Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2015 Nov 1.
Published in final edited form as: Dent Mater. 2014 Sep 6;30(11):1252–1262. doi: 10.1016/j.dental.2014.08.376

RAFT-mediated Control of Nanogel Structure and Reactivity: Chemical, Physical and Mechanical Properties of Monomer-dispersed Nanogel Compositions

JianCheng Liu a, Jeffrey W Stansbury a,b,*
PMCID: PMC4195848  NIHMSID: NIHMS627080  PMID: 25205366

Abstract

Objectives

This study examines how nanogel structure correlates with photopolymerization and key polymer properties upon addition of nanogels with latent reactivity into a monomer dispersant to produce polymer/polymer composites.

Methods

Two nanogels that retained RAFT functionality based on the synthetic approach were prepared to have different branching densities. These reactive nanogels were dispersed in triethylene glycol dimethacrylate at 0–40 wt%. Reaction kinetics, volumetric shrinkage and shrinkage stress associated with the photopolymerization of nanogel-modified formulations were measured in real time with mechanical properties of the polymers also evaluated. The basic structure of RAFT-derived nanogel particles was examined by the preparation of a separate nanogel constructed with degradable disulfide crosslinking groups. The model nanogel molecular weight and polydispersity were compared before and after degradation.

Results

Despite the controlled radical synthetic approach, the nanogels, which are composed of multiple interconnected, short primary chains presented relatively high polydispersity. Through addition of the reactive nanogels to a monomer that both infiltrates and disperses the nanogels, the photopolymerization rate was moderately reduced with the increase of nanogel loading levels. Volumetric shrinkage decreased proportionally with nanogel concentration; however, a greater than proportional reduction of polymerization-induced stress was observed. Mechanical properties, such as flexural strength, storage modulus were maintained at the same levels as the control resin for nanogel systems up to 40 wt%.

Significance

This study demonstrated that beyond the use of RAFT functionality to produce discrete nano-polymeric structures, the residual chain end groups are important to maintain reactivity and mechanical properties of nanogel-modified resin materials.

1. Introduction

Polymeric particles composed of multiple chains with relatively dense internal crosslinking or cyclization can be considered as highly branched polymers, or nanogels when the dimensions are below 100 nm. Dendrimers, represent a class of commonly studied highly branched polymers, which are monodisperse tree-like polymer with precise control over the polymer architecture. Significant attention has been directed towards the synthesis of dendrimers with different chemistries and properties for various applications1–3. However, it generally requires multiple steps to obtain the final dendrimer structure4 and the number of generations or final molecular weight is normally limited due to the de Gennes dense packing effect5. To overcome these limitations, one-pot synthesis has been applied to form hyperbranched polymers. A step-growth process is generally used with either single monomer (e.g., AB2) or multi-monomer (e.g., A2 + B3) methodology to generate imperfect (compared to dendrimers) hyperbranched structures due to the non-uniform growth of branching points6. Sherrington and coworkers7 developed a facile method to polymerize methyl methacrylate in the presence of modest amounts of a dimethacrylate crosslinker based on a free radical chain-growth mechanism. Polymeric nanoparticles were obtained with chain transfer agent and solvent used to prevent macrogelation. High molecular weight, multi-chain polymeric structures were formed by these reactions. Different from typical polycondensation reactions, branch points in these nanogels are based on cyclization and crosslinking introduced by the reaction of pendent vinyl groups with radicals either from the same molecule or another propagating polymer. As a simple technique with versatile monomer choices, this method has been applied to make a variety of nanogels with different monomers by either free radical8 or controlled radical polymerization methods9,10. Due to the low concentration of crosslinker (typically 1 – 2 mol% of monomers), limited crosslinks are formed for each individual chain so there is significant degree of linear polymer formation during nanogel synthesis. It was demonstrated that the nanogel synthesis often led to significant amounts of linear species formation with either free radical7 or controlled radical polymerization10,11. Recently, our group has extended the synthesis of nanogel structures through similar approaches but with the use of significantly higher concentrations of crosslinker in the system12. This generated nanogel structures with substantially higher crosslinking/branching density ([crosslinker]/[monomer] = 30/70 mole ratio) than other approaches resulting in high numbers of branching points per nanogel. In order to prevent macrogelation, relatively high concentrations of chain transfer agent were applied to limit primary chain length. An additional study with up to 50 mol% of crosslinker in the system was demonstrated for nanogel synthesis through a similar approach using higher solvent concentration to enhance cyclization and suppress macrogelation13.

These nanogel materials have been applied to form discontinuous and co-continuous nanocomposite morphologies with resin matrices as a means to modify polymer network properties for alleviation of shrinkage stress during photopolymerization. During polymerization, volumetric shrinkage14 is typically observed due to the decrease of free volume during the transition from a mobile liquid monomer to a constrained polymer. Particularly in crosslinking bulk polymerizations, polymerization-induced stress15 starts to accumulate beyond the gel point with a decrease of stress relaxation as modulus of the polymer network develops. Shrinkage stress can affect material performance and lead to premature failure based on the development of interfacial or bulk defects. A variety of methods have been applied to reduce shrinkage and stress including development of new monomers16, adoption of step-growth curing mechanisms17 and addition of secondary fillers18. Incorporation of inorganic fillers has a demonstrated ability of shrinkage reduction determined by volume fraction but depending on the effect of the filler on bulk modulus, higher residual shrinkage stress can result for these composite materials19. Nanogel, as an organic filler, was recently developed and showed significant potential for shrinkage stress reduction12,20. The utilization of nanogels swollen by monomers, has the advantage of forming well-controlled physically interpenetrated and covalently interconnected polymer networks. During the secondary polymerization between nanogel and matrix monomers, both the chemical and physical crosslinking between nanogel and resin are important for the mechanical performance of the final materials. In order to enhance crosslinks between nanogel and matrix for reinforcing the mechanical properties, either further modification of nanogel with reactive groups12 or increasing nanogel crosslinking density13 is required. Further modification requires additional reaction to introduce reactive groups that may complicate the nanogel synthesis procedure. Special attention is usually required for the synthesis of nanogels with very high branching densities due to the tendency for macrogelation.

As a part of this study, reactive nanogels were synthesized with a disulfide-containing crosslinker and monomethacrylate in the presence of a styrene dimer-based RAFT (reversible addition-fragmentation chain-transfer) agent. The internal disulfide connections allow subsequent intentional deconstruction of the nanogels as a means to probe their basic structure; however, these same linkages introduce difficulties and instabilities during the secondary photopolymerization of nanogel-modified resin formulations. As such, the disulfide dimethacrylate was replaced with a non-degradable urethane dimethacrylate crosslinker in the RAFT-based nanogel synthetic approach to provide a separate series of branched nanogels for study as potential shrinkage and stress-limiting resin additives.

2. Materials and Methods

Materials

Isobornyl methacrylate (IBMA), benzyl methacrylate (BzMA), 2,4-diphenyl-4-methyl-1-pentene (also known as α-methylstyrene dimer, AMSD), 2-hydroxyethyl disulfide, methacryloyl chloride, triethylamine (TEA), 4-dimethylaminopyridine (DMAP), dithiothreitol (DTT), azobisisobutyronitrile (AIBN) and triethylene glycol dimethacrylate (TEGDMA) were obtained from Sigma-Aldrich. 2,2-Dimethoxy-2-phenylacetophenone (DMPA) was obtained from BASF. The structures of the monomers and AMSD are shown as Figure 1. Urethane dimethacrylate (UDMA) was obtained from Esstech. (Essington, PA). The inhibitor was removed from IBMA by treatment with activated basic alumina (Brockman I, Sigma-Aldrich). AIBN was recrystallized from methanol. All other materials were used as received.

Figure 1.

Figure 1

Open in a new tab

Molecular structures for methacrylate monomers and AMSD RAFT agent.

Degradable disulfide dimethacrylate synthesis

Disulfide dimethacrylate (DSDMA) was synthesized according to a prior literature report21. Methacryloyl chloride (5.64 g, 36.5 mmol), TEA (11.09 g, 109.6 mmol) and DMAP (0.45 g, 3.65 mmol) was dissolved in 150 ml anhydrous dichloromethane (DCM). The solution was purged with nitrogen gas and cooled by ice bath for 30 min. 2-Hydroxyethyl disulfide (5.64 g, 36.5 mmol) was added to the solution in a dropwise fashion. The reaction mixture was allowed to warm to room temperature under nitrogen and held for 24 h. Afterwards, the reaction was quenched by addition of deionized water and subsequently washed twice with a DI H2O, Na2CO3 (10%), brine solution. The organic phase was dried over Na2SO4 and the solvent removed under vacuum. A yellow oil was obtained in 80.2% yield. The monomer structure was verified by 1H NMR (400 MHz, chloroform-d): δ (ppm) 6.15 – 6.14 (m, 1H), 5.69 – 5.54 (m, 1H), 4.43 (t, J = 6.6 Hz, 2H), 2.99 (t, J = 6.6 Hz, 2H), 1.96 (dd, J = 1.6, 1.0 Hz, 3H).

Polymer and nanogel preparation

A linear polymer of isobornyl methacrylate (PIBMA) was synthesized in the presence of 15 mol% of AMSD (relative to IBMA content) with 1 mol% AIBN in ethyl acetate (1:1 by weight) as solvent. Polymerization was carried out at 80 °C for 24 h in a nitrogen environment with 82.3% vinyl group conversion as determined by FT-IR. The homopolymer sample was isolated by precipitation from an excess of methanol with the residual solvent removed under vacuum to obtain the final oligomeric material as a powdery solid.

A linear block copolymer was then synthesized from the reaction of the isolated PIBMA with benzyl methacrylate (BzMA) using AIBN (0.5 mol% based on IBMA repeat units). The molar ratio of BzMA monomer to IBMA units in PIBMA was 2:1. A four-fold excess of ethyl acetate was used and the reaction was conducted for 12 h at 80 °C under nitrogen. With the BzMA vinyl group conversion at 80.4%, the copolymer was isolated by precipitation from methanol to yield a powdery solid material.

A degradable nanogel was prepared by the copolymerization of DSDMA and IBMA (30:70 mole ratio) with 15 mol% AMSD and 1 mol% AIBN (related to total monomers) in a four-fold excess of ethyl acetate. Polymerization was carried out at 80 °C for 24 h in a nitrogen environment with 81.4% final vinyl group conversion. Samples were taken at various intermediate time points with the nanogels being isolated by precipitation from an excess of methanol and then dried under vacuum to obtain a pale yellow powder.

Disulfide nanogel degradation

Nanogel prepared with disulfide dimethacrylate (0.10 g, about 0.15 mmol disulfide functionality) was dissolved in inhibitor-free tetrahydrofuran (THF; 6 ml). DTT (0.15 g, 1.0 mmol) was added under a nitrogen environment. The reaction was allowed to proceed at room temperature and aliquots of samples were withdrawn periodically and immediately diluted in THF for GPC analysis.

Non-degradable nanogel synthesis

Two additional nanogels were synthesized without the cleavable disulfide crosslinks. UDMA and IBMA were mixed at 10:90 or 30:70 molar ratios and 15 mol% of AMSD and 1 mol% of AIBN (relative to monomers) was added with a four-fold excess of ethyl acetate. Polymerization was continued for 4 h at 80 °C under a nitrogen environment. Samples were precipitated from an excess of methanol and dried under vacuum to obtain white powdery materials. The conversion of vinyl groups was 48.4 and 38.5 %, respectively, for RN19 and RN37 (RN refers to reactive nanogel, and the numbers indicate the crosslinker to monomer molar ratios).

RN19 and RN37 were added to TEGDMA in mass ratios of 10 % to 40 % in 10 % increments. DMPA as photoinitiator (0.5 wt%, relative to TEGDMA) was added to each sample. Unfilled TEGDMA with DMPA was used as control system. Nanogels were fully dispersed into TEGDMA and formed stable, transparent mixtures following mechanical agitation.

Polymer Structural Analysis

GPC (Waters 515) was employed for the analysis of linear polymer and nanogel molecular weight and polydispersity index (PDI) based on polystyrene calibration standards. Tetrahydrofuran was used as eluent at a flow rate of 1 mL/min at 35 °C. A 400 MHz NMR (Bruker Avance-III 400) was used for collection of 1H spectra of the monomers as well as polymeric structures in either CDCl3 or CD2Cl2.

Photopolymerization Kinetics

Real-time near-infrared (NIR) spectroscopy was applied to monitor the reaction kinetics of the TEGDMA-dispersed nanogel formulations during the photopolymerization process. The methacrylate group (=CH2) has an overtone absorbance peak at 6165 cm−1, which was monitored throughout the reaction. All specimens (n = 3) were irradiated under 365 (± 10) nm UV light with a 10 mW/cm2 irradiation intensity as determined by radiometer at the sample surface position.

Volumetric Polymerization Shrinkage

A non-contact linear variable differential transducer (LVDT)-based linometer (Academic Center for Dentistry Amsterdam, The Netherlands) was used for real-time volumetric shrinkage testing. Samples (compressed cylindrical shape of 3 – 4 mm diameter and 1 mm thickness) were placed on an aluminum disc with a glass slide covering the top. A thin film of applied grease on the aluminum disc and glass slide substrates allowed the sample to shrink freely in the horizontal as well as vertical dimensions. The transducer underneath the aluminum disc monitored the z-axis linear displacement of the sample during polymerization that was then used to calculate the real-time volumetric shrinkage. All specimens (n = 3) were irradiated under 365 (± 10) nm UV light with 10 mW/cm2 irradiant intensity with limiting shrinkage data reported at the end of the 10 min illumination interval.

Shrinkage Stress Measurement

A tensometer (Volpe Research Center, American Dental Association, Gaithersburg, MD) was utilized for measurement of real-time polymerization shrinkage stress22. All cylindrical samples (6 mm diameter and 1 mm thickness) were placed between two polished, silane-methacrylate-treated quartz rod surfaces (6 mm diameter) where the top rod was attached by a collet connected to the stainless steel beam of the cantilever tensometer. The bottom glass rod, which was fixed to the tensometer base with a second collet, was also used as a light guide extension for photocuring with an adapter connecting the tip of the UV light guide with the rod. Specimens (n = 3) were irradiated under 365 (± 10) nm UV light at intensity of 10 mW/cm2 (measured by radiometer at the exit of the glass rod) for 15 min. The deflection of the calibrated cantilever beam was monitored by a LVDT detector with displacement converted to the shrinkage stress output.

Mechanical Property Testing

Flexural mechanical properties were collected by a universal testing machine (Mini-Bionix II, MTS, Eden Prairie, MN). Specimens with dimensions of 25 × 2 × 2 mm were cured under 365 (± 10) nm light irradiation for 25 min with 5 mW/cm2 intensity. Flexural modulus and strength were determined in three-point bending tests on a 20 mm span at a crosshead speed of 1 mm/min after approximately 48 h storage in the dark at room temperature. Vinyl group conversion for the mechanical testing samples (n = 3) was measured with NIR shortly after the photopolymerization.

A dynamic mechanical analyzer (DMA, Perkin Elmer 8000, Waltham, MA, USA) was employed to measure the bulk nanogel glass transition temperature (Tg) values by sandwiching 10 mg of nanogel powder in aluminum pockets. A single cantilever cyclic displacement at 1 Hz was applied while the sample (n = 3) was heated to 180 °C and then cooled back to room temperature at 2 °C/min with tan δ data collected during the cooling process. The Tg was assigned as the maximum in the tan δ output. This technique has been validated using film and powdered polymer standards of known Tg.

A TA Q800 DMA was applied for measuring the storage moduli (G’) and tan δ behavior of fully-cured nanogel-modified compositions and the TEGDMA homopolymer control. Samples were prepared under the same curing conditions as employed with the flexural strength testing; however with different specimen dimensions of 25 × 3 × 1 mm. The photocured specimens (n = 3) were post-cured at 160 °C overnight to reach essentially full conversion (> 98%) to avoid any post-cure effects during the thermal analysis testing, which utilized a tensile test mode with a strain rate and frequency of 0.02 % and 1 Hz, respectively during the temperature scan from 0 °C to 220 °C at 2 °C/min.

3. Results

AMSD RAFT behavior

AMSD was applied initially for linear polymer synthesis. The PIBMA synthesized with 15 mol% AMSD showed a number average molecular weight from GPC (Mn) of 4.52E+3 Da with a PDI of 1.89, which corresponds to approximately 20 repeat units per polymer chain on average. The efficiency of AMSD as a chain transfer agent (CTA) is around 33.3% and the relatively high PDI indicates only a modest degree of structural control over final linear polymers. By assuming each polymer contained one AMSD end group, the molecular weight was calculated to be 4.0E+3 from the NMR data, which is in good agreement with the GPC results.

To further test the latent end group reactivity, the isolated PIBMA prepared with 15 mol% AMSD was further reacted with BzMA, which has been demonstrated to offer good reactivity and narrow polydispersity during RAFT polymerization. The BzMA structure also facilitates analysis by NMR due to the chemical shift differences of the benzyl groups from that of the alkyl groups in IBMA. From the NMR spectra of the block copolymer (Figure 2), the molar ratio of the BzMA segment relative to the IBMA segment is 1.77, which was calculated by a comparison of the methylene group peaks (5.0 ppm for BzMA and 4.5 ppm for IBMA, respectively). This indicates that 35 BzMA repeat units on average were added to the PIBMA chains during the block copolymer synthesis. This provides a molecular weight of 1.08E+4, which again is in good agreement with the Mn of 1.03E+4 Da (PDI: 2.34) from the GPC data (Figure 3).

Figure 2.

Figure 2

Open in a new tab

1H-NMR spectrum of PIBMA-b-PBzMA in deuterated DCM.

Figure 3.

Figure 3

Open in a new tab

GPC traces of PIBMA and PIBMA-b-PBzMA.

Based on the reasonable RAFT character demonstrated with the AIBN/AMSD initiated block copolymers, the same approach was undertaken using mixtures of monovinyl and divinyl monomers capable of producing nanogel structures. The synthesized nanogel from IBMA and DSDMA had a Mn of 8.35E+3, Mw of 4.4E+4 and PDI of 5.28.

Non-cleavable nanogels

To avoid the concerns of disulfide color, stability and its contribution to a relatively flexible spacer in the crosslinker, UDMA was used as an alternative divinyl monomer to form a separate series of nanogels with two different branching densities by changing the crosslinker to monovinyl monomer ratio. The resulting nanogel from 10 mol% UDMA and 90 mol% IBMA (RN19) had a Mw of 5.5E+04, PDI of 1.90 and 3.9 nm hydrodynamic radius. The more densely branched nanogel from 30 mol% UDMA and 70 mol% IBMA (RN37) had a Mw of 3.2E+05, PDI of 2.66 and Rh of 5.8 nm. For both non-degradable nanogels, the same 15 mol% of AMSD was used and if assumptions from the initial work with the linear PIBMA from degradable nanogel are applied, this predicts an average of 31.5 and 41.0 chains based on Mw data for RN19 and RN37 nanogel particles, respectively.

Dispersion of the two UDMA/IBMA nanogels into TEGDMA resin at 0 – 40 wt% loading levels provided transparent, colorless samples with viscosities that increased exponentially as a function of the nanogel content but still provided fluid resins. The UV-activated photopolymerization kinetics based on real-time NIR spectroscopic analyses were collected for the TEGDMA control and these nanogel-modified composite systems. As shown in Figure 4, the control polymerized rapidly and reached a plateau in less than 30 s and then continued to gradually increase towards a limiting conversion of just above 80 %. By the incorporation of either RN19 or RN37, the rate of polymerization (Rp) decreased gradually with increasing nanogel loading. At the 40 wt% nanogel loading level, considerably lower conversion was achieved compared with the control mainly due to the disappearance of the autoacceleration behavior or decreased exotherm generated during the slower reaction. The limiting conversion reached during ambient temperature photopolymerization was 63 % and 68 % for the 40 wt% loaded RN19 and RN37 systems.

Figure 4.

Figure 4

Open in a new tab

Kinetic profiles for RN19 and RN37 modified TEGDMA photopolymerizations. Conditions: 0.5 wt% DMPA; 365 nm light at 10 mW/cm2 with irradiation starting at 0 min.

From the three-point bending tests (Table 1), the unmodified control resin reached about 82 % conversion with flexural strength of 73.9 MPa and flexural modulus of 2.08 GPa. As previously mentioned, the final conversion gradually decreased with an increase in the nanogel content; however, despite the lower conversion levels, all the nanogel-modified samples showed statistically equivalent or improved mechanical strength and modulus values compared with the control.

Table 1.

Flexural strength, modulus and sample conversion data for RN19 and RN37 modified systems

Composite Formula Nanogel Loading (wt%) Flexural Strength (MPa) Flexural Modulus (GPa) Final Conversion (%)
Control 0 73.9 (9.3) 2.08 (0.03) 82.2 (3.0)
RN19 10 76.2 (5.3) 2.02 (0.38) 82.2 (1.0)
20 80.0 (6.2) 2.17 (0.12) 76.6 (1.5)
30 76.3 (2.6) 2.31 (0.18) 77.6 (1.7)
40 69.7 (12.8) 2.35 (0.37) 74.0 (2.1)
RN37 10 82.9 (8.5) 2.22 (0.11) 81.9 (0.8)
20 77.8 (7.1) 2.23 (0.06) 76.4 (2.3)
30 77.2 (15.0) 2.42 (0.12) 76.9 (1.3)
40 74.2 (4.2) 2.28 (0.10) 76.7 (0.1)
Open in a new tab

From DMA scans, all post-cured nanogel-modified materials showed unimodal distribution with maximum Tg values around 150 – 160 °C (Figure 5). More homogenous structures were formed with the increase of nanogel content indicated by a reduction of full-width-at-half-maximum values of tan δ peaks. The storage moduli (G’) of these materials evaluated by DMA have been determined and no significant change was observed for the control and all the nanogel systems except for the 40 % RN19 material (Figure 6).

Figure 5.

Figure 5

Open in a new tab

Tan δ data for RN19 and RN37 modified systems. All samples were prepared under 365 nm with 5 mW/cm2 for 15 min and then post-cured at 160 °C overnight to reach high conversion (>98%).

Figure 6.

Figure 6

Open in a new tab

Storage moduli (G′) for RN19 and RN37 modified systems. All samples were prepared under 365 nm with 5 mW/cm2 for 15 min and then post-cured at 160 °C overnight to reach high conversion (>98%).

A linear trend was found for the volumetric shrinkage for the nanogel-modified photopolymers up to the 40 wt% loading level (Figure 7). The 40 wt% loading of the RN37 nanogel showed a 33 ± 5 vol% reduction while a 38 ± 3 vol% decrease was noted for RN19 at the same 40 wt% loading.

Figure 7.

Figure 7

Open in a new tab

Volumetric shrinkage in relation to nanogel content (0 – 40 wt%) for RN19 and RN37 systems. The solid line plots represented linear regression fitting results of the actual shrinkage data. Conditions: 0.5 wt% of DMPA; irradiation with 365 nm light at 10 mW/cm2 for 10 min.

Shrinkage stress data was also measured in real-time during polymerization process. For the TEGDMA control, stress rapidly increased reaching a final stress value of about 3.2 MPa under the standardized curing conditions used here (Figure 8). By incorporation of RN19 or RN37 in the monomer, significant progressive reductions in stress were achieved as the nanogel concentrations increased. Unlike the effects on volumetric shrinkage, the stress reduction levels are proportionally greater than the corresponding nanogel weight fractions. For example, at the 40 wt% loading level, stress was reduced about 53 % and 56 % for the RN19 and RN37 samples, respectively.

Figure 8.

Figure 8

Open in a new tab

Shrinkage stress results for RN19 and RN37 modified systems. Light irradiation started at 0 min. Conditions: 0.5 wt% DMPA; continuous exposure to 365 nm light at 10 mW/cm2.

4. Discussion

AMSD has been used previously as an addition-fragmentation chain transfer agent for free radical polymerization23. AMSD has shown RAFT behavior due to the stabilized tertiary radical formed by the reaction of propagating radicals with the unsaturated double bond in AMSD (Scheme 1). The intermediate species can break either of the bonds adjacent to the tertiary radical to generate a highly reactive radical, which is capable of further propagation with monomers. Compared to the classic RAFT agent, like dithioester or trithiocarbonate, the tertiary radical formed from AMSD is less stable and more prone to further propagation or termination. Because of this, it doesn’t offer the precise control over molecular weight or polydispersity characteristic of typical RAFT agents. However, since thiol-based RAFT agents are typically colored, subsequent removal of the end groups may be necessary in many applications. Thiol-based RAFT agents as end groups are also not stable such that radicals can be generated under exposure to UV light24,25. On the other hand, AMSD is colorless and polymers generated in its presence don’t absorb light at wavelengths of 365 nm or higher, which among other advantages, avoids potential competition for light absorption during photopolymerizations.

The unimodal peak for the block copolymer (PIBMA-b-PBzMA) with a progressive increase in the Mn without significantly altering the PDI clearly shows that the AMSD residue at the chain ends remained active and displayed continued RAFT behavior during the secondary polymerization stage. During nanogel synthesis, when relatively low concentrations of crosslinker (1–2 mol% of overall monomer) have been used to prepare discrete nanogel structures, extended individual chain lengths can be accommodated. Macrogelation can be prevented in these cases with modest amounts of CTA or in some cases, just based on cyclization reactions promoted by the solution polymerization process. However, nanogels formed in this manner retain a significant amount of linear polymer character instead of highly branched nanogels due to the limited number or absence of crosslinking units in each chain. We have previously shown that a high crosslinker content was helpful to increase the physical entanglement between a nanogel and a secondary polymer matrix13. In this study, 30 mol% DSDMA with 70 mol% IBMA was applied to synthesize a highly branched/cyclized nanogel. The use of AMSD at 15 mol% was designed to provide reasonable control of the primary chain length as a means to avoid macrogelation during nanogel synthesis. The much higher PDI of this nanogel compared with that previously observed for PIBMA indicated a general lack of control over the size distribution of the multi-chain nanogel structures. The main peak in the GPC curve is probably related to the high molecular weight species – highly branched polymers or nanogels (Figure 9). Two shoulders appeared at lower molecular weight region, which might be related to the less branched polymers not captured by the more mature nanogels. To further examine the primary chain length involved in the nanogel, cleavage of disulfide bond in the crosslinker units was applied to degrade the nanogel to linear chain components. DDT has been commonly used for the cleavage of disulfide bonds due to the fast and efficient thiol exchange reactions10. The primary, high molecular weight peak in the nanogel GPC trace disappeared after a 2 h reaction with DDT to yield a Mn of 7.36E+3, Mw of 2.05E+4 and PDI of 2.79. After 48 h of reaction, the residual polymer had a Mn of 3.50E+3, Mw of 6.98E+3 and PDI of 1.99. It appears reasonable to assume that most if not all of the disulfide linkages were degraded by this point to generate linear polymers since the mass change effectively stopped at this point and the observed PDI matched well with that obtained in the homopolymerization of IBMA. These results indicate there were about 15 repeat units per primary chain with an average of 4.8 crosslinker units associated with each chain based on the final composition of the two monomer repeating units in nanogel (DSDMA:IBMA = 1:2.1) obtained from NMR data (not shown), which was similar to the molar feed ratio (DSDMA:IBMA = 3:7). By comparison of the Mn and Mw of the nanogel with the linear species, most of the nanogel ranged between 2.4 and 6.3 primary chains on average. It should be noted that disulfide compounds undergo inefficient chain transfer during free radical polymerization26 so some proportion of the disulfide crosslinkers may have cleaved during the nanogel synthesis process. Additionally, molecular weight data gathered from single-detection GPC is known to be lower than the actual value when analyzing multi-branched polymer structures due to the calibration from linear polymer standards. Therefore, this method likely underestimates the nanogel molecular weight and total number of primary chains.

Figure 9.

Figure 9

Open in a new tab

GPC traces for the polymer structures during degradation of disulfide nanogel from refractive index detector. DTT was added to the nanogel to cleave the disulfide bonds.

The reduction of rate of polymerization of RN19 and RN37 systems compared to the control is attributed mainly to a decrease in the active radical concentration due to reaction of the AMSD-based end groups with propagating radicals to form relatively stable tertiary radicals (Scheme 1). With the increase of nanogel loading, the overall concentration of the RAFT functionality on the nanogel chain ends also is increased so further reductions of Rp were observed for both RN19 and RN37 systems with somewhat lower final conversion values observed. Another factor influencing the final conversion is associated with earlier vitrification due to the high Tg contributed mainly by IBMA in the nanogel structures. A less than 5 % reduction in final conversion was reported before with similar nanogel additives lacking the AMSD reactive groups13. However, considering traditional thiol-based RAFT agents applied to photopolymerizations that significantly restrict reaction rates27, the Rp results here can be considered to be only modestly affected with all samples reaching reaction plateau conversions within 1 – 2 min.

Previous work has demonstrated the flexural strength of the final nanogel composite system is affected by nanogel Tg, branching density and the presence of reactive groups that allow covalent attachment between the nanogel and polymer matrix. Nanogel branching is important to allow nanoparticle swelling with monomer as well as to form physical crosslinking entanglements with the resin matrix. Chemical crosslinking can take place when reactive groups are present in the nanogel structure either as residual pendant groups arising from the divinyl monomer13 or through the introduction of polymerizable groups12,20 as the final step of the nanogel synthesis. High branching density within nanogel particles is required to maintain a similar level of flexural strength compared with the matrix polymer control, especially at high nanogel loading level (40 – 50 wt%) relative to low branching density nanogel. However, nanogels with high branching densities are more prone to macrogelation during their preparation. So enhancing the chemical reaction between the nanogel and matrix may provide an alternate route to maintain overall mechanical properties. Even though the RN19 and RN37 nanogels lack reactive methacrylate groups, chemical crosslinking can be achieved between the bulk matrix and nanogel through the residual reactive RAFT end groups. By functionalization of the nanogel chain ends with RAFT groups, it was possible to maintain the bulk mechanical properties by reliance on chemical crosslinking with the matrix network despite the relatively limited branching density associated with the RN19 nanogel. Functionalization of nanogels with secondary methacrylate groups has shown similar mechanical property behavior but using RAFT agent as the means to control primary chain length is advantageous since it avoids the need for thiol or other CTA during the nanogel synthesis and it requires no additional steps to produce reactive nanogels. A further potential advantage associated with the RAFT-based procedure, although not detailed here, involves the ability to modify nanogel size or more importantly to produce a variety of compositionally and structurally unique nanoparticles based on the abbreviated block copolymer approach as demonstrated with IBMA and BzMA in this study.

Dynamic mechanical analysis was applied to further characterize the mechanical properties and structure of these polymeric composite materials. Nanogel-modified resins below the nanogel dense packing state normally display two Tgs that arise from the different phases present. In this situation, one phase is associated with the bulk matrix and other is derived from the matrix-infiltrated nanogel regions. A single hybrid polymeric Tg is normally observed when the nanogel loading reaches confluence, which occurs at considerably lower loading levels compared with the dense packing limit associated with solid fillers due to the significant nanogel swelling based on monomer infiltration and allowed overlap between the globular, short-chain nanogel structures. Beyond confluence loading, the pure matrix phase does not exist. Prior experience has shown that this typically requires ~ 40 wt% nanogel loading, but also depends on the nanogel size and affinity for the matrix monomer13. For RN19 and RN37, different from non-reactive nanogel materials, the RAFT functionality within the nanogels can directly polymerize with infiltrated monomer as well as external matrix monomer in a continuous phase. This eliminates chain end mobility, which potentially can increase the Tg of the nanogel phase. The cured composite materials showed a single Tg (Figure 5) no matter the nanogel concentration in these materials indicating the monomer infiltrated nanogel domains effectively present the same Tg as the unaltered bulk matrix (in the thermally post-cured state). For the bulk nanogel particles, RN19 had a Tg of 122.5 ± 0.4 °C and RN37 presented a virtually identical Tg of 121.9 ± 1.1 °C despite the differences in branching density. The pure TEGDMA control homopolymer network has a Tg of 158.2 ± 3.2 °C after the thermal post-cure treatment. Chain ends in the nanogel structure have a greater degree of freedom than do the chain backbone or crosslinking sections so an isolated nanogel would be expected to have a lower Tg than that contributed by the corresponding crosslinked polymer region in which the chain ends are extended as part of a dense network structure. By copolymerization of UDMA and IBMA at 10:90 and 30:70 molar ratio without solvent and RAFT agents, the macrogelled polymer network that formed produced Tgs of 156.3 ± 0.3 °C and 159.7 ± 0.2 °C, respectively, after post-thermal treatment to achieve > 98% vinyl group conversion. These Tgs are very close to that of the TEGDMA homopolymer control. Therefore, it is reasonable to expect that the Tg of the fully cured resin-infused nanogel can increase about 30 °C to match that of the fully cured polymer matrix. With the increase in the nanogel content, the final materials are effectively more homogeneous as shown by the reduction of the half width of the tan δ peaks. It should be emphasized that these results concerning polymeric Tg refer to thermally post-cured materials. In the ambient photopolymerization process, the as formed polymer Tg is generally approximated as Tcure + 30 °C28, which correlates to a Tg of ~50–60 °C for methacrylate-based network polymers photocured at room temperature and experiencing a modest exotherm. Therefore, even allowing for somewhat lower degrees of conversion for monomer located within nanogel domains, it is reasonable to expect that the ambient-formed Tg of the nanogel regions may exceed that of the matrix polymer prior to any post-cure thermal treatments. The reduction of storage modulus for RN19 40% sample could be due to the incomplete reaction between nanogel RAFT groups with the matrix that can lead to a lower modulus and Tg. RN19 also has a lower internal crosslinking inside the nanogel structures than RN37, so the ability to form physical entanglement with the resin network would be more limited compared to the RN37 nanogel samples.

Volumetric shrinkage during polymerization is related to the initial functional group concentration and conversion achieved. By replacement of reactive monomers with relatively inert nanogels, it has been demonstrated that volumetric shrinkage can be reduced in a proportional manner with respect to the nanogel volume concentration. The weight and volume fractions of these nanogel additives in monomer is similar12. The RAFT end groups involved here likely can be neglected in terms of their effect on nanogel density. So it is expected that these reactive nanogels will follow similar behavior with regard to their effects of polymerization-based shrinkage. By fitting the shrinkage data with linear regression model (Figure 7), the shrinkage had a proportional reduction with RN19 samples while slightly less than proportional with the nanogel weight fraction in the RN37 samples. The lower overall degree of conversion of the nanogel-containing polymers would also factor in to these shrinkage results but that would only enhance the shrinkage reduction effect of the nanogel additives.

Stress developed as a consequence of polymerization is related to restricted shrinkage stain and elastic modulus. This implies a reduction in dynamic polymerization stress can be expected with nanogel additives since the final polymeric shrinkage is reduced here while the bulk modulus is maintained. The RAFT functionality at the chain ends is expected to keep secondary chain lengths more uniform than would be the case if nanogels without RAFT groups were used and this factor may enhance homogeneity as the nanogel content is increased, which might be important for the final mechanical properties. This may explain why even though lower degrees of conversion were achieved with these nanogel-based systems, bulk mechanical properties were maintained or improved compared with the control, which is not an expected result in typical glassy dimethacrylate networks.

Even though this preliminary study was based on UV light (365 nm) to trigger photopolymerization with a type I photoinitiator (DMPA), the use of visible-light initiation systems (e.g. camphorquinone/amine) would be expected to achieve similar physical and mechanical property results since the AMSD-based nanogel additives are reactive with propagating radicals from any source as demonstrated by the nanogel synthesis using AIBN as a thermal free radical initiator.

5. Conclusions

Nanogels composed of significant concentrations of crosslinker were prepared in the presence of a RAFT agent that allowed discrete nanoparticle formation with high branching densities although with relatively high polydispersity. The reactivity of ASDM-based RAFT groups attached to the nanogel structures was confirmed through synthesis of a simple block copolymer. By degradation of the crosslinks within a model nanogel, an average of at least 2.4 – 6.3 interconnected primary chains were found to constitute one nanogel particle. Two non-degradable nanogels with different branching densities were synthesized by the same technique and mixed with resin to form composite systems at 10 – 40 wt% loading. The reaction kinetics of nanogel-modified samples showed progressive decreases of rate of polymerization with an increase of nanogel loading likely due to the attached RAFT functionality in the nanogel structures. The volumetric shrinkage was reduced in a proportional manner with the nanogel concentration, while higher reductions in polymerization stress were found compared with the relative nanogel concentration. Mechanical properties of the final nanogel-modified glassy materials maintained or improved compared with the control for various nanogel concentrations.

Highlights.

  • A RAFT-based initiating system successfully provided nanogels with latent reactivity

  • The nanogel additives progressively reduced photopolymerization shrinkage and stress

  • RAFT residual functionality on nanogels affected conversion but not polymer strength

Acknowledgments

This work was supported by NIH/NIDCR RC1-DE020480 and R01-DE022348 as well as Septodont and Dentsply. The generous donation of monomers by Esstech is greatly appreciated. We thank Prof. Mark Stoykovich and Kate Morrissey for the assistance with GPC analysis.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  • 1.Astruc D, Ornelas C, Ruiz J. Metallocenyl dendrimers and their applications in molecular electronics, sensing, and catalysis. Accounts Chem Res. 2008;41:841–56. doi: 10.1021/ar8000074. [DOI] [PubMed] [Google Scholar]
  • 2.Boas U, Heegaard PMH. Dendrimers in drug research. Chem Soc Rev. 2004;33:43–63. doi: 10.1039/b309043b. [DOI] [PubMed] [Google Scholar]
  • 3.Patri AK, Kukowska-Latallo JF, Baker JR. Targeted drug delivery with dendrimers: Comparison of the release kinetics of covalently conjugated drug and non-covalent drug inclusion complex. Adv Drug Deliver Rev. 2005;57:2203–14. doi: 10.1016/j.addr.2005.09.014. [DOI] [PubMed] [Google Scholar]
  • 4.Walter MV, Malkoch M. Simplifying the synthesis of dendrimers: Accelerated approaches. Chem Soc Rev. 2012;41:4593–609. doi: 10.1039/c2cs35062a. [DOI] [PubMed] [Google Scholar]
  • 5.De Gennes PG, Hervet H. Statistics of starburst polymers. J Phys Lett-Paris. 1983;44:L351–L60. [Google Scholar]
  • 6.Voit B. Hyperbranched polymers - all problems solved after 15 years of research? J Polym Sci Pol Chem. 2005;43:2679–99. [Google Scholar]
  • 7.O’brien N, McKee A, Sherrington DC, Slark AT, Titterton A. Facile, versatile and cost effective route to branched vinyl polymers. Polymer. 2000;41:6027–31. [Google Scholar]
  • 8.Slark AT, Sherrington DC, Titterton A, Martin IK. Branched methacrylate copolymers from multifunctional comonomers: The effect of multifunctional monomer functionality on polymer architecture and properties. J Mater Chem. 2003;13:2711–20. [Google Scholar]
  • 9.Isaure F, Cormack PAG, Graham S, Sherrington DC, Armes SP, Butun V. Synthesis of branched poly(methyl methacrylate)s via controlled/living polymerisations exploiting ethylene glycol dimethacrylate as branching agent. Chem Commun. 2004:1138–9. doi: 10.1039/b401709a. [DOI] [PubMed] [Google Scholar]
  • 10.Li YT, Armes SP. Synthesis and chemical degradation of branched vinyl polymers prepared via atrp: Use of a cleavable disulfide-based branching agent. Macromolecules. 2005;38:8155–62. [Google Scholar]
  • 11.Vo CD, Rosselgong J, Armes SP, Billingham NC. Raft synthesis of branched acrylic copolymers. Macromolecules. 2007;40:7119–25. [Google Scholar]
  • 12.Moraes RR, Garcia JW, Barros MD, Lewis SH, Pfeifer CS, Liu JC, Stansbury JW. Control of polymerization shrinkage and stress in nanogel-modified monomer and composite materials. Dent Mater. 2011;27:509–19. doi: 10.1016/j.dental.2011.01.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Liu JC, Howard GD, Lewis SH, Barros MD, Stansbury JW. A study of shrinkage stress reduction and mechanical properties of nanogel-modified resin systems. Eur Polym J. 2012;48:1819–28. doi: 10.1016/j.eurpolymj.2012.08.009. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Labella R, Lambrechts P, Van Meerbeek B, Vanherle G. Polymerization shrinkage and elasticity of flowable composites and filled adhesives. Dent Mater. 1999;15:128–37. doi: 10.1016/s0109-5641(99)00022-6. [DOI] [PubMed] [Google Scholar]
  • 15.Davidson CL, Feilzer AJ. Polymerization shrinkage and polymerization shrinkage stress in polymer-based restoratives. J Dent. 1997;25:435–40. doi: 10.1016/s0300-5712(96)00063-2. [DOI] [PubMed] [Google Scholar]
  • 16.Weinmann W, Thalacker C, Guggenberger R. Siloranes in dental composites. Dent Mater. 2005;21:68–74. doi: 10.1016/j.dental.2004.10.007. [DOI] [PubMed] [Google Scholar]
  • 17.Carioscia JA, Lu H, Stanbury JW, Bowman CN. Thiol-ene oligomers as dental restorative materials. Dent Mater. 2005;21:1137–43. doi: 10.1016/j.dental.2005.04.002. [DOI] [PubMed] [Google Scholar]
  • 18.Ferracane JL. Developing a more complete understanding of stresses produced in dental composites during polymerization. Dent Mater. 2005;21:36–42. doi: 10.1016/j.dental.2004.10.004. [DOI] [PubMed] [Google Scholar]
  • 19.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. doi: 10.1016/j.dental.2014.02.015. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Liu JC, Rad IY, Sun F, Stansbury JW. Photo-reactive nanogels as a means to tune properties during polymer network formation. Polym Chem-Uk. 2014;5:227–33. doi: 10.1039/C3PY00870C. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Syrett JA, Haddleton DM, Whittaker MR, Davis TP, Boyer C. Functional, star polymeric molecular carriers, built from biodegradable microgel/nanogel cores. Chem Commun. 2011;47:1449–51. doi: 10.1039/c0cc04532b. [DOI] [PubMed] [Google Scholar]
  • 22.Lu H, Stansbury JW, Dickens SH, Eichmiller FC, Bowman CN. Probing the origins and control of shrinkage stress in dental resin-composites: I. Shrinkage stress characterization technique. J Mater Sci-Mater M. 2004;15:1097–103. doi: 10.1023/B:JMSM.0000046391.07274.e6. [DOI] [PubMed] [Google Scholar]
  • 23.Zetterlund PB, Yamazoe H, Yamada B. Addition-fragmentation chain transfer involving dimers of alpha-methylvinyl monomers studied by esr spectroscopy: Competition between fragmentation and bimolecular termination. Macromol Rapid Comm. 2003;24:197–201. [Google Scholar]
  • 24.Quinn JF, Barner L, Barner-Kowollik C, Rizzardo E, Davis TP. Reversible addition-fragmentation chain transfer polymerization initiated with ultraviolet radiation. Macromolecules. 2002;35:7620–7. [Google Scholar]
  • 25.Wang H, Li QB, Dai JW, Du FF, Zheng HT, Bai RK. Real-time and in situ investigation of “living”/controlled photopolymerization in the presence of a trithiocarbonate. Macromolecules. 2013;46:2576–82. [Google Scholar]
  • 26.Pryor WA, Pickering TL. Reactions of radicals. Rates of chain transfer of disulfides and peroxides with the polystyryl radical. J Am Chem Soc. 1962;84:2705–11. [Google Scholar]
  • 27.Yu Q, Zhu YS, Ding YH, Zhu SP. Reaction behavior and network development in raft radical polymerization of dimethacrylates. Macromol Chem Phys. 2008;209:551–6. [Google Scholar]
  • 28.Ye S, Cramer NB, Bowman CN. Relationship between glass transition temperature and polymerization temperature for cross-linked photopolymers. Macromolecules. 2011;44:490–4. [Google Scholar]

RESOURCES