Simultaneous Monitoring of Property Development and Reactive Group Conversion in Photopolymer Systems

By Parag K. Shah, Department of Chemical and Biological Engineering, University of Colorado Boulder and Jeffrey W. Stansbury, Department of Craniofacial Biology, School of Dental Medicine, University of Colorado Denver

The performance of materials in the UV industry for applications such as coatings, adhesives, dental materials and 3D printing depends on the ultimate properties of these materials. The properties are, in turn, dependent on the formulation and processing conditions that lead to the final product from the starting materials. A typical photopolymer formulation consists of a monomer mixture, photoinitiator(s), pigments, particles and other additives. The monomers can be of different types depending on the curing chemistries utilized to polymerize them¹. The geometry and cure conditions also affect the processing of the material due to differences in kinetics and the exothermic reaction. Ultimately, however, the final properties for a particular system are significantly determined by the extent of cure, other variables being held constant^2-5.

Measurement of functional group conversion is important in determining performance under a given set of conditions. Polymerization is accompanied by volumetric contraction and the subsequent development of shrinkage stress in hard polymers, which can be detrimental to performance. Shrinkage is more or less linearly dependent on conversion, but properties such as modulus and shrinkage stress are nonlinear. All these properties can be tied together with the measurement of conversion. This means it is important to measure the conversion and also to monitor the dynamic progress of conversion to more accurately gain information about the network development.

An often overlooked factor, especially for thin films, is the volatility of the photoinitiator. Previous work in our lab has shown that photoinitiators, especially camphorquinone that is widely used in visible light curing, can volatilize from the surface of the monomer film. This is especially critical in conditions that require purging to remove solvent, as with in dental adhesives⁶. This loss of photoinitiator can lead to a gradient in conversion with lower conversion at the surface, even in the absence of oxygen inhibition. This can significantly affect the reliability and performance of such materials.

Fourier transform infrared spectroscopy (FT-IR) can be used to quantitatively measure conversion for a wide variety of functional groups. Mid-IR spectroscopy spans the wavelength region of about 2500 to 25000 nm and is widely used. It contains the fundamental absorbance bands for a variety of functional groups, including the carbon-carbon double bond in acrylate and methacrylate monomers. In general, the absorbance of different functional groups in the mid-IR spectral region is very high, restricting the measurement to very thin samples. In addition, many of the fundamental absorption bands from some of the functional groups can overlap in the fingerprint region, making it difficult to accurately assess conversion. Some of these issues can be overcome by using the near-IR spectral region that spans 2500 to 800 nm, which consists of overtone and combination bands of the fundamental absorbances. The absorptivity in the near-infrared (NIR) region is much lower than in the mid-IR, which makes this region amenable to measurement of thicker samples than in mid-IR [7]. NIR signals also can be transmitted efficiently via fiber optics and utilized in unpurged environments, making it convenient to analyze samples at locations away from the FT-IR instrument and to conduct other measurements simultaneously with FT-IR measurements. This enables the coordinated measurement of material properties along with conversion measurements, providing valuable insights into differences among materials and processing conditions while helping to make informed decisions on their utility.

As an example of monitoring shrinkage, modulus and shrinkage stress along with conversion, we have studied a system of model dental composites with varying degrees of filler loading⁸. Differences in the aluminosilicate glass loading level significantly affect conversion and, with the plethora of commercial dental restorative materials that have a wide variety of particle loading levels, it is important to be able to predict the influence of the particulate filler on performance that is highly dependent on conversion. This study helped to show how, when properties are measured and indexed with conversion measurement, they can be compared with each other – even though the measurements are performed in different instruments and with different sample geometries. A model dimethacrylate resin formulation consisting of 2,2-bis[4-(2-hydroxy-3-methacrylyloxypropoxy)phenyl] propane (BisGMA) and triethylene glycol dimethacrylate (TEGDMA) (70/30 wt/wt ratio) was used and 0.7 µm glass particles surface functionalized with a methacrylate silane were mixed with the resin at various loading levels (0, 20, 40, 60 and 70 wt%). Shrinkage stress was measured using a tensometer, a cantilever-based instrument (Figure 1). NIR measurements of double bond conversion were done using fiber optic cables connected to an FT-IR spectrometer. This allowed real-time simultaneous measurement of shrinkage stress and methacrylate conversion. Flexural measurements were done on rectangular specimens (2x2x10 mm), and shrinkage was measured with a linometer using disc-shaped specimens (1mm thickness, 6 mm diameter). The incident irradiance was kept the same for all the measurements. For the flexure samples, it was not possible to measure real-time IR using the fiber optic cables due to limitations in aligning them with the sample geometry. For the shrinkage samples, the NIR cables being used were of a small diameter (100 µm) and were not sufficient for a signal to pass through the highly filled samples. Figure 2 shows the results of the different measurements as a function of double bond conversion.

Quite a few salient features can be extracted out of studying the property development in this manner. Shrinkage is fairly linear with respect to conversion for most of the range but tends to slow as the limiting conversion is reached [Figure 2(a)]. The deceleration of shrinkage is observed when the samples start to vitrify. While most of the sample is in a glassy state, there remain pockets of uncured monomers and pendant methacrylate groups that can still undergo polymerization, contributing to the increase in conversion while the shrinkage development is restricted by the glassy network. Flexural modulus develops exponentially with respect to conversion. At the same value of conversion, flexural modulus increases with particle loading, but the rate of increase is more pronounced when particle loading increases to more than 60 wt%. At these higher loadings, interparticle interactions increase, leading to the faster increase in modulus. Shrinkage stress followed a similar pattern to the modulus in varying with respect to conversion. At high conversion levels, it is observed that – for the same irradiation conditions – as particle loading increases from the unfilled resin, the shrinkage stress decreases initially, possibly attributable to the reduction in shrinkage due to the reduced resin volume and also lower final conversion.

Lower conversion in highly filled composites can be caused by a variety of factors. Light is attenuated with increasing thickness due to scattering and absorption by the particular fillers that generally have a different refractive index as compared to the resin mixture⁹. These particulate fillers act as a heat sink, reducing the overall exotherm, which also adversely affects conversion. Surface-bound methacrylate silanes provide a mobility-restricted environment at the interface between filler particles and the resin matrix, which also likely contributes to a reduced conversion. After going through a minimum at about 40 wt% particle loading, the shrinkage stress increases again, in spite of the lower conversion. This can be attributed to the contribution from higher particle loading to the modulus. Measurement of conversion, along with following the evolution of material properties, can lead to significant insights into the interplay of the various properties, thus helping to optimize the curing process depending on the final application.

Modulus also can be related to real-time conversion by directly coupling a DMA with NIR fiber optics. Figure 3 represents the schematic of achieving this in a light-accessible DMA (Perkin Elmer 8000). This study examined the effect of irradiance and irradiation times on the modulus development during dark cure¹⁰. Samples (BisGMA/TEGDMA at 70:30 wt/wt ratio) were partially cured to about 30 to 40% conversion to achieve reasonable stiffness so that they could be held in the single cantilever accessory in the DMA. They were then aligned with the NIR fiber optic cables and also a UV light source (365 nm).

Samples were further cured for various times at a low irradiance or at high irradiance in the DMA while flexural modulus was measured, along with conversion (Figure 4). Figure 4(a) shows the modulus as a function of time and Figure 4(b) shows the modulus as a function of conversion. At the low irradiance level with short irradiation times, there is significant amount of dark cure but only modest increase in the modulus. The more extensive dark cure can be attributed to the higher mobility of radicals at the lower modulus. At higher irradiation times, the vitrification stage is reached and there is significant rise in the modulus with very little amount of dark cure, this time due to the very mobility of the radicals and reactive sites in the network. At the higher irradiance level, the modulus increase is much higher than the corresponding low intensity sample, and the limiting conversion is also increased. This is due to the greater exotherm attributable to the higher irradiance used for curing, which increases the mobility of the network and subsequently delays the vitrification. The increased conversion further raises the final modulus of the network. This study points to the need for monitoring temperature during polymerization and studying the effect of irradiance and irradiation times to optimize the final properties of the network.

Photorheometry is an important tool to characterize viscoelastic properties of monomers and polymeric materials. We have connected NIR fiber optics to a photorheometer to monitor real-time conversion and shear moduli at the same time (Figure 5). This technique also helps to locate the crossover point between the shear storage and loss moduli that can be taken as an indication of the gel point of the network. In this study, mixtures of BisGMA/TEGDMA (70/30 wt/wt) and 2,2-bis[4-(2-hydroxy-3-acrylyloxypropoxy)phenyl] propane/triethylene glycol diacrylate (BisGA/TEGDA -70/30 wt/wt) were studied separately to determine differences in modulus development with respect to conversion. Figure 6 shows that the BisGMA/TEGDMA combination displays slower kinetics than the BisGA/TEGDA mixture, as expected (Figure 7). The modulus development shows a faster rise in the modulus of the BisGMA/TEGDMA mixture. When the modulus is plotted with respect to conversion, it is apparent that the vitrification point for the methacrylate mixture is reached earlier than that for the acrylate mixture. This is expected, as the methacrylate polymerization results in a much stiffer network than that formed with the acrylate polymerization at ambient curing conditions. This study confirms existing knowledge, and the technique is powerful when there is a need to test new monomers to elucidate the way the polymer network and its properties develop in real time to enable better design ofnovel monomers.

 

Photoinitiators are critical components of a photopolymerization system. It is important to use a photoinitiator that matches with the light source being used and also to ensure compatibility with the monomer chemistries. The photoinitiator being used should have high quantum yield and efficiency. Studies of photoinitators generally are done using UV-Vis and in solvent or at very low concentrations in a monomer. This is not optimal, as the behavior of the photoinitiator depends on its environment and may differ in going from the solvent to the monomer. Monitoring of the photoinitiator during a polymerization brings along its own set of challenges, such as having to prevent curing of the resin formulation by the UV-Vis source and ensuring that the polymerizing light source does not interfere with the UV-Vis signal. It becomes necessary to use a very low intensity for the UV-Vis source to be able avoid polymerizing the sample, leading to loss in sensitivity of measurement. Separate measurements of UV-Vis and FT-IR have been done previously, but the ability to reproduce the exact irradiation and polymerization conditions when testing a sample in two different experiments is unreliable, leading to possible misrepresentation of data. Figure 8 shows the setup in which UV-Vis and NIR cables are placed orthogonal to each other and in the same plane in front of a cuvette containing the solution to be tested. This avoids issues related to making separate measurements. The light source for photopolymerization is placed directly above the cuvette.

As an example, ethoxylated bisphenol-A dimethacrylate (BisEMA) was mixed with camphorquinone, and the effect of using different amine reductants was tested using this coupled UV-Vis/FT-IR system¹¹. The amine reductants tested were ethyl 4-(dimethylamino)benzoate (EDMAB), methyldiethanolamine (MDEA) and N-phenylglycine (PG). Figure 9(a) shows the effect of different irradiances on polymerization and the simultaneous photobleaching of CQ, demonstrating that it is possible to track both processes at the same time. The effect of using the different amine reductants on polymerization rate can be seen in Figure 9(b). Quantum yields for CQ using the different amine reductants with the overlapping conversion can be seen in Figure 9(c), showing the reduction in quantum yield as the system gets stiffer and vitrifies. Key parameters related to the photopolymerization process can be calculated using such experiments (data not shown).

In conclusion, we have briefly shown a variety of setups in which FT-IR instruments can be coupled to other instruments that measure properties such as shrinkage, shrinkage stress, flexural modulus, complex moduli and UV-Vis spectra. Combining each of these property measurements becomes possible due to the simultaneous measurement of conversion, which can be used as an indexing parameter for a given system. These studies can help evaluate monomers for particular applications and also test irradiation conditions to be used to get the correct match of mechanical properties for the final application.

Parag Shah received his Bachelor’s and Master’s degrees in chemical engineering from the Indian Institute of Technology Bombay. He then moved to University of Colorado at Boulder to complete his PhD in chemical engineering. His PhD thesis work under Dr. Stansbury involved surface modification of nanoparticles using polymer brushes for shrinkage stress reduction in dental restorative materials. After receiving his PhD degree, Shah decided to continue working in the Stansbury lab as a post-doctoral research associate, where he is currently working on developing polymer nanoparticles for various applications and developing techniques to study the fundamentals behind the effect of these nanoparticles. His research interests include photopolymerization, surface science and nanocomposite applications.

Jeffery Stansbury received his undergraduate degree in chemistry from the University of Maryland, followed by a position in the Polymers Division at the National Institute of Standards and Technology (NIST). He returned to University of Maryland in College Park for graduate school in organic/polymer chemistry while continuing at NIST. His PhD thesis work involved free radical and cationic ring-opening polymerizations under William Bailey (graduated 1988). He remained at NIST a total of 21 years, working on a wide variety of polymeric biomaterials and materials characterization techniques. Stansbury then moved to the University of Colorado in 2000 to join the School of Dental Medicine to develop a biomaterials program there while also being appointed in the Department of Chemical and Biological Engineering in Boulder. In the dental school, he serves as vice-chair of the Department of Craniofacial Biology and senior associate dean for research. He has research laboratories on both the Boulder and Anschutz Medical Campuses with work focused on a variety of fundamental and applied areas involving dental materials, polymer networks, photopolymerization and bioengineering. He has been the recipient of the bronze medal award from the US Department of Commerce, the Souder Distinguished Scientist award from the International Association for Dental Research (IADR) and the New Inventor of the Year Award from the University of Colorado. He currently is president of the Dental Materials Group at IADR.

References

1. Carioscia JA, Stansbury JW, Bowman CN. Evaluation and Control of Thiol-ene/Thiol-epoxy Hybrid Networks. Polymer (Guildf) 2007;48:1526-32.
2. Howard B, Wilson ND, Newman SM, Pfeifer CS, Stansbury JW. Relationships between conversion, temperature and optical properties during composite photopolymerization. Acta Biomater 2010;6:2053-9.
3. Lovell LG, Lu H, Elliott JE, Stansbury JW, Bowman CN. The effect of cure rate on the mechanical properties of dental resins. Dent Mater 2001;17:504-11.
4. Stansbury JW. Dimethacrylate network formation and polymer property evolution as determined by the selection of monomers and curing conditions. Dent Mater 2012;28:13-22.
5. Stansbury JW, Trujillo-Lemon M, Lu H, Ding X, Lin Y, Ge J. Conversion-dependent shrinkage stress and strain in dental resins and composites. Dent Mater 2005;21:56-67.
6. Forman DL, McLeod RR, Shah PK, Stansbury JW. Evaporation of low-volatility components in polymeric dental resins. Dent Mater 2015;31:1090-9.
7. Stansbury JW, Dickens SH. Determination of double bond conversion in dental resins by near infrared spectroscopy. Dent Mater 2001;17:71-9.
8. Shah PK, Stansbury JW. Role of filler and functional group conversion in the evolution of properties in polymeric dental restoratives. Dent Mater 2014;30:586-93.
9. Emami N, Sjödahl M, Söderholm K-JM. How filler properties, filler fraction, sample thickness and light source affect light attenuation in particulate filled resin composites. Dent Mater 2005;21:721-30.
10. Abu-elenain DA, Lewis SH, Stansbury JW. Property evolution during vitrification of dimethacrylate photopolymer networks. Dent Mater 2013;29:1173-81.
11. Aguirre-Soto A, Hwang AT, Glugla D, Wydra JW, McLeod RR, Bowman CN, et al. Coupled UV-Vis/FT-NIR Spectroscopy for Kinetic Analysis of Multiple Reaction Steps in Polymerizations. Macromolecules 2015;48:6781-90.