pubs.acs.org/Langmuir © 2009 American Chemical Society
Diffusion Behavior of Isobornyl Acrylate into Photopolymerized Urethane Acrylate Films: Influence of Surface Oxidation during Curing Ali Mougharbel,† Jacky Mallegol,‡ and Xavier Coqueret*,† †
Universit e de Reims Champagne Ardenne, Institut de Chimie Mol eculaire de Reims CNRS UMR 6229, Moulin de la Housse, BP 1039, 51687 Reims Cedex 2, France, and ‡ArcelorMittal Li ege, Boulevard De Colonster B57, 4000 Li ege, Belgium Received March 17, 2009. Revised Manuscript Received May 12, 2009 The diffusion of liquid isobornyl acrylate (iBoA) into photopolymerized films of urethane acrylate was shown to be strongly dependent on curing conditions. Fourier transform infrared (FTIR) monitoring of iBoA sorption into films photopolymerized in oxygen-free conditions reveals that diffusion proceeds according to Fick’s laws. The diffusion coefficient and the total amount of absorbed monomer are dependent on the degree of cure of the polymer network. When polymerized in air, the films exhibit an anomalous behavior, with a lag time increasing with the dose of applied radiation. At a comparable degree of cure, films photopolymerized in air showed a retarded diffusion with a characteristic diffusion constant close or equal to the Fickian diffusion coefficient observed with the film polymerized under nitrogen, suggesting that the retardation phenomenon was due to differences in surface interactions with liquid the monomer. X-ray photoelectron spectroscopy (XPS) measurements clearly revealed the occurrence of surface oxidation attributed to direct photolysis of urethane functions by the short wavelength component of the incident UV light. The lag time observed in samples polymerized in aerobic conditions is interpreted by the existence of H-bonded hydroperoxides forming a temporary barrier retarding monomer diffusion.
Introduction Radiation induced polymerization is now a well-established method for elaborating polymeric networks with unique possibilities in terms of temporal and spatial control.1 Beyond the numerous applications in graphic arts2 and microelectronics3 where a single layer of reactive formulation is subject to curing, some basic or advanced applications include the sequential application and curing of layers having distinct compositions in view of specific properties. This situation is encountered for UVcured overprint varnishes, two-layered claddings for optical fibers, and coatings of steel coils. By repeating the application and curing cycle, thick parts can be gradually produced as for layer-by-layer production of high-performance fiber reinforced composites cured by application of low-energy electron beam treatment. There is also the unique possibility of producing gradient materials with tailored properties in the depth by changing the composition of the coated formulation. A common issue in the multilayered materials produced by the sequential application and polymerization process is obtaining a satisfactory adhesion between the layers without altering the properties of the individual materials. In order to ensure good adhesion through anchoring, the diffusion of one or some reactive constituents from the liquid coating into the cured underlying film is a key parameter needing accurate quantification. The phenomena occurring at the interface between the liquid and the cured layer have only received limited attention to date.4 In this context, the control of diffusion of some reactive components of a liquid layer into the underlying radiation cured layer is a key
point for ensuring satisfactory anchoring of the top material, possibly in the form of an interpenetrating network. We have therefore examined the case of multilayered organic coating systems involving more specifically the diffusion of an acrylic monomer, isobornyl acrylate (iBoA), into a urethane acrylate substrate cured by free radical photoinitiated polymerization. Our study was focused on the determination of the transport mechanism in the model system and of the associated diffusion parameters, as function of curing conditions and of sorption temperature. Various configurations of IR spectroscopy are increasingly used to study in a quantitative manner transport phenomena in polymers materials, particularly time-resolved ATR-IR5 and IR imaging.6 We made use of conventional Fourier transform infrared (FTIR) spectrophotometry in the transmission mode, that allows a direct and accurate determination of the amount of permeant in the entire film sample, cast and photopolymerized on a sodium chloride plate.
Experimental Section Materials. The acrylate monomers used as swelling agents for the UV-cured urethane acrylate films are isobornyl acrylate (iBoA), 1,6-hexanediol diacrylate (HDDA), both obtained from Cytec (Belgium), and n-butyl acrylate (nBuA) purchased from Aldrich (France).
*To whom correspondence should be addressed. Telphone/fax: 33 3 26913338. E-mail:
[email protected]. (1) Scranton, A. C.; Bowman, C. N.; Peiffer, R. W. Photopolymerization: Fundamentals and Applications; ACS Symposium series 673; American Chemical Society: Washington, DC, 1997. (2) Haverkamp, R. G.; Siew, D. C. W.; Barton, T. F. Surf. Interface Anal. 2002, 33, 330. (3) Fahlman, M.; Salaneck, W. R. Surf. Sci. 2002, 500, 904. (4) Hinder, S. J.; Lowe, C.; Maxted, J. T.; Perruchot, C.; Watts, J. F. Prog. Org. Coat. 2005, 54, 20.
Langmuir 2009, 25(17), 9831–9839
(5) Elabd, Y. A.; Baschetti, M. G.; Barbari, T. A. J. Polym. Sci., Part B: Polym. Phys. 2003, 41, 2794. (6) Challa, S. R.; Wang, S. Q.; Koenig, J. L Appl. Spectrosc. 1996, 50, 1339.
Published on Web 06/09/2009
DOI: 10.1021/la9009338
9831
Article The urethane acrylate resin (Cytec, Belgium) is trifunctional aliphatic prepolymers (PUA) diluted with HDDA (25 wt %). The photopolymerization was initiated by 2-hydroxy-2-methyl-1phenyl-propan-1-one (HMPP) introduced in the urethane acrylate formulation at a concentration of 0.5 wt %.
Irradiation Sources. The samples prepared for recording the IR spectral changes upon polymerization and for determining the network Tg by differential scanning calorimetry (DSC) were treated in a Vilber Lourmat irradiation chamber (Biolink BLXE365) equipped with 365 nm fluorescent tubes delivering an irradiance of 8 mW 3 cm-2. Film specimens prepared for sorption measurements were cured with an IST Metz UV source (medium pressure Hg bulb) operated at a conveyor speed ranging between 10 and 110 m 3 min-1 and an electric power of 100 W 3 cm-1. Spectrocopic Measurements. Photopolymerization kinetics were monitored by discontinuous FTIR measurements on 10 μm thick films cast on polished NaCl plates. Acrylate monomer diffusion into the cured films was also monitored by FTIR with the same type of substrate. A Bruker Alpha-T IR spectrometer was used in the transmission mode. The variations of absorbance at 810 cm-1 was used to quantify the consumption of acrylate functions upon photoinitiated polymerization or the increase of acrylate concentration by gradual sorption into samples immersed in the liquid monomers. The urethane deformation band located at 770 cm-1 is invariant during the photopolymerization reaction and during sorption. The changes of absorbance at 810 cm-1 were thus normalized relatively to the 770 cm-1 absorbance taken as reference value, so as to minimize the influence cell manipulation. Characterization of Thermal Properties of Cured PUA Films. The glass transition temperature (Tg) in PUA materials exhibiting various degrees of cure was determined using a TA Q100 differential scanning calorimeter. Two samples of PUA formulation were cured simultaneously in DSC pans under the same irradiation conditions under nitrogen flow at a temperature (60-120 C) controlled by using a Linkam hot-stage (England) covered with a glass plate and placed in the Biolink irradiation chamber. One sample was used to measure the conversion ratio by FTIR (powdered sample in KBr pellets), and the other one was used to measure Tg. The calorimetric scans were carried out from -50 to 150 C at a rate of 10 C 3 min-1 immediately after sample preparation. Sorption Kinetics. IR plates were used as support to monitor sorption kinetics. The polyurethane formulation was coated on IR plates with a bar coater designed for 10 μm wet films. Photopolymerization was carried out with an IST mercury lamp system (at a speed ranging between 10 and 110 m 3 min-1, at an incident light intensity of 100 W 3 cm-1). Various UV-curing conditions corresponding to different radiant energies (from 0.033 to 1 J 3 cm-2) led to several films with different values of conversion ratios. The lamp intensity (UV dose) was measured with a radiometer (Powerpuck multispecter). One side contact occurred between the cured polyurethane films (25 mm 10 mm 0.015 mm) and mono- or difunctional monomer as the immersion was taking place in 15 mL flasks. The volume of the polymer film compared to the liquid volume was negligible. Sorption kinetics were monitored at 22 and 46 C. At controlled time intervals, the film was withdrawn from the flask, wiped smoothly, and analyzed by FTIR spectroscopy. Diffusion phenomena were followed until stabilization of the amount of absorbed acrylate. 9832 DOI: 10.1021/la9009338
Mougharbel et al.
Determination of Diffusion Coefficients. The cured polyurethane film cast on the sodium chloride plate is immerged in the flask containing the liquid monomer so as to have a direct contact between the solid and liquid. The experimental data can be treated with the formalism of the second Fick’s law, considering an unsteady one-dimensional diffusion under isothermal conditions, as described by eq 1: Dc D Dc ¼ D ð1Þ Dt Dx Dt where c is the concentration of monomers in the network, x is the distance from the surface, t is the time, and D is the diffusion coefficient for liquid into the polymer, assuming that it is independent of monomer concentration. For a plane sheet of thickness e, Fick’s second law of diffusion was solved7 to obtain the expression of eq 2: ! ¥ mt 8 X 1 ð2n þ 1Þ2 π2 ¼ 1- 2 exp Dt ð2Þ m¥ e2 π n ¼0 ð2n þ 1Þ2 where mt is the sorbed mass at time t, m¥ is the equilibrium mass, and e is the thickness of the studied film. For short contact times, mt/m¥ is small, and eq 2 is approximated by the expression of eq 3: mt 4 Dt 1=2 ¼ ð3Þ e π m¥ By using the FTIR data obtained in the present study, mt/m¥ can be changed for ΔRt/ΔR¥ and eq 3 is rewritten in the form of eq 4: ΔRt Rt -R0 4 Dt 1=2 ¼ ¼ ð4Þ e π ΔR¥ R¥ -R0 where Rt = Atvariable/Atinvariable is the ratio of the absorbance of the variable band Atvariable to the absorbance of the invariable band ¥ Atinvariable at time t and R¥=A¥ variable/Ainvariable is the ratio of the absorbance of the variable band A¥ variable to the absorbance of the invariable band A¥ invariable at equilibrium (t f ¥). Diffusion coefficients were determined from the presentation of ΔRt/ΔR¥ versus the square root of time. For a Fickian behavior, the beginning of the sorption kinetics is linear and the diffusion coefficient D is proportional to the square of the slope β as expressed in eq 5. D ¼
πe2 β2 16
ð5Þ
Surface Characterization. X-ray Photoelectron Spectroscopy (XPS) Measurements. The surface composition of the polyurethane films was analyzed using an ESCALAB 250 X-ray Photoelectron spectrometer. The analysis was carried out at room temperature. The binding energy (BE) scale was calibrated by setting the C1s transition at 285.0 eV. The accuracy of BE values was (0.2 eV. The BE values corresponding to all contributions were obtained by using the Peak-fit program implemented in the control software of the spectrometer. ATR-IR Characterization. Characterization of the surface of the polyurethane films was also carried out with an attenuated total reflection (ATR-IR) accessory (Eurolabo). The film samples were placed on the two sides of the KRS-5 crystal (multiple reflexions). Each sample was scanned 160 times with a resolution of 4 cm-1 to cover the 400-4000 cm-1 spectral domain. The probed depth was estimated to range between 0.3 and 3 μm in the investigated spectral domain. (7) Crank, J. The mathematics of diffusion, 2nd ed.; Oxford Science Publications: Oxford, 1975.
Langmuir 2009, 25(17), 9831–9839
Mougharbel et al.
Article
Results and Discussion The diffusion behavior of a penetrant in polymers has been studied extensively, both theoretically and experimentally. Diffusion of a penetrant in a rubbery polymer is often Fickian, whereas in a glassy polymer diffusion can exhibit non-Fickian (or anomalous) behavior, especially when there is extensive swelling of the polymer matrix with the penetrant. For molecular migration in an amorphous medium, the permeant needs free volume space of sufficient size adjacent to a molecule, and this molecule must possess enough energy to overcome intermolecular interactions to jump into the hole. For a polymer above its glass transition temperature, polymer chains can adjust their segment motion by thermal vibration, and the free volume usually becomes the predominant factor. Based on this conceptual framework, several versions of free-volume theory were proposed to predict the solvent diffusion coefficient in polymers.8-10 The quantitative investigation of the effects of polymerization conditions on the sorption of simple acrylate monomers into model polyurethane acrylate films requires a sufficient knowledge and mastering of the degree of curing. Particular attention was paid to film preparation so as to evaluate the properties of film samples prepared under well-controlled conditions. PUA Film Preparation. The films based on polyurethane acrylate networks were prepared by photopolymerization in air or a nitrogen atmosphere. FTIR spectroscopy was implemented in the transmission mode to determine the average fractional conversion π of acrylate functions. The spectra shown in Figure 1a were recorded during the curing under a lower power fluorescent source of a 15 μm thick sample PUA sample containing 0.5 wt % HMPP. The gradual consumption of the unsaturations is deduced from the decrease of the absorbance at 810 cm-1, whereas the urethane band at 770 cm-1 appears essentially unaffected by the photopolymerization process. The observed variations of absorbance for a sample submitted to repeated irradiations were transformed into conversion values by using eq 6 770 A810 DUV =ADUV π ¼ 1 - 810 A0 =A770 0
ð6Þ
where π is the acrylate conversion ratio after exposure to dose and ADUV810 are the absorbances of the DUV (mJ.cm-2), A810 0 acrylate band at the selected wavenumbers, before and after exposure to dose DUV, respectively. The corresponding conversion profile is drawn in Figure 1b. The PUA resin exhibits a high level of reactivity under UV irradiation, even upon exposure to the 365 nm fluorescent Biolink source, that activates the weakly absorbing n-π* transition of the initiator HMPP.11 The steep slope observed for doses below 0.5 J 3 cm-2 indicates a very fast process until conversion π ≈ 0.60, and then the rate levels off with an almost linear dependence from π=0.75 to 0.80 when increasing the cumulated light dose from 4 to 16 J 3 cm-2. The presence of a multifunctional prepolymer diluted with a difunctional monomer in the formulation allows the formation of a three-dimensional polymer network and a gel effect contributing strongly for the observed initial efficiency.12 The later stage is marked by the gradual vitrification effect that limits the progress of the reaction in the densifying network.13 (8) Vrentas, J. S.; Duda, J. L. J. Polym. Sci., Polym. Phys. Ed. 1977, 15, 403. (9) Vrentas, J. S.; Duda, J. L. J. Polym. Sci., Polym. Phys. Ed. 1977, 15, 417. (10) Paul, C. W. J. Polym. Sci., Polym. Phys. Ed. 1983, 21, 425. (11) Fouassier, J. P. Radiation curing in polymer science and technology; Elsevier Science Publishers LTD: England, 1993; Chapter 2. (12) Andrzewejska, E. Prog. Polym. Sci. 2001, 26, 605. (13) Jager, W. F.; Lungu, A.; Chen, D. Y.; Neckers, D. C. Macromolecules 1997, 30, 780.
Langmuir 2009, 25(17), 9831–9839
Figure 1. (a) FTIR monitoring of the photopolymerization kinetics of PUA resin exposed to low-intensity UV radiation. (b) Progress of polymerization as a function of cumulated exposure to low-intensity UV radiation.
The thermal properties of the resulting polymer networks were analyzed by DSC for different polymerization temperatures (between 20 and 120 C). The evolution of glass temperatures from 60 to 100 C for conversion ratios π increasing from 0.85 to more than 0.98 (Figure 2) clearly illustrates the final control of the curing process by the thermal conditions. This set of data allowed us to produce in a reproducible manner films cured in air or in an inert atmosphere with various degrees of curing, in the perspective of a detailed sorption study. For obtaining the desired levels of acrylate conversion in samples polymerized in air, the UV dose was adjusted to compensate the inhibiting effect of O2. Sorption Kinetics. Preliminary experiments carried out at 22 C for evaluating the adequate conversion range for the films and for determining appropriate swelling acrylates have shown that n-butyl acrylate as well as hexanediol diacrylate were severely damaging the films, by inducing cracks and delamination from the underlying NaCl substrate. This phenomenon becomes even more critical when experiments are carried out at 46 C. Nevertheless, by using isobornyl acrylate (iBoA) as the liquid medium, the interaction parameters and the molar volume of the swelling liquid were more convenient and allowed monitoring satisfactorily the sorption kinetics of iBoA into PUA films with conversion ratios π ranging between 0.50 and 0.70. iBoA Sorption into Films Photopolymerized in OxygenFree Conditions. The FTIR spectra recorded to monitor single side sorption of iBoA at 46 C into a film cured in oxygen-free conditions to a π value of 0.64 are plotted in Figure 3a. The band at 770 cm-1 is essentially unaffected by the diffusion process, and DOI: 10.1021/la9009338
9833
Article
Mougharbel et al.
Figure 2. Variations of the glass transition temperature in photopolymerized PUA samples as a function conversion ratio (samples cured in DSC pans under low-intensity UV radiation at temperatures ranging between 20 and 120 C).
its absorbance was used as a reference value for measuring the increase of absorbance at 810 cm-1 as a consequence of swelling in iBoA. The normalized variation of the amount of sorbed iBoA versus the square root of time is represented in the graph of Figure 3b, exhibiting a linear variation over a broad time domain before leveling off to a value denoting sample saturation. These different features support a typical Fickian diffusion behavior for the system under study. Various additional experiments achieved with samples cured at different levels and immersed in iBoA at 22 C confirmed such a behavior. The diffusion and swelling parameters deduced from these experiments will be discussed in a forthcoming section of this paper. iBoA Sorption into Films Photopolymerized in Air. A different situation was observed for PUA films irradiated under air and placed in iBoA for sorption experiments at 22 C or at 46 C. Figure 4 shows that plots representing the variations of ratios Rt/R¥ as a function of t0.5 have a sigmoidal shape not consistent with the Fickian diffusion process. A number of curves drawn on the basis of 8-12 measurement points consistently revealed the retarded starting of iBoA uptake, with a lag time increasing with the sample conversion increase. A linear segment was identified, though the corresponding slope is determined with fewer points and subject to higher experimental errors. The sigmoid shape in plots featuring a deviation from the Fickian linear dependence of penetrant uptake with t0.5 can be related to various phenomena.14 In some situations, as for case II diffusion, swelling influences mass transport by changing the solubility or the diffusion coefficient of the penetrant in the host polymer. In other cases, the concentration profiles lead to a swelling gradient across the film, with generation of a swelling stress that modifies the geometry of the sample. In the present situation, the difference in diffusion behavior is expected to mainly result from superficial effects during the preparation of the networks, with no or very limited effects on the bulk structure of the host network. The comparison between the diffusion profiles at 46 C for two samples with approximately the same degree of cure and prepared under similar conditions, except on the presence of oxygen during the photopolymerization, reveals a distinct feature (Figure 5). The difference appears in terms of retardation of the diffusion phenomenon, and in a lower value for the diffusion coefficient for the sample cured in air, but equilibrium swelling ratio is unchanged.
(14) Sanopoulou, M.; Petropoulos, J. H. Macromolecules 2001, 34, 1400.
9834 DOI: 10.1021/la9009338
Figure 3. Sorption kinetics at 46 C of iBoA into PUA films photopolymerized in oxygen-free conditions (acrylate conversion π= 0.64): (a) FTIR monitoring of the increase of acrylate absorbance as a consequence of sorption and (b) profile of the relative iBoA uptake ΔRt/ΔR¥ versus square root of time.
Irradiation under air is likely to induce a modification in the chemical composition of the extreme surface, which can generate new types of interactions with the isobornyl acrylate that were absent in the PUA films irradiated under nitrogen. This interpretation will be further supported by experimental facts and discussed in the last section of this paper. Lag Time. Lag times τ quantifying the retardation to iBoA diffusion were calculated for all experiments by projecting the intermediate linear segment of the sigmoid to the abscissa axis. As already mentioned, the limited number of points for determining the straight line representing the Fickian transient yields τ values with moderately low precision. The dependence of τ values, typically ranging between 1 and 10 min, with the conversion ratio of the host network is plotted in Figure 6. The main general trend appearing in this graph is the increase of the lag time with increasing degree of cure. Second, on the basis of sorption experiments conducted at 22 and 46 C, it seems that a higher temperature tends to enhance the retarding effect. The influence of temperature on the lag time is unexpected if one considers a superficial barrier effect to diffusion, sensitive to a thermally activated permeation. The observed behavior will find a possible explanation in the discussion to come later. Finally, the dispersion of experimental points is larger for samples having a higher level of conversion. In the higher conversion range, we found that the networks did not accommodate well large swelling ratios, with some propensity to cracking, the swell-induced deformation exceeding the elasticity limit of this type of material. For this reason, only the diffusion kinetics and swelling data of films with a moderate degree of cure will be considered as intrinsic Langmuir 2009, 25(17), 9831–9839
Mougharbel et al.
Figure 4. Evolution of the relative iBoA uptake ΔRt/ΔR¥ as a function of the square root of time for PUA films photopolymerized in air: sorption at 22 C, acrylate conversion ratio π = 0.54; sorption at 46 C, acrylate conversion π = 0.55.
Figure 5. Evolution of the relative iBoA uptake ΔRt/ΔR¥ as a function of the square root of time during sorption kinetics at 46 C (PUA films photopolymerized in oxygen-free conditions (9) and in air (b), conversion in both samples π = 0.64).
values of the materials. Above the critical value of π=0.65, the determined quantities are considered as apparent diffusion values, resulting from various transport and sorption mechanisms in a geometrically complex samples. Swelling Ratio at Equilibrium. The measurement of sorption values at equilibrium allows determination of the swelling ratio (SR) for various samples immersed in iBoA at different temperatures (Figure 7). Defining SR as the ratio of the amount of iBoA sorbed at equilibrium into the PU film (mtf¥ iBoA) to the total mass of PU film after sorption (mt = mfilm + mtf¥ iBoA), we obtained SR values between 0.02 and 0.08. The results indicate an important effect of temperature and conversion ratio on the amount of iBoA penetrated into the polyurethane films. As expected, SR values are lower when the conversion ratio increases and hence the crosslinking density. A temperature increase of 24 C was shown to typically increase the swelling in iBoA by a factor of 2. Diffusion Coefficients. Diffusion coefficients D were calculated from eq 5 with an acceptable accuracy from the slope of the regular or retarded Fickian plots, allowing us to examine in some detail the effects of curing conditions and of temperature on the dependence of D as a function of sample conversion. Error bars added to the plots indicate the uncertainty on D values, as evaluated from the linear regression calculation. The data reported in Figure 8 show the evolution of diffusion coefficients calculated at 46 C. The diffusion coefficients of iBoA Langmuir 2009, 25(17), 9831–9839
Article
Figure 6. Influence of acrylate conversion on the lag time observed at the beginning of all sorption kinetics carried out at 22 C and at 46 C in PUA films photopolymerized in air.
Figure 7. Influence on acrylate conversion on the equilibrium swelling ratio in iBoA at 22 and 46 C for PUA films photopolymerized in air and in free-oxygen conditions.
in PUA films photopolymerized in oxygen-free conditions are larger than those obtained in films irradiated under air. For a given value of conversion ratio (π =0.60), a factor of 2 can be assessed. The penetration of iBoA into PUA films photopolymerized in air seems to be hindered by superficial barrier effects related to a possible change in the chemical composition of the top surface after curing in air. The inhibiting effect of oxygen in the bulk is more likely believed to generate defects in the network structure, resulting in an easier diffusion of the penetrant. In line with the dependence observed for swelling ratios at equilibrium, the diffusion coefficients decrease when the conversion ratio increases, whatever the irradiation and the sorption conditions. The conversion dependence of the apparent diffusion coefficients for the two sorption temperatures is presented in Figure 9. The D value range covers more than one decade, between 1 10-15 and 210-14 m2 3 s-1. These values are two magnitude orders below those estimated in the study of the diffusion of inks into polyurethane acrylate at 25 C15 and the diffusion of benzene vapor into polyurethane films at 35 C.16 The conversion dependence of diffusion coefficients is more important for the series of experiments conducted at 46 C than for those performed at 22 C. This observation can be tentatively interpreted by the difference between sorption temperature and (15) Tey, J. N.; Soutar, A. M.; Priyadarshi, A.; Mhaisalkar, S. G.; Hew, K. M. J. Appl. Polym. Sci. 2007, 103, 1985. (16) Yang, Y.; Huang, Y.; Chen, Y.; Wang, D.; Liu, H.; Hu, C. J. Appl. Polym. Sci. 2004, 91, 2984.
DOI: 10.1021/la9009338
9835
Article
Mougharbel et al. Table 1. Elemental Compositions on the PUA Surface of Films Photopolymerized in Nitrogen Atmosphere and in Air atomic composition oxygen nitrogen carbon oxygen/carbon
Figure 8. Influence of acrylate conversion on the apparent diffusion coefficients (m2 3 s-1) at 46 C of iBoA into films photopolymerized in oxygen-free conditions and in air.
Figure 9. Influence of sorption temperature on the conversion dependence of iBoA diffusion coefficients D (m2 3 s-1) into UVcured PUA films.
the glass transition of the networks. Whereas for experiments conducted at 22 C, the materials are in the glassy state in the whole range of conversion covered by the selected film samples, At 46 C, the samples with the lower degree of cure (0.52 < π < 0.56) are much closer to their Tg, but as the conversion increases the Tg shifts to higher values (Figure 2) with a strong limitation on the molecular diffusion mechanism.17 The various features revealed by the study of iBoA sorption kinetics into photopolymerized PUA films highlighted the effect of the irradiation conditions, particularly (under air or in nitrogen atmosphere). Among the various issues raising basic questions of interest and requiring further explanation, the influence of the presence of oxygen at the film surface during free radical curing appears essential. We have consequently further explored the above-mentioned assumption on the effect of the irradiation conditions on the chemical composition of PUA film surfaces. Surface Characterization. Two types of samples were prepared by exposure to the medium pressure Hg source (IST system) so as to obtain the same conversion degree (π = 0.63 ( 0.02). A flat gastight irradiation chamber covered with a 30 μm thick PE film was used, allowing similar exposure to the UV light and control of the surrounding atmosphere (air for one type of samples and nitrogen for the second type). The chemical composition on the top surface in contact with ambient atmosphere during photopolymerization was analyzed (17) Chuda, K.; Smolinski, W.; Defoort, B.; Rudz, W.; Gawdzik, B.; Rayss, J.; Coqueret, X. Polimery 2004, 49, 505.
9836 DOI: 10.1021/la9009338
PUA exposed under N2
PUA exposed in air
22.8 5.2 72.0 0.32
23.5 4.6 71.8 0.33
by FTIR in ATR mode. The presence of additional oxygencontaining functions to be related with the UV irradiation in the presence of air was searched for, but no significant new band was observed by probing the samples within a few micrometers in depth and comparing their mid-infrared spectra in the 4000-400 cm-1 range (see the Supporting Information). Significant differences in the superficial conversion ratio measured with the 810 cm-1 acrylate deformation band were however measured. For the film irradiated under air, we determined the superficial conversion ratio π = 0.65, whereas the film irradiated in oxygen-free conditions presented a higher superficial conversion value π = 0.70. This observation is consistent with the expected inhibiting effect of O2 at the film surface, but cannot be simply related to the retardation effect on iBoA sorption. More likely, photopolymerization in air could modify chemical constitution of the extreme surface, generating a very thin barrier to the penetration of iBoA. XPS was used to analyze the chemical composition of the top 1-10 nm of the film surface in the two types of samples similar to those used for the diffusion study. The low-resolution XPS spectra exhibit peaks at 533, 400, and 285 eV due to oxygen (O1s), nitrogen (N1s), and carbon (C1s), respectively. The corresponding atomic compositions are reported in Table 1. The elemental composition at the surface is essentially the same, independent of the irradiation conditions. The O/C atomic ratio is close to 0.32 in both samples. The binding energy in the high-resolution C1s XPS spectra ranges from 284 to 290 eV. The C1s spectra can be resolved into four contributions (see the Supporting Information).18,19 On the basis of literature data, we have assigned the peaks obtained by deconvolution with maxima at 284.9, 285.5 ( 0.1, 286.5 ( 0.1, and 289.3 eV to the functional groups of C-C/C-H, C-O, C-N, and X-C=O (X = O, N), respectively, as shown in Table 2. The distribution of C atoms in different types of functions is consistent with the chemical constitution of the aliphatic polyurethane acrylate diluted in hexanediol diacrylate shown in the Experimental Section. The most salient feature is the difference between atomic percents corresponding to the C-O and the C-N carbon atoms, whereas the content in carbonyls is slightly lower in the sample treated in air (Table 2). Films irradiated in the presence of oxygen present a higher contribution of C-O bonds (32.9.7%) compared to the other one calculated for films irradiated under nitrogen (26.4%). On the other hand, the elemental composition associated to the C-N bonds is more important in the case of films irradiated under nitrogen (12.5%) compared to films irradiated under air (8.0%). The high-resolution O1s XPS spectra are consistent overall with the C1s profiles and comprise mainly CdO (532.3 ( 0.1 eV) and C-O (533.8 ( 0.1 eV). The relative content of O atoms in C-O bonds is higher in the films irradiated in air and increases at the expense of the content in carbonyl oxygens (Table 3).
(18) Mishra, A. K.; Chattopadhyay, D. K.; Sreedhar, B.; Raju, K. V. S. N. Prog. Org. Coat. 2006, 55, 231. (19) Sanchis, M. R.; Calvo, O.; Fenollar, O.; Garcia, D.; Balart, R. J. Appl. Polym. Sci. 2007, 105, 1077.
Langmuir 2009, 25(17), 9831–9839
Mougharbel et al.
Article
Table 2. Contributions of the Different Types of Atomic Environments in High-Resolution C1s Spectra for PU Films Photopolymerized in Air (PUA-Air) and in Nitrogen (PUA-N2) PUA exposed under N2
PUA exposed in air
C atom environment
binding energy (eV)
atom %
binding energy (eV)
atom %
C-C and C-H C-N C-O -COO- and -CON-
284.9 285.6 286.4 289.3
46.7 12.2 26.4 14.4
284.9 285.5 286.4 289.3
46.0 8.0 32.9 13.1
Table 3. Contributions of the Different Types of Atomic Environments in High-Resolution O1s Spectra for PU Films Photopolymerized in Air (PUA-Air) and in Nitrogen (PUA-N2) PUA exposed under N2
PUA exposed in air
O atom environment
binding energy (eV)
atom %
binding energy (eV)
atom %
O-C OdC
532.4 532.6
44.8 55.2
532.3 533.8
50.3 49.7
The XPS data only allow one to compare the two materials after UV-curing, without any existing reference for the uncured liquid film. It is assumed that, in both situations, initiation and cross-linking polymerization proceed efficiently The comparison of the functional group distributions supports unambiguously the oxidation of the extreme surface of polyurethane films during UV irradiation under air, though not in a straightforward manner. The data indicate that, simultaneously to the free radical polymerization resulting from initiator photolysis, additional reactions take place in the presence of oxygen, producing (i) the decrease of C-C and C-H functions to a slight but significant extent, (ii) the disappearance of a larger amount of C-N bonds, (iii) the decrease of the C-O bonds and, (iv) a weak decrease of carbonyls of amide, ester, or acid groups. During irradiation in the presence of air, the additional reactions parallel to polymerization can be considered at first as similar to the long wavelength photooxidation mechanism of aliphatic polyurethanes,20-22 which proceed somewhat differently from the photodegradation of aromatic analogues.23,24 As a consequence of free radical generation by initiator photolysis, the polyurethane segments would be subject to hydrogen abstraction at the methylene or methine groups R to the nitrogen atom (pathway 1 of Scheme 1). Free radical quenching by dioxygen would then convert the C-N units into products of type I, with geminate heteroatoms O-C-N. The data of Table 2 do not support this pathway as the dominant process, since the mole fraction of C atoms with a binding energy of 289 eV is lower in the samples cured in air. The short wavelength photolytic mechanism undergone by aliphatic urethane functions is expected to yield products in a much better agreement with the present observations (Scheme 1, pathways 2 and 3). Direct homolysis of urethane functions at short wavelength irradiation (λ=254 nm) is indeed documented.15 It is worth it to notice that the UV source we have used (IST mercury lamp system) for photopolymerization presents a large spectral response between 250 and 450 nm compatible with such a degradative mechanism that operates on the surface, due to the high absorbance of the PUA formulation at the corresponding wavelength range. (20) (21) (22) (23) 1879. (24) 2631.
Wilhelm, C.; Gardette, J.-L. Polymer 1997, 38, 4019. Wilhelm, C.; Gardette, J.-L. Polymer 1998, 39, 5973. Posada, F.; Gardette, J.-L. Polym. Degrad. Stab. 2000, 70, 17. Hoyle, C. E.; Kim, K. J. J. Polym. Sci., Part A: Polym. Chem. 1985, 24(8), Hoyle, C. E.; Kim, K. J. J. Polym. Sci., Part A: Polym. Chem. 1986, 25(10),
Langmuir 2009, 25(17), 9831–9839
The scission is reported to occur at N-CH as well as at N-CdO bonds,20 presumably with little dependence on ambient atmosphere. Under nitrogen, quenching of the free radicals by O2 is avoided, favoring initiation by addition to acrylate monomers. We can consider at first the photolysis of urethane function at the alkyl-nitrogen bond (CH-N), as depicted for pathway 2 of Scheme 1. By addition of the carbon radical on the monomer, another C-N is produced, without specific signature in the analyzed XPS data. In the presence of oxygen, the primary free radical is expected to be quenched and converted into a peroxyl radical and further to a more stable hydroperoxide by hydrogen abstraction. This would yield to new CH-O bonds as in products of type II, at the expense of the starting CH-N groups, as is observed in the recorded XPS spectra. When photolysis occurs also at the urethane N-CO bond (pathway 3), the functions derived from the nitrogen-based free radical should not decrease the number of CH-N bonds, whereas decarboxylation is expected to follow readily the photolytic step. It would be difficult to explain the observed differences in function distribution depending on the type of gaseous atmosphere. Thus, the first cleavage mechanism leading to product II (pathway 2) is more consistent with the set of observations, and can be considered as the main oxidation route. The formation of hydroperoxide groups on the polyurethane surface is a very likely interpretation of the presence of lag time preceding the diffusion of iBoA into polyurethane films with high conversion ratios (π > 0.60). The presence of additional polar groups interacting by strong hydrogen bonds with the polar segments of the network can exert, during the first moments of immersion, a superficial barrier effect to the penetration of iBoA into the polyurethane films. This is illustrated by the presence of lag time at the beginning of the sorption kinetics of iBoA. It provides a coherent interpretation of the distorted Fickian sorption profiles we have observed. The increase of surface oxidation was not observed in IR spectra, likely because the depth of the domains probed by the ATR technique (some micrometers) is several orders of magnitude larger than the extreme surface analyzed by the XPS beam. The results are consistent with a surface oxidation mechanism taking place upon irradiation in air with a UV source including an intense short wavelength contribution. The various polar groups produced according to the reaction pathways depicted in Scheme 1 likely contribute to a temporary barrier effect through hydrogen bonding. Solvation by adsorbed iBoA would then gradually suppress the interactions during the observed lag time in the diffusion process. The relative stability of the hydroperoxides within the time scale of our experiments where sorption kinetics were monitored DOI: 10.1021/la9009338
9837
Article
Mougharbel et al.
Scheme 1. Direct Photolysis and Secondary Free Radical Reactions Affecting the Urethane Functions in Polyurethane Acrylates Submitted to Photopolymerization in Air under Short Wavelength UV Radiation
less than 12 h after sample preparation is supported by the absence of new carbonyls in the XPS spectra of samples polymerized in air. However, one can suspect, particularly for sorption kinetics conducted at 46 C, that some peroxide decomposition takes place at that time and might contribute to enhance the cross-link density at the surface of the films, thus reducing the value of the apparent diffusion coefficients associated with such samples. The direct photolysis of urethane also takes place when irradiation is conducted in an inert atmosphere, and probably contributes as an additional initiating process to the triggering of the free radical polymerization. To our knowledge, such a process is not documented in the literature. The present study emphasizes this short wavelength activation that contributes to an important extent to the efficient surface curing of polyurethane formulations. Indeed, the XPS data suggest that a large fraction of the CH-N bonds present in the probed domain have been converted as a consequence of direct photolysis.
Conclusion Typical Fickian behavior was observed for the sorption of isobornyl acrylate into UV-cured polyurethane acrylate films exhibiting a conversion level π ranging between 0.50 and 0.70. 9838 DOI: 10.1021/la9009338
The degree of swelling as well as the diffusion rates were shown to decrease with an increasing cross-link density in the cured material, in agreement with the general behavior of networks. Anomalous sorption profiles were obtained when studying samples photopolymerized in the presence of air, with lag time increasing with the radiation dose. Apparent diffusion coefficients were also affected by the irradiation conditions. For a given conversion degree (π=0.60), the apparent D value of iBoA into a PUA film photopolymerized in oxygen-free conditions is about twice as high as the one obtained in films cured under air. This phenomenon was interpreted by an additional resistance to mass transfer at the surface, as a consequence of hydroperoxides generated in the presence of air, that enhances the cross-link density by hydrogen bonding and/or additional covalent interchain bridges. Apparent diffusion coefficients evaluated in this work range between 10-15 and 2 10-14 m2 3 s-1. The quantification allows prediction of the diffusion effects with regard to time and to the depth of iBoA into substrates. The associated numerical model with experimentally determined data inputs provides a useful tool for understanding and designing in a controlled manner multilayered objects and coatings prepared by radiation-induced polymerization. The determination of the chemical composition at the film surface by XPS showed that photopolymerization in air induces a surperficial photooxidation, attributed to the Langmuir 2009, 25(17), 9831–9839
Mougharbel et al.
short wavelength photolysis of the urethane bond, revealed by a strong decrease of the C-N bonds without formation of additional carbonyl groups, when irradiation takes place in the presence of oxygen. This result suggests that the short wavelength irradiation of polyurethane contributes to a significant extent to the surface initiation of free radical polymerization, particularly when performed in an inert atmosphere. The high reactivity in cross-linking photopolymerization coined to acrylated polyurethane is worth it to be reinvestigated in the light of the high level of surface photolysis reported in the present paper.
Langmuir 2009, 25(17), 9831–9839
Article
Acknowledgment. Financial support by Region Champagne Ardenne (Programme PlAneT) is gratefully acknowledged. The authors express their thanks to Dr P. Nicolas for his interest and support to this work. Supporting Information Available: Representative spectra recorded for the surface characterization of UV-cured PUA films by attenuated total reflection infrared spectroscopy (ATR-IR) and X-ray photoelectron spectroscopy (XPS). This material is available free of charge via the Internet at http://pubs.acs.org.
DOI: 10.1021/la9009338
9839
