiethanolamine/diphenyliodonium chloride (MB/MDEA/DPI), 2) eosin Y/MDEA/DPI (EY/MDEA/DPI), and 3) eosin Y, spirit soluble/MDEA/DPI (EYss/MDEA/DPI). Kinetic studies reveal that the photopolymerization behavior for eosin dyes is quite similar to that observed for methylene blue. The fastest polymerization rate is obtained when all three components are present, the next fastest with the dye/amine pair, and the slowest with the dye/iodonium pair. In the MB/MDEA/DPI system, it is concluded that the primary photochemical reaction involves electron transfer from the amine to the dye.

© 2003 American Chemical Society

Belfield and Crivello; Photoinitiated Polymerization ACS Symposium Series; American Chemical Society: Washington, DC, 2003.

15


16 We suggest that the iodonium salt reacts with the resulting dye-based radical (which is active only for termination) to generate the original dye and simultaneously produce a phenyl radical that is active in initiation. Moreover, oxygen is found to quench the triplet state of the dye leading to retardation of the reaction. In the eosin/MDEA/DPI system, experimental results showed that the two eosin dyes function quite similarly in the three component system eosin/MDEA/DPI. We propose a sequence of steps in which the dye is rapidly oxidized by DPI in a reaction that is not highly efficient at producing initiating radicals. However, the dye may also be reduced by MDEA producing active initiating amine radicals. Another active radical may be produced if the reduced dye radical is further reduced to its leuco form by an additional molecule of amine. Additionally, the reduction of the bleached dye radical (formed via reaction with DPI) by the amine has a two fold effect. First it regenerates the original eosin dye, and second, it produces an active amine radical in place of the less active dye radical. Dual cell UV-visible absorption spectroscopy yielded no evidence for the formation of a ground state complex between the eosin dyes and the iodonium salt.

Introduction Free radical photoinitiators, which produce active centers upon absorption of light, have traditionally been built around the benzoyl chromophore, which absorbs light in the ultraviolet (UV) region of the spectrum. Some common classes of these photoinitiators include benzoin ethers, dialkoxyacteophenones, hydroxy alkyl ketones, benzoyl oxime esters, amino ketones, and morpholino ketones. When illuminated with U V light, these photoinitiators produce active centers efficiently by the well-known α-cleavage process. However, the traditional α-cleavage initiators have some important limitations. For example, ultraviolet light is undesirable for applications such as photopolymer master printing plates (7-3) and photopolymerizations in biological systems (4,5). The widespread availability and low cost of visible light sources has made visible initiating systems very attractive for use in photopolymer plates. In addition, for biological applications such as dental composites and bone cements, visible light is preferable due to the damaging effects of UV radiation.



17 Until recently, the need for visible light radical photoinitiators has been addressed primarily by using bimolecular initiator systems. In these systems, active radicals are generally produced via electron transfer and subsequent proton transfer from an amine to a photo-excited molecule. Using this approach, the system can be tailored to a specific wavelength range by choice of the lightabsorbing species. In the last decade, three-component photoinitiation systems have emerged as an improvement over two-component electron transfer initiating systems. Like the two component systems, the three-component initiators include a light absorbing moiety which is typically a dye and an electron donor which is almost always an amine. The third component is usually an iodonium salt. Three-component systems are extremely flexible since a wide variety of dyes may be used. As in the two-component systems, the selection of the dye determines the active wavelength; classes of dyes that have been reported for three-component systems include ketones, xanthenes, thioxanthenes, xanthones, thioxanthones, coumarins, ketocoumarins, thiazines, and merocyanines. These three-component systems have consistently been found to be faster, more efficient, and more sensitive than their two-component counterparts (2,6-14). In addition to the advantages of enhanced cure rate and higher sensitivity mentioned above, three component systems have some other significant features which make them quite interesting. For example, a single combination of electron donor and iodonium salt can be used with a variety of different dyes thereby providing tremendous flexibility in selection of the initiating wavelength. In addition, with proper selection of the components the same initiating systems may be effective for initiation of cationic polymerizations as well as radical polymerizations (75). In this contribution we report a series of spectroscopic studies of three different three-component radical photoinitiator systems. The first system consists of a methylene blue dye (MB), N-methyldiethanolamine (MDEA) as the electron donor component, and diphenyliodonium chloride (DPI) as the third component. Diphenyliodonium chloride and N-methyldiethanolamine were chosen because they are essentially standard constituents in three-component systems. Methylene blue, however, has not been well studied and was selected due to its unique characteristics. For example, since methylene blue is a cationic dye, electrostatic repulsion precludes direct reaction with the iodonium cation. Furthermore, the entire class of cationic dyes has not been well studied in threecomponent systems. Finally, methylene blue is of extremely low toxicity, making it an appealing choice for biological applications of three-component systems. The second system we investigated consisted of the neutral dye Eosin Y (spirit soluble), N-methyldiethanolamine, and diphenyliodonium chloride. The third system was eosin Y (doubly charged anionic dye), Nmethyldiethanoiamine, and diphenyliodonium chloride. The amine and iodonium components are identical to those in the previous study, and are relatively


18 standard constituent choices in three-component systems because their efficacy is well-demonstrated. The use of the eosin Y dyes provided some interesting comparisons to the system containing methylene blue.

Experimental


Materials As in previous publications (16-18), the monomer 2-hydroxyethylmethacrylate was obtained from Aldrich. It was treated with DeHibit (from Polysciences) to remove hydroquinone inhibitor and filtered prior to use. Methylene blue and Eosin Y were obtained from Aldrich, and Nmethyldiethanolamine (MDEA) and diphenyliodonium chloride (DPI) were obtained from Fluka, and were used as received.

Photo-differential scanning calorimetry By continuously monitoring the heat flow from the reacting sample, differential scanning calorimetry (DSC) provides a complete polymerization exotherm which yields both instantaneous reaction rate and conversion data. Photo-DSC was conducted using a Hewlett-Packard DSC 7 modified in-house for photo-experiments. The light source was a 200 W Oriel mercury-xenon lamp. The beam was passed through a waterfilteroutfitted with a thermostatted recirculating jacket to reduce infrared radiation and limit sample heating. A glass cover on the DSC sample block was used to eliminate any ultravioletinduced photoinitiation. The total radiant power incident on the sample was 63 mW/cm , as measured by graphite disc absorption. This measurement includes all visible region light, not all of which can be absorbed by the methylene blue photosensitive component. Comparison of the lamp's spectral characteristics with the absorbance profile of the methylene blue dye reveals that only about 6% of this total radiant intensity can be absorbed by the dye, for an effective incident power of approximately 4 mW/cm . This low light intensity was used in order to produce a relatively slow reaction rate that can be accurately monitored using DSC. A nitrogen purge, flowing at approximately 25 cmVmin, was used in the DSC to eliminate oxygen inhibition of the polymerization. Samples were degassed for a minimum of five minutes prior to illumination. The heat flow data collected by the DSC can be easily converted to the polymerization rate data reported in this contribution using the heat of polymerization, which is 49.8 kJ/mol for HEMA. 2

2


19

DSC Dark-Cure Experiments The DSC apparatus, described above, was used to obtain dark polymerization exotherms for samples that contained 0.00010 M methylene blue and 0.013 M MDEA in ethylene glycol dimethacrylate (EGDMA)/HEMA 8:1 mass ratio at various DPI concentrations between 0 and 5 mM. Each sample was illuminated for 2 minutes then the light was shuttered. The decay in the heat flow was then recorded until the curve was flat at 0 W/g. All experiments were performed at a temperature of 25°C and light intensity I = 0.3 mW/cm . 2


a

Time-resolved steady state fluorescence The experimental details for the in situ time-resolved steady state fluorescence monitoring studies have been described previously (16-18). Fluorescence may be used to characterize the dye concentration as a function of time because the fluorescence signal decays as the dye is consumed in the initiation reaction. Most studies were carried out in a thin film geometry using a custom-made sample cell that allowed both temperature control and a controlled atmosphere. Typically, fifty complete fluorescence spectra in the wavelength range 700-850 nm were taken at time intervals chosen to capture the entire decay of the methylene blue or Eosin fluorescence. To investigate the effect of oxygen, some studies were performed using a capillary cell which consists of a 1 mm diameter capillary tube held perpendicular to the laser beam.

Dual-cell UV-visible absorption spectroscopy Ultraviolet-visible absorption spectroscopy was used to investigate the possible formation of ground state complexes in all of the three component initiator systems in this study. The experimental protocol was as follows: Solutions of each component were prepared in methanol. For each pair of components, a reference scan was carried out using two cuvettes sandwiched together, each containing one of the components. The samples were then mixed and placed in a 2.0 cm pathlength cuvette, and the difference between the resulting absorbance spectrum and that obtained with the sandwiched cuvettes was obtained. In this way, any signal observed in the sample can be attributed to the effect of mixing the two components.


20

Results and Discussion


Methylene blue, N-methyldiethanolamine and diphenyliodonium chloride system The photo-DSC data illustrated that each component plays an important role in the photopolymerization reaction (16). The maximum rate was observed when all three components were present (maximum rate 0.019 mol/L/s observed at 10.5 min). The two component system containing MDEA and MB exhibits the second fastest reaction rate (maximum rate 0.011 mol/L/s observed at 18 min), while a two-component system containing DPI and MB exhibits a relatively slow reaction (maximum rate 0.005 mol/L/s observed at 26 min). Next, a more detailed series of studies on the three-component system was performed to investigate the effects of the concentrations of DPI and MDEA on the observed polymerization rate. With the concentrations of MB and DPI held constant, as the MDEA concentration is increased from zero to 0.051 molar, the maximum polymerization rate was observed to increase from 0.005 to 0.024 mol/L/s. Similarly, the reaction rate also increases dramatically with increasing DPI concentration. With the concentrations of MB and MDEA held constant, as the DPI concentration is increased from zero to 0.0049 molar, the maximum polymerization rate was observed to increase from 0.010 to 0.025 mol/L/s. Therefore, these kinetic studies based upon photo-differential scanning calorimetry revealed a significant increase in polymerization rate with increasing concentration of either the amine or the iodonium salt. However, the laserinduced fluorescence experiments show that while increasing the amine concentration dramatically increases the rate of dye fluorescence decay, increasing the DPI concentration actually slows consumption of the dye (16). Based upon the experimental trends described above, we concluded that the primary photochemical reaction involves electron transfer/proton transfer from the amine to the dye. We suggest that the iodonium salt reacts with the resulting dye-based radical (which is active only for termination) to regenerate the original dye and simultaneously produce a phenyl radical (active in initiation) derived from the diphenyliodonium salt as shown in scheme I. The role of the iodonium salt proposed in scheme I has two primary implications: First, since the DPI serves to regenerate the methylene blue, it will reduce the rate at which the dye is consumed in the photoinitiation reaction (consistent with the fluorescence observations described above); Second, since the DPI consumes the methylene blue radical (which is active for termination but not for initiation) an increase in the DPI concentration should lead to a decrease in the rate of termination. In this way, DPI can enhance the polymerization rate by two distinct effects.


21

N

[MB—Cf]* +

* "ΌΗ C

2

2

H C-N^ CH CH OH

2

2

-·• C H - < 0 >

2

2

3

2

3

. CH CH OH MB* + H C—Ν + HCI CH CH OH

^Η ΰΗ ΟΗ 2

[MB^-Clf

H

—

-'

+

2

+

MB—CI*


cr

Scheme I: Major

reaction steps for the three component system

MB/MDEA/DPI

To test the effect of DPI on the rate of termination, dark cure studies were performed using differential scanning calorimetry. The dark-cure results shown in Figure 1 illustrate that DPI does indeed affect the rate of termination. Samples with higher DPI concentrations exhibit greater and more persistent levels of dark cure than samples with lower DPI concentration. There results suggest that termination is reduced in the samples with higher DPI concentration. Furthermore, as shown in Figure 1, the initial rate when the dark cure experiment begins (at time = 0 seconds) is higher for samples with higher DPI concentrations. This is a natural effect of the increase in photopolymerization rate with increasing DPI concentration. However, the observed effect on the termination step cannot be attributed to hindered diffusion of growing chains arising from a higher degree of crosslinking for the samples with higher DPI concentrations because the overall conversion is quite low for these samples (the sample with the highest DPI concentration reaches only 7% of its final conversion during the first two minutes of exposure). For this reason, variation in degree of crosslinking is expected to be minimal and should not significantly affect the termination step. Hence, the primary cause of the variation in the rate of termination illustrated in Figure 1 is the consumption of terminating MB radicals via reaction with DPI. A series of studies were also performed to characterize the influence of oxygen on the rate of consumption of the methylene blue dye (17). Oxygen quenches the triplet state of the dye, leading to retardation of the reaction. We use time-resolved steady state fluorescence monitoring to observe the methylene blue concentration in-situ in both a constant-oxygen environment and a capillary


22 tube cell as the dye is consumed via photoreaction. In the capillary tube cell, we observed a retardation period (attributed to the presence of oxygen) followed by rapid exponential decay of the methylene blue fluorescence after the oxygen was depleted. Based on the impact of the amine and iodonium concentrations on the fluorescence intensity and the duration of the retardation period, our proposed mechanism includes an oxygen-scavenging pathway in which the tertiary amine radicals formed in the primary photochemical process consume the oxygen via a cyclic reaction mechanism. 3 0.24

Ί


DPI Concentration

ε

o.2o

5.0 mM 1.0 mM 0.0 mM

I ο-™ LU

•g CO Ν

0.12

C

£ 0.08 Ο

Q- 0.04

•s Û

0.00 o.o

0.5

—r1.0

1.5

2.0

2.5

Time (min)

Figure 1: Dark polymerization exotherms for various DPI concentrations. Increasing DPI slows termination of the polymerization reaction. [MB] = 0.00010 M, [MDEAJ = 0.013 M in EGDMA/HEMA 8:1 mass ratio. Τ = 25°C I = 0.3 mW/cm . Each curve is the average of three replicate trials. 2

a

Eosin Y, N-methyldiethanolamine and diphenyliodonium chloride system The dyes eosin Υ (ΕΥ) and eosin Y, spirit soluble (EYss) (18) were chosen as model anionic and neutral dyes with charges of -2 and 0, respectively. In contrast to the cationic methylene blue dye, the eosin dyes are able to react directly with both the amine and the iodonium components. Although the ΕΥ/DPI reaction is expected to proceed more vigorously than the EYss/DPI reaction because of the possible electrostatic attraction between the negatively



23

charged ΕΥ dye moiety and the positively charged iodonium ion, experimental results showed that the two dyes function quite similarly in the three component system eosin/MDE A/DPI. For both of the eosin dyes, kinetic studies based upon photo-differential scanning calorimetry reveal that the fastest polymerization occurs when all three components are present, the next fastest with the dye/amine pair, and the slowest with the dye/iodonium pair. However, the laser-induced fluorescence experiments show that the pairwise reaction between eosin and iodonium bleaches the dye much more rapidly than does the reaction between eosin and amine. In this system, we concluded that although direct eosin/amine reaction can produce active radicals, in the three-component system this reaction is largely overshadowed by the eosin/iodonium reaction, which does not produce active radicals as effectively. We propose that the amine reduces the oxidized dye radical formed in the eosin/iodonium reaction back to its original state and simultaneously produces an active initiating amine-based radical. Due to the difference in the pairwise reaction rates for eosin/amine and eosin/iodonium, it seems likely that this regeneration reaction is the primary source of active radicals in the three-component eosin/amine/iodonium system. Although, as indicated above, EY and EYss were found to react similarly in the eosin/MDEA/DPI three component system, direct reaction between DPI and EY appears to be somewhat more efficient than that between DPI and EYss. The rate of photocuring for the two-component EYss/DPI pair is significantly slower than that for the ΕΥ/DPI pair. This is most likely due to increased initiation efficiency of the radicals generated rather than a significant difference in the rate of the ΕΥ/DPI reaction, since the decay profiles for the ΕΥ/DPI and EYss/DPI pairs are essentially identical within experimental errors, as shown in Figure 2. (The characteristic decay constants determined from the exponential fit are also equal within experimental errors). The experimental results described above are not inconsistent with the formation of a ground-state complex between eosin dye and the iodonium salt. In fact, Fouassier and Chesneau (19) report the formation of a ground state complex between a singly charged eosin dye and diphenyliodonium complex under certain conditions. For our two eosin dyes (which are anionically double charged and neutral, respectively) we performed studies using dual-cell UVvisible absorption spectroscopy to seek evidence for the formation of a ground state complex. In these studies, for each of the eosin dyes, the visible absorption spectrum was unchanged by the addition of DPI. Therefore, while the DSC evidence suggests that the ΕΥ/DPI produces faster photocuring than EYss/DPI, it seems unlikely that this difference is due to ground-state pre-association of the negatively charged EY dye moiety and the positively charged DPI. Instead, it is likely that difference is due to an enhanced initiation efficiency of the radicals produced from the ΕΥ/DPI reaction as compared to the EYss/DPI pair. This


24


difference may be attributable to the dye charge. After being oxidized by DPI, the charge on EY is reduced from -2 to -1. However, the charge on the EYss dye moiety changes from neutral to +1. MDEA has a dipole moment due to the lone electron pair on the central nitrogen atom. This region of high electron density will experience a greater attraction to the positively charged EYss radical than the negatively charged EY radical.

0.2-1

, 0

.

, 1

. 2

,

.

1 3

.

r 4

Time (sec)

Figure 2: Comparison of Eosin Y and Eosin Y, spirit soluble photobleaching DPI. [EY] = [EYss] = 0.0010 M, [DPI] = 0.001 M, all in HEMA, Τ = 25 V I = 50 mW/cm . No MDEA is present. Each curve represents the average three trials 2

a

Conclusions We have presented experimental characterization of three different three component initiator systems. The three-component system comprised of methylene blue, MDEA, and DPI appears to proceed by a mechanism involving electron transfer/proton transfer from the amine to the dye as the primary photochemical reaction. Interaction between the dye and the iodonium salt is precluded by charge repulsion. The iodonium salt is an electron acceptor, acting to re-oxidize the neutral dye radical back to its original state and allowing it to re-enter the primary photochemical process. This reaction simultaneously generates phenyl radicals that can serve as initiating radicals. Therefore, the



25 iodonium has a double role: to regenerate the methylene blue and to replace the inactive, terminating dye radical with an active, initiating radical. Based on the impact of the amine and iodonium concentrations on the fluorescence intensity and the duration of the oxygen-induced retardation period, the reaction scheme includes an oxygen-scavenging pathway in which the tertiary amine radicals formed in the primary photochemical process consume the oxygen via a cyclic reaction mechanism. The three-component systems comprised of neutral or anionically doubly charged eosin Y dyes, MDEA, and DPI appear to produce active radical centers by an electron transfer/proton transfer reaction from the amine to photoexcited dye, and/or to dye that has been reduced by the iodonium. The eosin may be oxidized by the DPI to produce an eosin radical, which may be in turn reduced by an amine back to the original state. This reaction has a twofold effect: first, it regenerates the original eosin dye, and second, it produces an active amine radical in place of the presumably less active dye radical. Because of the difference in reaction rates for the pairwise reactions, this dark amine-mediated dye regeneration reaction may be the primary source of active radicals in the three-component system. No spectroscopic evidence for the formation of ground state complexes between the eosin dyes and the iodonium salt was observed.

Acknowledgements This material is based upon work supported under a National Science Foundation Graduate Fellowship. (KSP) The authors also acknowledge the assistance of Dr. Ruiping Huang in developing and setting up the thin film fluorescence monitoring experiments. This work was partially supported by the IUCRC Center for Fundamentals and Applications of Photopolymerization.

References 1. 2. 3. 4.

Kawabata, M., and Takimoto, Y., J. Photopolym. Sci. Tech., 1990, 3, 147. Fouassier, J. P.; Ruhlmann, D.; Graff, B.; Takimoto, Y.; Kawabata, M.; Harada, M. J. Imaging Sci. and Tech., 1993, 37, 208. Erddalane, Α.; Fouassier, J. P.; Morlet-Savary, F.; Takimoto. Y. J. Polymer Sci., A: Polym. Chem., 1996, 34, 633. Anseth, K. S., Svaldi, D. C., Laurencin, C. T., and Langer, R., Photopolymerization: Fundamentals and Applications, A. B. Scranton, C. N. Bowman, and R. W. Peiffer (eds.), American Chemical Society, Washington DC' 189, 1997˙


26 5.

Kucybala, Ζ., Pietrzak, M., Paczkowski, J., Lindén, L.-Å., and Rabek, J.,

Polymer, 1996, 37, 4585. Oxman, J. D.; Ubel III, Α.; Larson, E. G., Minnesota Mining and Manufacturing Company 1988, U.S. Patent 4,735,632. 7. Padon, K. S.; Scranton, A. B. In Recent Research Developments in Polymer


6.

Science, vol. 3; Pandalai, S. G., Ed.; Transworld Research Network, p 369, 1999. 8. Palazzotto, M . C.; Ubel III, F. A.; Oxman, J. D.; Ali, M . Z. Minnesota Mining and Manufacturing Company 1996, U.S. Patent 5,545,676. 9. Oxman, J. D.; Ubel III, F. Α.; Larson, E. G. Minnesota Mining and Manufacturing Company 1989, U.S. Patent 4,828,583. 10. Palazzotto, M. C.; Ubel III, F. Α.; Oxman, J. D.; Ali, Μ. Ζ. 3M Innovative Properties Company 2000, U.S. Patent 6,017,660. 11. Fouassier, J. P.; Ruhlmann, D.; Takimoto, Y.; Harada, M.; Kawabata, M. J. Polymer Sci. A: Polym. Chem. 1993,31,2245.

12. Fouassier J. P.; Wu, S. K. J. Appl. Polm. Sci., 1992, 44, 1779. 13. Fouassier J. P.; Erddalane, Α.; Morlet-Savary, F. Macromolecule., 1994, 27, 3349. 14. Fouassier J. P.; Morlet-Savary, F.; Yamashita K.; Imahashi S. J. Appl.

Polym. Sci. 1996, 62, 1877. 15. Bi, Y.; and Neckers, D.C. Macromolecules, 1994, 27, 3683. 16. Padon K. S.; Scranton, A. B. J. Polymer 2057. 17. Padon K. S.; Scranton, A. B. J. Polymer 3336. 18. Padon K. S.; Scranton, A. B. J. Polymer 715.

Sci. A: Polym. Chem. 2000, 38, Sci. A: Polym. Chem. 2000, 38, Sci. A: Polym. Chem. 2001, 39,

19. Fouassier J. P.; Chesneau E. Macromol. Chem. 1991, 192, 1307.


Photoinitiated Polymerization - ACS Publications - American Chemical

Recommend Documents