Acrylate nanolatex via self-initiated photopolymerization

Acrylate Nanolatex via Self-Initiated Photopolymerization Florent Jasinski,1 Emeline Lobry,1 L�ena€ıg Lefevre,1 Abraham Chemtob,1 C�eline Croutxe-Barghorn,1 Xavier Allonas,1 Adrien Criqui2 1Laboratory of Photochemistry and Macromolecular Engineering, ENSCMu, University of Haute-Alsace, 3 bis rue, Alfred Werner 68093 Mulhouse Cedex, France 2M€ader Research - M€ADER GROUP, 130 rue de la Mer Rouge, 68200 Mulhouse, France Correspondence to: A. Chemtob (E-mail: [email protected]) Received 3 February 2014; accepted 22 March 2014; published online 00 Month 2014 DOI: 10.1002/pola.27190 ABSTRACT: The use of UV light to initiate emulsion polymeriza- tion processes is generally overlooked, whilst extensive litera- ture exists on photocuring of monomer films. In this study, the unique potential of UV light to produce at ambient temperature polyacrylate latexes without initiator was exploited. Although radical initiators are utilized at low concentration, their cost, toxicity, and odor provide incentives for finding alternatives. Starting with concentrated (30 wt %) and low scattering acry- late miniemulsions (droplet diameter 100 �C), most radical monomers when exhaustively puri- fied cannot undergo a purely thermal self-initiated polymer- ization.23,24 As a result, thermal initiators or redox systems are indispensable ingredients to bring about a radical chain polymerization in aqueous dispersed media where reaction temperatures are generally limited.25 In contrast, a much broader range of monomers are able to self-initiate at ambient temperature under UV exposure. Several thiol-ene26 and elec- tron donor–acceptor systems27 have demonstrated their self- photoinitiation ability as well as some specific vinyl28 or bro- minated acrylates.29 Recently, even simple alkyl acrylate and methacrylate systems have proven to polymerize without pho- toinitiator through deep UV irradiation provided by excimer lamp30–33 (172 or 222 nm) or even medium-pressure mercury Additional Supporting Information may be found in the online version of this article. VC 2014 Wiley Periodicals, Inc. WWW.MATERIALSVIEWS.COM JOURNAL OF POLYMER SCIENCE, PART A: POLYMER CHEMISTRY 2014, 00, 000–000 1 JOURNAL OF POLYMER SCIENCE WWW.POLYMERCHEMISTRY.ORG ARTICLE arc lamps (250–600 nm).34,35 In the continuity of these latter works focused on UV curing; we show that miniemulsion pho- topolymerization can be used as an efficient method to making initiatorless water-based acrylate dispersions. In polymer industry, acrylate latexes are currently employed at large scale in diversified customer markets ranging from coatings to adhesives.36 In addition to reducing cost and odor, the elimina- tion of initiator could open up new opportunities for medical, food, or microelectronic applications in which nontoxic poly- mer materials are in great demand.37 A final polymer film without initiator residues may exhibit a decreased tendency to yellowing and sunlight degradation, therefore, providing the benefits of a durable material suitable for outdoor applica- tions, which is rare in photopolymer materials.37 With regard to process, initiator-less monomer miniemulsions are likely to have a prolonged colloidal stability. In addition, since light is attenuated only by droplet scattering and absorption of mono- mers rather than photoinitiator molecules, enhanced polymer- ization depth may be achieved when the extinction coefficients of monomers are not too high. Following an initial feasibility study,19 we investigate exhaus- tively in this paper all the kinetic, colloidal and mechanistic aspects of initiator-free acrylate photopolymerizations using high-solid content miniemulsions of 30 wt % required for commercial applications. The dependence on initiating wave- length, droplet size, optical path, and irradiance has been reviewed thoroughly as well as the critical issue of monomer miniemulsion stability. Photopolymerizations have been per- formed in small volume spectroscopic cell, and then scaled- up in an annular photoreactor. In a last part, the mechanism of initiation in spontaneous photoinduced polymerization of acrylate and methacrylate is discussed. To date, there has been a lack of evidence to draw any distinct conclusion about the acrylate self-photoinitiation mechanism. EXPERIMENTAL Materials All miniemulsions were prepared with distilled water. Technical grade monomers, butyl acrylate (BA), methyl methacrylate (MMA), and acrylic acid (AA), n-butyl methacrylate (BMA), methyl acrylate (MA) and ethyl acrylate (EA) were supplied by Aldrich and used without further purification. Sodium dodecyl sulfate (SDS, Aldrich) was used as received. Stearyl acrylate (SA, Aldrich) was added in the formulation as a reactive costabilizer. 2-Methyl-1-[4-(methylthio)phenyl]-2-(4-morpholinyl)21-prop- anone (I 907) from BASF was used as an oil-soluble photoinitia- tor, benzophenone (BP, Aldrich), and thioxanthone (TX, Aldrich) as photosensitizers. Thermal polymerizations were performed with potassium persulfate (KPS, Aldrich). In electron spin reso- nance experiments, tert-butylbenzene (TBB, Aldrich) was used as inert solvent and 5,50-dimethyl-1-pyrroline-N-oxide (DMPO, Aldrich) as spin trap.38 Preparation of Monomer Miniemulsions An organic phase was first prepared by adding SA (4 wt % with respect to the monomer phase) to the monomer acry- late mixture (BA/MMA/AA, 49.5/49.5/1 wt %). The aqueous phase was prepared separately by dissolving SDS in distilled water (3.5 to 0.15 wt % with respect to the monomer phase). The weight concentration of the monomer phase labeled as Cmonomer was kept constant at 30 wt % in all experiments. Both phases were mixed for 10 min using a magnetic stirrer at 600 rpm. The coarse emulsion was then homogenized for 5 min with the aid of a Branson Sonifier 450 (450 W/L) at 90% amplitude, while maintaining the agitation. Initiatorless Miniemulsion Photopolymerization in Spectroscopic Quartz Cell In a typical procedure, the photopolymerization of the mono- mer in the miniemulsion was carried out in a capped quartz rectangular cell (1 mm thick, 340 mL volume) without nitro- gen bubbling and stirring. Irradiation was applied with the polychromatic light of a medium-pressure Hg–Xe arc lamp (Hamamatsu L8252, 200 W) coupled to a flexible light-guide. A picture of the illumination set-up is given in Supporting Information Figure S1 in the Additional Supporting Informa- tion (ASI). The lamp is backed by a semi-elliptical mirror or reflector to focus radiation and minimize irradiance loss. In this study, a 254 or 365 nm reflector was used, each one enhancing the reflection of the mentioned wavelength. The end of the optical guide was placed at a distance of 4.2 cm from the sample and directed at an incident angle of 90� . In the spectral region below 300 nm in which acrylate mono- mers absorb, the light irradiance was respectively 150 mW cm22 (254 nm reflector) and 100 mW cm22 (365 nm reflec- tor). This irradiation set-up was used for the kinetic analysis of the polymerization by real-time Fourier transform near IR spectroscopy (RT-FTNIR) described in the characterization section. For comparison, thermally induced polymerizations were also performed by heating the same spectroscopic cell (70 �C) containing the miniemulsion inside an environmental chamber. After photopolymerization, the resulting latex was precipitated in methanol. After filtration and washing, the solid polymer was then dissolved in filtered and distilled tet- rahydrofuran for molecular weight analysis. Initiatorless Miniemulsion Photopolymerization in an Annular UV Reactor The annular photoreactor (UV-Consulting Peschl) shown in Figure 1 is composed of three parts: first, a standard medium-pressure Hg arc lamp (Heraeus Noblelight TQ 150, arc length: 4.4 cm) emits a series of rays from 250 to 600 nm (emission spectrum in Supporting Information Figure S2 of ASI). This lamp is housed in a fused quartz sleeve offering an excellence transmittance in the UV region down to ca. 200 nm, which is essential to monomer excitation and self- initiation. Because of the heat liberation during lamp proc- essing, an external cooling jacket surrounds the sleeve vessel in order to hold the photoreactor contents at a temperature between 20 and 25 �C throughout the polymerization reac- tion. Third, a borosilicate cylindrical section (outer annulus) is then installed around the sleeve to accommodate and irra- diate 300 mL of monomer miniemulsion when the reactor is ARTICLE WWW.POLYMERCHEMISTRY.ORG JOURNAL OF POLYMER SCIENCE 2 JOURNAL OF POLYMER SCIENCE, PART A: POLYMER CHEMISTRY 2014, 00, 000–000 full. The distance between the two annular sections (external sleeve wall and inner reactor wall) is 9 mm and defines the optical path length. Magnetic stirring of the miniemulsion is maintained throughout the polymerization. Prior to the UV exposure, the monomer miniemulsion was degassed for 30 min through nitrogen bubbling, which was also maintained during the irradiation process. Samples were drawn at differ- ent time intervals in order to determine the polymerization kinetics (FTNIR spectroscopy) and the evolution of molecu- lar weights. Polymerization was stopped when a complete conversion was achieved or after a maximum exposure time of 6 h. Methods Colloidal Characterization The miniemulsion stability was characterized using a Turbis- can instrument (Formulaction). In a typical measurement, 20 mL of freshly prepared miniemulsion was placed in vial and placed in the instrument. The temporal evolution of the backscattered light irradiance normalized with respect to a non-absorbing standard reflector was assessed during 4 h. The evolution of the backscattered light (at 45� from the incident beam) was determined by scanning the entire vol- ume of the miniemulsion from the bottom to the cap of the cell.39 Measurements were typically carried out at 25 �C at the middle of the miniemulsion (�20 mm). Droplet and par- ticles diameters, respectively labeled as Dd and Dp, were determined through dynamic light scattering (DLS) with a Zetasizer Nano ZS (Malvern Instrument). Typically, the monomer miniemulsion or the resultant latex was diluted 125 times in filtered and distilled water before measurement. Polymer Characterization and Reaction Kinetics Acrylate conversion was followed in situ by RT-FTNIR. In these rapid scan experiments (temporal resolution: 0.5 s, spectral resolution: 4 cm21), a NIR probe beam and a UV exciting beam irradiated simultaneously the spectroscopic cell containing the miniemulsion (see Supporting Information Figure S1 in ASI). The NIR region (k5 970–1940 nm) is well suited to thick samples (1000 mm) and can accommodate the high water concentrations of monomer miniemulsions without detector saturation. In this spectral range, the acry- late monomers exhibit a band at 6170 cm21 which is assumed to be a combination of two C–H stretching bands.40 The band is isolated enough from the other vibrational over- tones of water so that it can be used for quantitative pur- poses. Thus, polymerization kinetics was followed in situ using RT-FTIR by calculating the integrated absorbance of this band and monitoring its decrease during irradiation. The limited signal to noise ratio, especially at high conver- sion, provided motivation to fit the conversion–time plots using a sigmoidal function described in reference.41 Molecu- lar weights were determined by gel permeation chromatog- raphy (GPC). The GPC column was calibrated with polystyrene standards, implying that all the molecular weight values (Mn) are considered as polystyrene equivalent. Electron Spin Resonance (ESR) Spin Trapping Experiments ESR spectra were measured with an X-band spectrometer (Miniscope 200 spectrometer, Magnettech). The acquisition was performed in 4 mm diameter capillary with 5 accumu- lations and a 400 gain. Sample solutions contained 9 3 1023 mol L21 of DMPO and 0.445 mol L21 of an acrylate monomer. After 15 min bubbling with argon, the spectrum of the solutions were recorded following 100 s of UV irradi- ation with the same Hg–Xe lamp (I250–3005 100 mW cm 22) used for the photopolymerization in spectrophotometric cuv- ettes. The main issue arising from the acrylate and methac- rylate study in ESR is the very short lifetime of the generated radicals ( RESULTS AND DISCUSSION Monomer Miniemulsion Metastability Four initiatorless acrylate miniemulsions (MMA/BA/AA/SA) with diameters ranging from 40 to 115 nm were prepared using different SDS concentrations, but at a constant organic phase content (30 wt %) and costabilizer concentration (SA, 4% wt/wtmonomer). As shown in Figure 2, the average drop- let diameter declines with higher surfactant concentrations due to the increase in surfactant interfacial area and decrease in interfacial tension. The generation of nanosized droplets (Dd< 100 nm) is important to reduce UV light attenuation caused by scattering, but very small droplet sizes can result also in stability problems.42 Although their poor deformability makes them relatively insensitive to coales- cence,43 the main source of instability in nanoemulsions is generally Ostwald ripening (diffusional degradation) arising from the difference in Laplace pressure (chemical potentials) between droplets of different sizes.44,45 The stability of these monomer miniemulsions (40, 75, 90, and 115 nm) was examined by monitoring the temporal evo- lution of the diffuse reflectance R (fraction of incident light reflected back at the sample interface). Figure 3 is a plot representing R in the sample middle and at ambient temper- ature as a function of the ageing time (4 h). As expected, we clearly see that miniemulsions having the largest droplets exhibit greater values of R due to a stronger scattering effi- ciency. But more importantly, all monomer miniemulsions having an average diameter larger than 75 nm were stable within the first 4 h of analysis, as suggested by a relatively steady R during this period. As expected, such metastability is strongly dependent on the presence of costabilizer mole- cules in the monomer phase. As an evidence, the removal of SA from the initially stable 90 miniemulsion led to a rapid destabilization (see � in Fig. 3 and also Supporting Informa- tion Figure S346). The case of the smallest miniemulsion (Dd5 40 nm) prepared with the highest concentration in SDS (3.5% wt/wtmonomer) is particular since slight signs of destabilization were observed even in the presence of SA. Two factors are responsible for the enhanced Ostwald Ripen- ing. (i) The size distribution is broader in this case as dis- played in Figure 2, whereas other miniemulsions appear more monodisperse (note that strictly monodisperse samples will not exhibit diffusional degradation). The increase in pol- ydispersity with decreasing miniemulsion size has been already reported in the literature47 and related to micelles formation or the technical limitations of a sonicator-induced emulsification. (ii) Additionally, the high concentration in SDS may drive the migration of monomer molecules from the smaller droplets, across the aqueous phase, into the larger droplets.48 Indeed, if too much surfactant is used, the excess present in the aqueous continuous phase can enhance the rate of monomer transfer between the droplets. Proof of Acrylate Self-Photoinitiation in Miniemulsion Figure 4 compares the absorbance spectra of monomer mini- emulsions (Dd5 40, 75, and 115 nm) with that of an equiva- lent solution prepared in acetonitrile at similar monomer concentration (30 wt %). As expected, monomer miniemul- sions had higher absorbance values which tend to increase with droplet size owing to an enhanced scattering. The solu- tion system reveals that the acrylates absorption range starts below 280–300 nm, allowing them to match at least partially the emission spectrum of the Hg–Xe arc lamp plotted on the same graphic. Although the overlap concerns only the FIGURE 2 Droplet size distribution obtained by DLS for acry- late monomer miniemulsions including different weight con- centrations in SDS: 3.5% (Dd540 nm, W), 0.75% (Dd5 75 nm, �), 0.5% (Dd5 90 nm, �), 0.25% (Dd5 115 nm, �). Cmonomer530 wt %, organic phase: MMA/BA/AA/SA. FIGURE 3 Monomer miniemulsions stability assessed by the reflectance measurements during ageing time in the middle of the sample vial (Turbiscan data). The miniemulsions comprise different weight concentrations in SDS: 3.5% (Dd5 40 nm, W), 0.75% (Dd5 75 nm, �), 0.5% (Dd590 nm, �) and 0.25% (Dd5 115 nm, �). Full symbols are for miniemulsion containing SA (4% wt/wtmonomer) and open symbols are costabilizer-free systems. All droplet size data refer to “stable” miniemulsions containing SA. Cmonomer530 wt %, organic phase: MMA/BA/ AA/(SA). ARTICLE WWW.POLYMERCHEMISTRY.ORG JOURNAL OF POLYMER SCIENCE 4 JOURNAL OF POLYMER SCIENCE, PART A: POLYMER CHEMISTRY 2014, 00, 000–000 shortest wavelengths (250–300 nm), there are strong incen- tives to utilize mercury arc lamps, which are the workhorse of the photopolymerization in industry, and existing in a variety of size, geometry, and power at limited cost. Figure 5 shows graphically the acrylate conversion of a 40 nm miniemulsion without initiator as a function of irradia- tion time under different irradiation conditions. All kinetic curves were obtained by RT-FTNIR spectroscopy allowing the real-time acquisition of NIR spectra during the irradia- tion process (see characterization section). A complete con- sumption of monomer was achieved in less than 6 min with the Hg–Xe lamp fitted with a 365 nm reflector providing an irradiance of 100 mW cm22 in the 250–300 nm range. As a proof of concept demonstrating the self-initiation ability of acrylates, the addition of a cutoff filter blocking the wave- lengths below 300 nm impedes the course of the polymer- ization. Because of differences in light absorption, the polymerization rate is also predicted to be highly dependent on the wavelength of the irradiating light. As seen in Figure 4, acrylates absorb light much more readily at lower wave- lengths. Accordingly, a 254 nm reflector emitting a higher irradiance of 150 mW cm22 (250–300 nm) induced a greater radical generation rate, thereby increasing polymerization rates. Hence, the same acrylate miniemulsion reached a full conversion in 4 min instead of 6 min. Only faster kinetics was achieved when a water-insoluble photoinitiator (a-ami- noacetophenone radical) was added to the monomer organic phase. In this case, initiating radicals originate both from acrylates excitation and initiator a-cleavage. Noteworthy is that temperature does not increase more than 10 �C above ambient temperature during irradiation, thus indicating that the polymerization is primarily photoinitiated. Such result is confirmed in Figure 6 by the absence of poly- merization upon heating the miniemulsion at 70 �C without irradiation. In contrast, the addition of KPS, a conventional water-soluble peroxide initiator, triggers a thermal initiated polymerization at 70 �C. However, the polymerization is much slower in this case compared to a photoinitiated path- way. It is assumed that faster initiation rates are promoted in acrylate self-initiated polymerization compared to KPS exhibiting a slow decomposition rate at 70 �C (half life of 10 h at this T). FIGURE 4 Emission spectra of the Hg–Xe arc lamp fitted with a 365 nm reflector (solid) or a 254 nm reflector (dot). In this latter case, the light output irradiance below 300 nm is increased. Absorption spectra of acrylate monomer miniemulsions with four different droplet sizes: 40 nm (W), 75 nm (�), 90 nm (�), and 115 nm (�). For comparison, the absorption spectrum of an acrylate monomer solution in acetonitrile at similar concen- tration is provided (*). Cmonomer5 30 wt %, organic phase: MMA/BA/AA/SA, 1 mm thick quartz cuvette. FIGURE 5 Conversion–time curves of an initiatorless miniemul- sion (Dd5 40 nm) irradiated by the medium-pressure Hg–Xe lamp fitted with a 365 nm reflector (1), a 254 nm reflector (3) or a 300 nm cutoff filter (�). For comparison, the photopolyme- rization kinetics (365 nm reflector) of the same miniemulsion containing I907 as oil-soluble PI (2% wt/wtmonomer) is also dis- played (*). Cmonomer: 30 wt %, 1 mm thick quartz cuvette. FIGURE 6 Acrylate conversion evolution for acrylate miniemul- sions heated at 70 �C without initiator (W) and with water- soluble KPS initiator (�). The reference PI-free miniemulsions irradiated using a Hg–Xe lamp (365 nm reflector) is also repro- duced for comparison (3). Cmonomer5 30 wt %, [Initiator]5 2% wt/wtmonomer, 1 mm thick quartz cuvette. JOURNAL OF POLYMER SCIENCE WWW.POLYMERCHEMISTRY.ORG ARTICLE WWW.MATERIALSVIEWS.COM JOURNAL OF POLYMER SCIENCE, PART A: POLYMER CHEMISTRY 2014, 00, 000–000 5 Effect of Different Experimental Parameters on Initiatorless Acrylate Photopolymerization Effect of Droplet Size Figure 7 illustrates the strong dependency between polymer- ization rates and miniemulsion droplet size. The reaction rates are fairly enhanced upon decreasing the average drop- let diameter by the following factors: (i) Light penetration depth through the miniemulsion is improved when decreasing droplet size (Fig. 4). A larger vol- ume of miniemulsion is thus illuminated, enabling a greater fraction of monomer droplets to be initiated. Optically, both monomer absorption and droplet scattering impact the pene- tration of light inside the reaction medium. Recent spectro- photometric studies on acrylate miniemulsion have proved that change in droplet diameter, in the range 40–300 nm, affects scattering, but not absorption.22 (ii) The decrease in droplet size is known to favor a radical compartmentalization effect.49 Very small droplet sizes ( a decreased number of initiating radicals promoted at low irradiance caused an increase in molecular weight. Miniemulsion Photopolymerization without Photoinitiator in a Labscale Photoreactor Self-initiated miniemulsion photopolymerization was attempted in an annular photoreactor. Such immersion-type photochemical reactor is one of the most common reactor configuration used at laboratory scale, and has been com- monly involved in many photolysis (water purification),52 photocatalysis,53 and photoinduced organic reactions.54 How- ever, there have been fewer examples of polymerizations performed in this vessel or even in other photochemical reactors.55–59 As displayed in Figure 1, the lamp in this pho- toreactor is centered parallel to the axis of the reactor vessel and separated from the miniemulsion by a cooling tube. To make monomer self-initiation effective, a medium-pressure Hg arc lamp is employed (emission spectrum in Supporting Information Fig. S2) and housed in a quartz cooling tube with a transmittance extending down to 200 nm. Obviously, such scaling-up imposes different irradiation conditions in comparison with our initial model system based on the local- ized irradiation of miniemulsions contained in spectroscopic cells. The photoreactor set-up is notably characterized by a lower irradiance measured at the cooling tube surface (5 mW cm22 vs. 100 mW cm22 in the 250–300 nm spectral range), a larger optical path (9000 mm vs. 1000 mm), and a much greater irradiated volume (300 vs. 0.34 mL). TABLE 1 Effect of Droplet Size (Dd), Irradiance (I250–300 nm), and Optical Path (e) on Colloidal Properties (Dp, Np/Nd), Reactions Kinetics (Conv, Rp max) and Number Average Molecular Weight (Mn) SDS concentration (% wt/wtmonomer) I250–300 nm (mW cm22) e (mm) Dd (nm) Dp (nm) Np/Nd Conv after 16 min (%) Rp max (mol L21 s21) Mn (310 3 g mol21) 3.5 100 1 40 60 0.30 100 1.28 33.1 1.5 100 1 55 90 0.20 100 0.91 66.8 0.75 100 1 75 95 0.43 100 0.57 107 0.5 100 1 90 110 0.42 80 0.44 96.9 0.25 100 1 110 (150) – 64 0.34 51.8 0.15 100 1 140 (185) – 44 0.22 57.0 0.05 100 1 210 (265) – 16 0.14 45.1 3.5 100 0.5 40 55 0.33 100 2.29 24.4 3.5 100 0.1 40 55 0.32 100 4.34 16.1 3.5 75 1 40 55 0.33 100 0.51 47.9 3.5 50 1 40 60 0.21 100 0.41 58.9 3.5 25 1 40 55 0.36 100 0.30 69.1 FIGURE 8 Acrylate conversion as a function of irradiation time at different cell thicknesses: 100 mm (W), 500 mm (�), and 1000 mm (�). Cmonomer5 30 wt %, I250–300 (365 nm reflector)5 100 mW cm22, Dd5 40 nm. FIGURE 9 Acrylate conversion as a function of irradiation time for different irradiance I250–300 nm (irradiance is given in the 250–300 nm range where acrylate absorption is effective): 100 mW cm22 (W), 75 mW cm22 (�), 50 mW cm22 (�), and 25 mW cm22 (�), Cmonomer5 30% w/wmonomer, 1 mm thick quartz cuv- ette, Dd5 40 nm. JOURNAL OF POLYMER SCIENCE WWW.POLYMERCHEMISTRY.ORG ARTICLE WWW.MATERIALSVIEWS.COM JOURNAL OF POLYMER SCIENCE, PART A: POLYMER CHEMISTRY 2014, 00, 000–000 7 Our main efforts have focused on investigating the impact of droplet size on polymerization kinetics as the first experi- ments have revealed the prominent role of this parameter. Despite major differences in experimental conditions, the conversion–time plots shown in Figure 10 revealed strong analogies with the polymerization behavior found in irradi- ated cuvettes. Similarly, it is emphasized that small enough miniemulsion droplets are essential to achieve high reaction rates and full conversion. In addition, Table 2 indicated again a lower number of particles in comparison with the initial number of droplets (Np/Nd< 1), translating a limited droplet nucleation efficiency. However, the most salient event occurs with the smallest monomer miniemulsion (Dd5 40 nm) undergoing a very high increase of the miniemulsion viscos- ity during the polymerization. The high density of particles in this case led to a short particle-particle distance and strong particle interactions. This caused a substantial decel- eration of the polymerization at the end of the reaction as well as filming problems on the cooling and immersion tubes indicative that diffusion was insufficient inside the reactor (see images in Supporting Information Fig. S4). In our monomer miniemulsions, the ionic strength is low because no ions other than those of the SDS are present in the aqueous phase. Therefore, the electrical double layer sur- rounding the particles is relatively thick and the electrostatic interactions are strong in a densely packed system compris- ing a high number of droplets. Consequently, the problem was overcome by adding a small concentration of electrolyte ([NaHCO3]5 0.03 mol L 21) to the aqueous phase.60 The addition of salt compresses the double layer and lowers the zeta potential; thereby, reducing the electrostatic repulsion. However, it induces further coalescence after the sonication process, and thus increases the average droplet size. In pres- ence of salt, the miniemulsion prepared with 3.5 wt % SDS showed a diameter increasing from 40 to 60 nm, without detrimental effects on its relative stability (Supporting Infor- mation Fig. S5). Remarkably, no filming was now reported and a higher polymerization rate was even obtained because of the higher diffusion capacity in a low viscous medium. As seen in Figure 10, a complete conversion was reached after 1 h of irradiation, whereas the same salt-free miniemulsion required almost 2 h. Self-Initiation Mechanism of Acrylate and Methacrylate Monomer Initiating Radicals When photoinitiators are not utilized, the mechanism by which initiation of acrylate polymerization occurs has not been clearly identified. Attempts were made to identify the transient species through laser flash photolysis supple- mented by computational analysis.30,61 Under excitation at 222 nm, butyl acrylate was found to exhibit a transient absorption decaying within 10 ms in acetonitrile solution. This transient was assigned to a triplet state. This was con- firmed by density functional theory (DFT) computations which further indicated that the spin density was highly localized on the double bond. Photolysis products were stud- ied using chromatography techniques showing that a- cleavage can occur from the singlet state and from unrelaxed triplet state, leading to the formation of initiating radicals.61 Regarding the bimolecular reaction with a second acrylate moiety, it was shown to occur preferentially through an addi- tion reaction rather than H-abstraction. The rate constant of addition being 7 3 108 mol L21 s21, this reaction was thought to be predominant.30 According to the authors, a 1,4-biradical in the triplet state was formed. Further infor- mation on the evolution of this triplet biradical can be found FIGURE 10 Effect of the initial droplet size on polymerization kinetics in a self-induced miniemulsion photopolymerization carried out in a photoreactor. A range of droplet sizes was obtained by changing the weight concentrations in SDS: 3.5% (Dd5 40 nm, W), 0.75% (Dd575 nm, �), 0.5% (Dd5 90 nm, �), and 0.25% (Dd5 115 nm, �). The open symbol is for miniemul- sion prepared with an aqueous phase containing 3.5 wt % SDS and 0.03 mol L21 of NaHCO3 (Dd5 60 nm, w). Cmonomer5 30 wt %, organic phase: MMA/BA/AA/SA. TABLE 2 Effect of Droplet Size on Colloidal Properties (Dp, Np/Nd) and Reactions Kinetics (Rp max) for a Self-Initiated Polymerization Performed in an Annular Photoreactor [SDS] concentration (% wt/wtmonomer) Dd (nm) Dp (nm) Np/Nd Conv after 2 h (%) Rp max (mol L21 s21) 3.5 40 70 0.21 94 0.10 3.5 [0.03 mol L21 of NaHCO3(aq)] 60 70 0.51 100 0.25 0.75 75 100 0.32 98 0.10 0.5 90 120 0.35 66 0.04 0.25 115 150 0.30 33 0.03 ARTICLE WWW.POLYMERCHEMISTRY.ORG JOURNAL OF POLYMER SCIENCE 8 JOURNAL OF POLYMER SCIENCE, PART A: POLYMER CHEMISTRY 2014, 00, 000–000 from works devoted to thermal self-initiation of alkyl acryl- ates and methacrylates.23,24,62 DFT conclusively proved the occurrence of a Flory mechanism during the thermal self- initiation of acrylate or methacrylate by evaluation of the chemical reaction transition state energies. The initiation mechanism was thought to proceed from a 1,4-biradical in the triplet state which subsequently reacts with a third monomer molecule to create two monoradical initiating spe- cies via hydrogen abstraction (favored with methacrylate) or hydrogen transfer (favored with acrylates). The proposed ini- tiation mechanism was depicted in Scheme 1. A qualitative insight into whether triplet states can react with acrylates to initiate the polymerization was provided by a sim- ple experiment. A monomer miniemulsion including a sensi- tizer (2% wt/wtmonomer) such as benzophenone (BP) or thioxanthone (TX) was irradiated through a longpass filter (k>300 nm) to hinder acrylate excitation. The kinetic analysis in Figure 11 showed a slow but complete conversion. The deactivation of aromatic ketones by acrylates is known to occur with relatively low rate constants (5.4 3 107 and 1.5 3 107 mol L21 s21 for the quenching of the triplet state of ben- zophenone and thioxanthone by methyl methacrylate, respec- tively).63 The direct hydrogen transfer from MMA to SCHEME 1 Proposed mechanism for photochemical self- initiation of alkyl acrylates and methacrylates. FIGURE 11 (a) Proposed initiation mechanism of (meth)acry- late monomers in the presence of an aromatic ketone. Although the initiation pathway shows a limited efficiency, the kinetic study indicates qualitatively that excited BP or TX can also react with the monomer, leading to initiating monoradical generation (X@H for acrylates and X@CH3 for methacrylates). (b) Acrylate conversion as function of irradiation time for differ- ent acrylate miniemulsion containing a photosensitizer: BP (W) and TX (�). FIGURE 12 Experimental ESR spectrum of a MMA solution containing DMPO spin trap (a). Simulated spectrum of the spin adducts arising from MMA reaction only (b). JOURNAL OF POLYMER SCIENCE WWW.POLYMERCHEMISTRY.ORG ARTICLE WWW.MATERIALSVIEWS.COM JOURNAL OF POLYMER SCIENCE, PART A: POLYMER CHEMISTRY 2014, 00, 000–000 9 benzophenone triplet state was suspected in,63 although this reaction was ruled out in a more recent paper.64 Besides this possible hydrogen transfer reaction, the possible formation of a 1,4-biradical between the ketone triplet state and the mono- mer was pointed out as the main reaction pathway.63 This assumption was further supported by semi-empirical calcula- tions showing that (i) the thermodynamics of the process depends on the ketone triplet state energy and (ii) a nice rela- tionship can be drawn between the quenching rate constant and the ketone triplet state energy.64 Hence, the 1,4-biradical is expected to be involved in the polymerization reaction, in a quite similar way as the acrylate triplet state (Fig. 11). Propagating Radicals ESR experiments were performed by irradiating monomer organic solutions in the presence of a spin trap (DMPO) to identify the propagating radical. Figure 12 shows the results in the case of MMA (but comparable results were obtained with other acrylates or methacrylates). The spectrum b of the main radical was simulated from the experimental spec- trum a after eliminating peaks identified as the degradation products of DMPO. The hyperfine coupling constants found (aN5 1.4 mT, aH5 2.1 mT) were fully consistent with those of carbon centered radical species.65,66 Similar results were obtained during the self-initiation of other. In all likelihood, this radical is thought to be the conventional methacrylate propagating radical (Scheme 2). As additional evidences, sim- ilar constants were obtained when adding a photoinitiator to the MMA. CONCLUSIONS High-solid content acrylic latexes (30 wt %) were produced without initiator via the UV-driven self-initiation of alkyl acrylate/methacrylate miniemulsions. A high reactivity was found for photopolymerizations performed both in spectro- photometric cells (a few minutes) and photoreactor (�1 h) when using small enough monomer droplets (40 nm) to minimize the light attenuation. Industrially, the strength is that such high polymerization rates could be achieved with a conventional medium-pressure Hg arc lamp displaying only a limited match with the absorption spectrum of the mono- mers. However, the necessity of handling concentrated nano- sized emulsions implies that issues of colloidal stability and viscosity are addressed. Other key parameters affecting molecular weight and reaction kinetics included the optical path and the irradiance. Mechanistically, the generation of initiating radicals by hydrogen abstraction or transfer from a diradical intermediate was proposed as a likely explanation for experimentally observed spontaneous initiation. REFERENCES AND NOTES 1 J. Ugelstad, M. S. El-Aasser, J. W. Vanderhoff, J. Polym. Sci. Lett. Ed. 1973, 11, 503–513. 2 S. V. Kostjuk, F. Ganachaud, Acc. Chem. Res. 2009, 43, 357–367. 3 F. M. Bauers, S. Mecking, Angew. Chem. Int. Ed. 2001, 40, 3020–3022. 4 M. Antonietti, K. Landfester, Prog. Polym. Sci. 2002, 27, 689– 757. 5 M. F. Cunningham, Prog. Polym. Sci. 2008, 33, 365–398. 6 P. B. Zetterlund, Y. Kagawa, M. Okubo, Chem. Rev. 2008, 108, 3747–3794. 7 Y. W. Luo, X. D. Zhou, J. Polym. Sci., Part A: Polym. Chem. 2004, 42, 2145–2154. 8 M. Goikoetxea, R. J. Minari, I. Beristain, M. Paulis, M. J. Barandiaran, J. M. Asua, J. Polym. Sci. Part A: Polym. Chem. 2009, 47, 4871–4885. 9 E. Bourgeat-Lami, J. Nanosci. Nanotechnol. 2002, 2, 1–24. 10 J. M. Asua, Prog. Polym. Sci. 2002, 27, 1283–1346. 11 F. J. Schork, Y. W. Luo, W. Smulders, J. P. Russum, A. Butte, K. Fontenot, In Polymer Particles; M. Okubo, Ed.; 2005; pp 129–255. 12 K. Landfester, Angew. Chem. Int. Ed. 2009, 48, 4488–4507. 13 J. Tonnar, E. Pouget, P. Lacroix-Desmazes, B. Boutevin, Macromol. Symp. 2009, 281, 20–30. 14 A. V. Fuchs, G. D. Will, Polymer 2010, 51, 2119–2124. 15 J. Dou, Q. Zhang, L. Jian, J. Gu, Colloid Polym. Sci. 2010, 288, 1751–1756. 16 A. Chemtob, B. Kunstler, C. Croutxe-Barghorn, S. Fouchard, Colloid Polym. Sci. 2010, 288, 579–587. 17 V. Daniloska, R. Tomovska, J. M. Asua, Chem. Eng. J. 2012, 184, 308–314. 18 I. Capek, Des. Monomers Polym. 2012, 15, 345–355. 19 P. A. Hoijemberg, A. Chemtob, C. Croutx�e-Barghorn, Macro- mol. Chem. Phys. 2011, 212, 2417–2422. 20 P. A. Hoijemberg, A. Chemtob, J. Poly, C. Croutx�e-Barghorn, A. M. Braun, Macromolecules 2011, 44, 8727–8738. 21 V. Daniloska, R. Tomovska, J. M. Asua, Chem. Eng. J. 2013, 222, 136–141. 22 F. Jasinski, E. Lobry, A. Chemtob, C. Croutxe-Barghorn, A. Criqui, Macromol. Chem. Phys. 2013, 214, 1669–1676. 23 S. Srinivasan, M. W. Lee, M. C. Grady, M. Soroush, A. M. Rappe, J. Phys. Chem. A 2009, 113, 10787–10794. 24 S. Srinivasan, M. W. Lee, M. C. Grady, M. Soroush, A. M. Rappe, J. Phys. Chem. A 2011, 115, 1125–1132. 25 J. M. Asua, Polymer Reaction Engineering; Wiley-Blackwell: Oxford, 2007. 26 N. B. Cramer, J. P. Scott, C. N. Bowman, Macromolecules 2002, 35, 5361–5365. 27 J. P. Fouassier, J. Lalev�ee. Photoinitiators for Polymer Synthe- sis: Scope, Reactivity and Efficiency; Wiley-VCH: Weinheim, 2012. 28 T. Y. Lee, T. M. Roper, E. S. Jonsson, I. Kudyakov, K. Viswanathan, C. Nason, C. A. Guymon, C. E. Hoyle, Polymer 2003, 44, 2859–2865. 29 T. Scherzer, W. Knolle, S. Naumov, C. Elsner, M. R. Buchmeiser, J. Polym. Sci., Part A: Polym. Chem. 2008, 46, 4905–4916. 30 W. Knolle, T. Scherzer, S. Naumov, R. Mehnert, Radiat. Phys. Chem. 2003, 67, 341–345. 31 T. Scherzer, J. Polym. Sci. Part A: Polym. Chem. 2004, 42, 894–901. SCHEME 2 Propagating radical arising from methacrylate polymerization. ARTICLE WWW.POLYMERCHEMISTRY.ORG JOURNAL OF POLYMER SCIENCE 10 JOURNAL OF POLYMER SCIENCE, PART A: POLYMER CHEMISTRY 2014, 00, 000–000 32 F. Bauer, U. Decker, K. Czihal, R. Mehnert, C. Riedel, M. Riemschneider, R. Schubert, M. R. Buchmeiser, Prog. Org. Coat. 2009, 64, 474–481. 33 T. Scherzer, Macromol. Chem. Phys. 2012, 213, 324–334. 34 H. Wang, H. R. Brown, Macromol. Rapid Commun. 2004, 25, 1095–1099. 35 L. Huang, Y. Li, J. Yang, Z. Zeng, Y. Chen, Polymer 2009, 50, 4325–4333. 36 M. Schwartz, R. Baumstark. Waterbased Acrylates for Deco- rative Coatings; Vincentz Verlag: Hannover, 2001. 37 K. Dietliker. Compilation of Photoinitiators Commercially Available for UV Today, Sita technology Ltd: Edinburgh, Lon- don, 2002. 38 P. Ionita, B. C. Gilbert, A. C. Whitwood, Lett. Org. Chem. 2004, 1, 70–74. 39 O. Mengual, G. Meunier, I. Cayre, K. Puech, P. Snabre, Col- loids Surf., A 1999, 152, 111–123. 40 J. Workman, L. Weyer. Practical Guide to Interpretive Near- Infrared Spectroscopy; CRC Press: Boca Raton, 2007. 41 A. Ibrahim, V. Maurin, C. Ley, X. Allonas, C. Croutxe- Barghorn, F. Jasinski, Eur. Polym. J. 2012, 48, 1475–1484. 42 N. Anton, T. F. Vandamme, Int. J. Pharm. 2009, 377, 142–147. 43 C. Solans, P. Izquierdo, J. Nolla, N. Azemar, M. J. Garcia- Celma, Curr. Opin. Colloid Interface Sci. 2005, 10, 102–110. 44 T. Tadros, R. Izquierdo, J. Esquena, C. Solans, Adv. Colloid Interface Sci. 2004, 108, 303–318. 45 T. J. Wooster, M. Golding, P. Sanguansri, Langmuir 2008, 24, 12758–12765. 46 The progressive decrease of the reflectance is indicative of an enhanced diffusional degradation causing a depletion of droplets in the middle of the sample (clarification). The con- comitant increase in R at the top of the sample (see Supporting Information Figure S3 of ESI) demonstrates that a creaming destabilization process occurs practically. 47 K. Landfester, F. J. Schork, V. A. Kusuma, C. R. Chim. 2003, 6, 1337–1342. 48 T. Cosgrove, Colloid Science, Principles, Methods and Appli- cations; John Wiley and Sons: Chichester (UK), 2010. 49 B. Charleux, Macromolecules 2000, 33, 5358–5365. 50 J. P. Fouassier, Photoinitiation, Photopolymerization and Photocuring: Fundamental and Applications; Hanser Publisher: New York, 1995. 51 To be close to polymerization conditions performed indus- trially, distillation of the monomers to remove inhibitors and degassing to remove oxygen dissolved in the aqueous phase were not carried out. 52 J. E. Duran, M. Mohseni, F. Taghipour, Aiche J. 2011, 57, 1860–1872. 53 M. N. Chong, B. Jin, C. W. K. Chow, C. Saint, Water Res. 2010, 44, 2997–3027. 54 A. M. Braun, M.-T. Maurette, E. Oliveros, Photochemical Technology; John Wiley and Sons: Chichester, 1991. 55 S. Teixeira, R. Giudici, S. H. Bossmann, J. Lang, A. M. Braun, Chem. Eng. Process. 2004, 43, 1317–1328. 56 A. F. M. Silvares, C. A. O. Nascimento, E. Oliveros, S. H. Bossmann, A. M. Braun, Chem. Eng. Process. 2006, 45, 1001–1010. 57 R. J. J. Jachuck, V. Nekkanti, Macromolecules 2008, 41, 3053–3062. 58 V. Nekkanti, R. J. J. Jachuck, J. Appl. Polym. Sci. 2010, 116, 1940–1947. 59 C. G. Dobie, MarijaVicevic, K. V. K. Boodhoo, Chem. Eng. Process. 2013, 71, 97–106. 60 M. Manea, A. Chemtob, M. Paulis, J. C. de la Cal, M. J. Barandiaran, J. M. Asua, AIChE J. 2008, 54, 289–297. 61 T. Scherzer, W. Knolle, S. Naumov, R. Mehnert, Nucl. Ins- trum. Methods Phys. Res. Sect. B 2003, 208, 271–276. 62 S. Srinivasan, M. W. Lee, M. C. Grady, M. Soroush, A. M. Rappe, J. Phys. Chem. A 2010, 114, 7975–7983. 63 H. J. Timpe, K. P. Kronfeld, J. Photochem. Photobiol. 1989, 46, 253–267. 64 V. Lemee, D. Burget, P. Jacques, J. P. Fouassier, J. Polym. Sci. Part A: Polym. Chem. 2000, 38, 1785–1794. 65 M. Pinteala, S. Schlick, Polym. Degrad. Stab. 2009, 94, 1779–1787. 66 See the the spin trap database of the National Institure of Envi- ronmental Health Science (http://www.tools.niehs.nih.gov/stdb/). JOURNAL OF POLYMER SCIENCE WWW.POLYMERCHEMISTRY.ORG ARTICLE WWW.MATERIALSVIEWS.COM JOURNAL OF POLYMER SCIENCE, PART A: POLYMER CHEMISTRY 2014, 00, 000–000 11