Article Cite This: Macromolecules XXXX, XXX, XXX−XXX
pubs.acs.org/Macromolecules
α‑Ketoesters as Nonaromatic Photoinitiators for Radical Polymerization of (Meth)acrylates Paul Gauss,† Markus Griesser,§ Marica Markovic,‡ Aleksandr Ovsianikov,‡ Georg Gescheidt,§ Patrick Knaack,† and Robert Liska*,† Institute of Applied Synthetic Chemistry and ‡Institute of Materials Science and Technology, TU Wien, Getreidemarkt 9/163, 1060 Vienna, Austria § Institute of Physical and Theoretical Chemistry, TU Graz, Stremayrgasse 9/I, 8010 Graz, Austria Downloaded via UNIV OF EDINBURGH on March 22, 2019 at 18:06:34 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
†
S Supporting Information *
ABSTRACT: Photopolymerization of (meth)acrylate-based formulations has become a widespread method for industry due to the high energy efficiency and low curing times of this technology. Various products from simple coatings to more complex applications such as additive manufacturing technologies are based on this versatile method. Common industrial radical photoinitiators are generally based on aromatic ketones with the benzoyl chromophore as the key constituent. In medical or food packaging applications, residual photoinitiator or photoproducts migrating into the product have to be avoided, particularly for toxicological reasons. In this paper we present a new generation of nonaromatic initiator systems. Besides their good reactivity in (meth)acrylic formulations, they show good bleaching properties and high biocompatibility.
■
INTRODUCTION During the past half-century free radical photopolymerization has become one of the most important technologies for protective and decorative coatings. Advantages are the fast and energy-efficient curing of volatile organic compound (VOC) free formulations. In this technology, UV−vis light is used to activate monomolecular type I or bimolecular type II photoinitiators for the curing of (meth)acrylate-based coatings and inks. Type I photoinitiators undergo homolytic cleavage from the triplet state after excitation by light, while type II systems generate radicals by hydrogen abstraction or electron/proton transfer with a co-initiator from the triplet state.1 The initiation by cleavable type I systems is generally more efficient because type II reactions with co-initiators are diffusion controlled and electron/proton transfer reactions suffer from back electron transfer. Regulations for food packaging and medicinal products are especially strict when it comes to substances migrating through the cured material into the product.2,3 These restrictions result from the fact that low molecular weight photoinitiators and sensitizers together with their photoproducts are not environmentally friendly and may harm the human body. 2 Benzophenones are among the most used photoinitiators for UV curing in the range of 230−350 nm due to the low price and the rather good performance, especially in the case of bathochromic shifted derivatives.4 The International Agency for Research on Cancer (IARC) published a monograph in 2013 about the negative health effects and cancerogenity of benzophenone.5 © XXXX American Chemical Society
There is evidence that benzophenone and its derivatives as well as photoproducts act as estrogenic endocrine disruptors in mammals.6−8 The IACR also found some indications that orally absorbed benzophenones can induce cancer growth in rats. Therefore, benzophenone was categorized as “possibly carcinogenic to humans (Group 2B)”.5 In 2005, Italian authorities discovered traces of the photoinitiator isopropylthioxanthone (ITX) within baby milk cartons, resulting in the seizure of more than 30 million liters of milk which has been contaminated by the curable ink from outside the packaging.9 Furthermore, in a recent US study different photo(co)initiators could be found in 100% of 1000 tested serum samples, showing that these substance are ubiquitous contaminants.10 The benzoyl chromophores of cleavable photoinitiators are generally also problematic as various photoproducts are generated during the curing reaction. Especially volatile and odorous compounds such as benzaldehyde can be problematic at the production site or when it comes to food packaging.2 On the other hand, degradation and recombination products of aromatic initiators are potentially mutagenic or toxic to the human body. Therefore, even safe initiators can lead to substances migrating in the resulting polymer network and becoming hazardous. Therefore, nonaromatic PIs are of high interest for industrial applications. Up to now, camphorquiReceived: December 11, 2018 Revised: March 8, 2019
A
DOI: 10.1021/acs.macromol.8b02640 Macromolecules XXXX, XXX, XXX−XXX
Article
Macromolecules Scheme 1. Concept of Changing Benzoyl Chromophores to Ester Chromophores
Scheme 2. Two Different Type II Photodecomposition Pathways of Aliphatic α-Ketoesters Are Known: Hydrogen Abstraction from Alkyl Chains on the Ketone Side (a) and Hydrogen Abstraction from Alkyls on the Ester Moiety (b)
Namely, aliphatic α-ketoacids may be characterized by high rates of both intersystem crossing and photoreaction.19 Two years later, Nakanish et al. published the photoreduction of a non-enolizable α-ketoester linked to a ginkgolide. As reaction route an intramolecular hydrogen abstraction creating a biradical was suggested.20 The radicals produced upon irradiation of pyruvates in isopropanol were investigated by ESR in 1972 by Samuni et al. They could observe not only the resulting α-hydroxy radical but also a cis/trans isomerization with the ester moiety.21 In 1981, Davidson et al. published a paper on the investigation of the Norrish type II reaction of long chain aliphatic α-ketoesters and -acids. They were able to show that intramolecular hydrogen abstraction leads to a biradical intermediate which further decomposes to alkenes and a pyruvate22 (Scheme 2a). Finally, in the year 2000, Hermann et al. could verify that also the ester moiety can be used for the intramolecular hydrogen abstraction. The biradical further decomposes under CO release of the intermediate ketal to aldehydes or an aldehyde and ketone (Scheme 2b). In both cases radicals are generated which potentially could initiate radical polymerization. The decomposition pathways make a Norrish type II reaction mechanism most probable, so co-initiators might improve the reaction performance by creating more reactive radicals and preventing cleavage reactions that quench the radicals. It has to be mentioned that some α-ketoesters such as ethyl pyruvate are Food and Drug Administration (FDA) approved food additives, which makes them especially suitable for photopolymers on food packaging. The Norrish type II photoproducts of α-ketoesters would consist of α-hydroxy ester derivatives which are considered harmless to the human body.23 α-Ketoglutaric acid is a known water-soluble aliphatic photoinitiator for inverse emulsion polymerization and preparation of hydrogels.24,25 Hitherto ethyl pyruvate, ethyl 3-methyl-2-oxobutanoate, and the cyclic compound 4,4dimethyldihydrofuran-2,3-dione have never been tested as
none is the only aromat-free photoinitiator of significant importance.11 Searching for new chromophores, we came across phenylglyoxylates as well-known as low-yellowing type II initiators (e.g., Irgacure 754).12,13 Compared to benzophenone, one phenyl moiety is replaced with an ester group (Scheme 1). The n−π* transition is in the same range (around 330 nm) as typical for a benzoyl chromophore.14 Interestingly, aliphatic αketoesters also show a comparable absorption pattern. This means that the n−π* transition of the ketone carbonyl is not dependent on aromatic moieties, and aliphatic α-ketoesters could be used as photoinitiators at the same wavelength as the aromatic derivatives. It was hitherto unclear whether the aromatic moiety is necessary at all for an efficient initiation. Hence, we were interested to examine the efficiency of aliphatic ketoesters as photoinitiators for radical polymerization. Up to now, only a few studies have been reported describing their photodecomposition. The first paper was published by Hammond et al. in 1961.15 In these experiments they were able to show that ethyl pyruvate can be photolyzed to CO and acetaldehyde. They also proved that the decomposition can be sensitized by benzophenone and interpreted this as proof for the decomposition of ethyl pyruvate via the triplet state. In this publication, neither was a reaction pathway suggested nor were radicals mentioned.15 Subsequently, Leermakers et al. found similar quantum yields for the decomposition reaction for all alkyl pyruvates and a narrow range of triplet energies around 272 kJ mol−1, which is slightly less than benzophenone (287 kJ mol−1).16 Lower triplet energies make it easier to use sensitizers and reduce the probability of side reactions during the initiation process.17 They could also show that a direct hydrogen transfer reaction is not favored upon irradiation, favoring a biradical mechanism for the decomposition and CO release.18 In 1967, Evans and Leermakers could verify the type II elimination reaction of αketododecanoic acid, decomposing to heptene and pyruvic acid. They also could show that α-ketoacids exhibit a unique behavior in comparison to other classes of keto-compounds. B
DOI: 10.1021/acs.macromol.8b02640 Macromolecules XXXX, XXX, XXX−XXX
Article
Macromolecules
and tested with all initiators. To investigate the novel initiators in a different monomer environment, hexanediol diacrylate (HDDA), a dental dimethacrylate mixture (DDM), and PEG700 diacrylate in 50 wt % water for hydrogels were chosen as monomer systems (see Chart 2).
photoinitiator for bulk photopolymerization of (meth)acrylates. Therefore, linear aliphatic α-ketoesters and a cyclic αketoester were tested as low-cost, biocompatible type II photoinitiators in comparison to industrial standard initiators to find new aliphatic compounds, which can replace aromatic photoinitiators.
■
Chart 2. Used Monomer Systems for Testing the Initiation Efficiency of Initiators
EXPERIMENTAL SECTION
Material and General Methods. Ethyl pyruvate (EP; CAS: 61735-6, TCI), ethyl 3-methyl-2-oxobutyrate (EMOB; CAS: 20201-24-5, Aldrich), 4,4-dimethyldihydrofuran-2,3-dione (DDFD; CAS: 1303104-4, Sigma-Aldrich), methyldiethanolamine (MDEA; CAS: 105-599, Fluka), benzophenone (BP; CAS: 119-61-9, TCI), and 2-hydroxy2-methylpropiophenone (SC73; CAS: 7473-98-5, Lambson) were purchased at the mentioned suppliers and used as received. 4Benzoyl-4′-methyldiphenyl sulfide (BMS; CAS: 83846-85-9) was a gift by Lambson and used as received. Hexandiol diacrylate (HDDA; CAS: 13048-33-4, ρ = 1.01 g mL−1) was purchased from Alfa Aesar and used as received. Urethane dimethacrylate (UDMA; CAS: 7286986-4, ρ = 1.11 g mL−1), and 1,10-decandiol dimethacrylate (D3MA; CAS: 72829-09-5, ρ = 0.963 g mL−1) were kindly obtained from Ivoclar Vivadent and used as received in an equimolar mixture (DMM). Polyethylene glycol diacrylate (PEG700DA; CAS: 2657048-9, Sigma-Aldrich, Mn ∼ 700 Da, ρ = 1.12 g mL−1) was purchased and used as received as a 50 wt % aqueous solution. The water-soluble initiators 2-hydroxy-4′-(2-hydroxyethoxy)-2-methylpropiophenone (IC2959; CAS: 106797-53-9, Ciba) and 3-(4-benzoylphenoxy)-2hydroxy-N,N,N-trimethyl-1-propanaminium chloride (Quantacure BPQ, Rahn) were purchased and used as received. UV−Vis Measurements. For the UV−vis experiments, samples were dissolved in pure acetonitrile and measured in 10 mm quartz cells on a Lambda 750 UV−vis photometer (scanning mode) from 250 to 450 nm at a slit width of 2 nm. The concentration was chosen so that the maximal absorption does not exceed 1 AU (absorption unit) and Lambert−Beer’s law can be applied. Concentrations in acetonitrile for aliphatic compounds EP, EMOB, and DDFD were 1 × 10−2 M, for BP and PGO 1 × 10−3 M, for ITX 1 × 10−4 M, and for BMS 1 × 10−5 M. Photo-DSC Measurements. As the photoinitiation performance and mechanism of ketoesters is unclear, a basic set of ketoesters with different substituents at the α-position was chosen for a broad photoreactivity study by photo-DSC. Benzophenone (BP), 4-benzoyl4′-methyldiphenyl sulfide (BMS), one of the most reactive benzophenone derivatives, and ethyl phenylglyoxylate (PGO) were used as industrially applied type II reference initiators (Chart 1). Generally, ketoesters were expected to be Norrish type II initiators; therefore, a typical type II amine co-initiator (MDEA) was chosen
To ensure equal molar ratio of initiators to monomer for all tested compounds, ethyl pyruvate (EP) as smallest molecule was used as benchmark. For the photo-DSC measurements, 10 mg (1 wt %, 0.086 mmol) of EP was weighed into a brown glass vial with 1000 mg of monomer. All other (co)initiators were weighed into the monomer equimolar to the amount of 1 wt % ethyl pyruvate. All samples were thoroughly mixed on a vortex mixer at highest power for 10 s. The photo-DSC measurements were conducted on a Netzsch DSC 204 F1 with autosampler at 25 °C under a nitrogen atmosphere. Photoreactivity of the monomers was tested by weighing accurately 10.5 ± 0.5 mg of the sample into an aluminum DSC pan with a pipet, which was subsequently placed in the DSC chamber after closing it with a glass lid. The sample chamber was purged with a N2 flow (∼20 mL min−1) for 4 min. Afterward, the samples were irradiated with filtered UV-light (320−500 nm) from an Exfo OmniCure S2000 broadband Hg lamp with a light intensity of 1 W cm−2 at the end of the light guide. The samples were irradiated twice for 300 s, and the heat flow was recorded as a function of time. After subtraction of the second irradiation period for baseline correction, a calorigram was obtained. The maximal rate of polymerization was calculated from the maximum heat flow, density, and theoretical heat of polymerization (Table S2) by eq 1.26
Chart 1. Different α-Ketoesters and Industrial Type II Photoinitiators with Tertiary Amine Coinitiator MDEA
RP =
q × δ × 1000 ΔH0
(1) −1
where RP is the maximal rate of polymerization in mM s , q is the maximal heat flow mJ s−1 mg−1, δ is the density of the sample in g L−1 provided by the supplier, and ΔH0 is the theoretical heat of polymerization in kJ mol−1. The time to attain 95% conversion (t95%) is defined as the time required for generated heat to equal 95% of ΔHp (peak area). The double-bond conversion DBC was calculated from ΔHp (peak area) and the theoretical heat of polymerization according to eq 2.26−28 The theoretical heat of polymerization for acrylates (80.5 kJ mol−1) was found in the literature29 and used for diacrylate HDDA. For the 50 wt % diacrylate hydrogel formulation the weight factor f = 0.5. For the methacrylate mixture DMM ΔH0,DMM = 58.5 kJ mol−1 per reactive group, the known value from the literature, was used.27 DBC = 100 × C
ΔHP ΔH0 × f × n
(2) DOI: 10.1021/acs.macromol.8b02640 Macromolecules XXXX, XXX, XXX−XXX
Article
Macromolecules where DBC is the double bond conversion in %, ΔHP is the measured heat of polymerization of the formulation in kJ mol−1, ΔH0is the theoretical heat of polymerization of the formulation in kJ mol−1, f is the factor for wt % monomer = 1 for pure monomer and 0.5 for 50 wt % aqueous formulation, and n is the number of polymerizable groups. To ensure the stability of the measurement, mixtures with HDDA/DMM and 1 equiv of Speedcure 73 (SC73) to 1 wt % of ethyl pyruvate were measured as reference as first, middle, and last sample with every series. All mixtures were measured as triplicates. Long Wavelength Experiments. Mixtures of HDDA with ITX and co-initiator MDEA as well as DDFD were prepared and measured as mentioned above. In this case, a 400 nm LED light source was used at a power of 0.38 W cm−2 at the exit of the light guide to irradiate the DSC samples. Preparation of Hydrogel Samples. A stock solution of 50 wt % polyethylene glycol diacrylate (Mn 700) in water was prepared. Into 1000 mg of the stock solution 1 wt % of ethyl pyruvate were weighed in. All other (co)initiators were weighed into the formulations equimolar to the amount of ethyl pyruvate to ensure same molar concentrations for all measurements (Chart 3).
(Synergy BioTek; excitation 560 nm, emission 590 nm). After correction for background fluorescence, the results of the cells exposed to different concentrations of photoinitiators were compared to each other and to the controls (nonstimulated cells) using the One Way ANOVA Dunnett’s Multiple Comparisons Test.30 It was assumed that metabolic activity of the 24 h control not exposed to photoinitiators is 100%. Statistical evaluation of data was performed using the software package IBM SPSS Statistic 25.0 and Excel 2016 (Microsoft Office). NMR Photolysis Studies. 1H NMR spectra were recorded on a 200 MHz Bruker AVANCE DPX spectrometer. Irradiation of the solutions in CD3CN was performed using 365 nm LEDs (100 mW), and spectra were recorded after 0, 30, and 60 min.
Chart 3. Commercial Initiators for Hydrogels
Table 1. Maxima of the n−π* Transition and Extinction Coefficient of the Tested Initiators
■
RESULTS AND DISCUSSION UV−Vis Absorption. The UV−vis spectra of the initiators were recorded in acetonitrile to determine the maxima of the n−π* transition and the corresponding extinction coefficient ε. As can be seen in Table 1 and Figure 1, the initiators exhibit
initiator BP BMS PGO ITX EP EMOB DDFD
concn (M) 1 1 1 1 1 1 1
× × × × × × ×
−3
10 10−5 10−3 10−4 10−2 10−2 10−2
λmax (nm)
ε (M−1 cm−1)
338 312 343 383 330 331 375
140 21600 53 2792 17 21 28
large differences in their extinction coefficients. The benzophenone derivative BMS shows an extinction at least 2 orders of magnitude higher than that of the other initiators. This can be explained by the overlap of the n−π* transition with the red-shifted π−π* transition due to the donor thioether substituent. Interestingly, the aliphatic ketoesters show an ε of the same magnitude as the aromatic ketoester PGO. Photo-DSC Studies. To investigate the photoreactivity of the different monomer/initiator systems, photodifferential scanning calorimetry was chosen as a powerful tool. As there is no literature available about the photoreactivity of aliphatic ketoesters, three simple monomer systems were selected (Chart 2). Ethyl pyruvate (EP) was used as benchmark at a concentration of 1 wt %; all other (co)initiators were used equimolar to 1 wt % EP providing identical molar concentrations of the initiators in all experiments. BP and BMS were measured with MDEA as co-initiator. The strongly basic amine MDEA was not compatible with the ketoesters, as it lead to fast decomposition of the initiators, especially in aqueous formulations. Therefore, the ketoesters were only measured without co-initiator. Detailed results can be found in the Supporting Information. Acrylates. The maximal rate of polymerization (RP) of mixtures with acrylate HDDA is summarized in Figure 2A. As benzophenones BP and BMS are classical Norrish type II initiators, mixtures with co-initiator MDEA were prepared and tested. The phenyl glyoxylate PGO is a known unimolecular type II initiator and was also used without co-initiator like all other ketoesters.31,32 As expected, BMS with the amine MDEA shows a much higher maximal rate of polymerization (RP) than BP.
All samples were thoroughly mixed on a vortex mixer at the highest power for 10 s. Afterward, DSC measurements were conducted with filtered UV light (320−500 nm) from an Exfo OmniCure S2000 broadband Hg lamp with a light intensity of 1 W cm−2 at the end of the light guide, as already mentioned. UV Aging. A silicon mold (1 × 2 × 0.2 cm3) was filled with the same mixtures as used for DSC measurements and afterward irradiated in an Intelliray 600 UV flood curing system. The samples were exposed to the UV light for 120 min, and afterward the color was documented by digital photography. Cytotoxicity Tests. Mouse fibroblast cell line L929 was purchased from Sigma-Aldrich and cultured in Dulbecco’s Modified Eagle’s medium (DMEM, Sigma-Aldrich) supplemented with 10% fetal bovine serum (Sigma) and 1% of 10000 units of penicillin and 10 mg mL−1 streptomycin (Sigma-Aldrich). Cells were cultivated in an incubator in a humid atmosphere with 5% carbon dioxide at 37 °C. The medium was refreshed every second day. To evaluate the cytocompatibility of photoinitiators, PrestoBlue cell viability reagent (Thermo Fisher) was used. For this test ten 96well plates were seeded with 5000 cells per well and left in the incubator overnight for cells to attach. The next day, the cells were incubated with 100 μL of different dilutions of PI in DMEM, each plate in duplicate: one to evaluate the cytotoxicity of substance only and another to evaluate the cytotoxicity after exposure to UV for 10 min in a Boekel Scientific UV cross-linker at 365 nm, equivalent to 1 J of accumulated energy. All plates had wells with nontreated cells as a control samples; every concentration had at least six repetitions. After 24 h incubation period with photoinitiators, the culture medium was exchanged twice to remove residues of photoinitiator, and cell viability was evaluated. Resazurin-based reagent PrestoBlue was diluted 1:10 with medium, and 100 μL was applied per well and incubated for 1 h. Because of the reducing environment of viable cells, this reagent is transformed and turns red, while becoming highly fluorescent. The fluorescence was then measured with a plate reader D
DOI: 10.1021/acs.macromol.8b02640 Macromolecules XXXX, XXX, XXX−XXX
Article
Macromolecules
Figure 1. UV−vis spectra of diaryl ketones BMS and ITX and ketoesters in acetonitrile.
Figure 2. Maximal rate of polymerization RP of industrial initiators (□) and α-ketoesters (■), with co-initiator MDEA (orange □) and without coinitiator (green ■) in (A) HDDA mixtures at 1 W cm−2 (320−500 nm) with industrial type II initiators BP and BMS with MDEA as co-initiator and ketoester formulations containing 1 equiv initiator relative to 1 wt % EP. (B) DMM methacrylate mixtures at 1 W cm−2 (320−500 nm) with industrial type II initiators BP and BMS and ketoesters containing 1 equiv of initiator relative to 1 wt % EP. (C) HDDA mixtures at 0.38 W cm−2 (400 nm) with industrial type II initiator ITX and ketoester formulation DDFD containing 1 equiv of initiator relative to 1 wt % EP. (D) PEG700 diacrylate in 50% H2O at 1 W cm−2 (320−500 nm) with industrial initiators IC2959 and BPQ/MDEA and ketoesters EP and DDFD containing 1 equiv of (co)initiator relative to 1 wt % EP.
As can be seen in Figure 2B, the maximal rate of polymerization of mixtures of ketoesters without co-initiator (green) perform at the same level as the industrial initiators with MDEA. This least sterically hindered compound ethyl pyruvate EP shows significantly higher reactivity than BMS and can even reach higher maximal rates of polymerization than the phenyl glyoxylate PGO. This shows that aliphatic ketoesters can readily compete with industrial standards even in less reactive monomers. BP with MDEA as co-initiator showed the lowest conversion with only 48 ±
