Evaporation of low-volatility components in polymeric dental resins.

DENTAL-2579; No. of Pages 10

ARTICLE IN PRESS d e n t a l m a t e r i a l s x x x ( 2 0 1 5 ) xxx–xxx

Available online at www.sciencedirect.com

ScienceDirect journal homepage: www.intl.elsevierhealth.com/journals/dema

Evaporation of low-volatility components in polymeric dental resins Darren L. Forman a , Robert R. McLeod a , Parag K. Shah b , Jeffery W. Stansbury b,c,∗ a b c

Department of Electrical, Computer and Energy Engineering, University of Colorado at Boulder, Boulder, CO, USA Department of Chemical and Biological Engineering, University of Colorado Boulder, Boulder, CO, USA Department of Craniofacial Biology, School of Dental Medicine, Anschutz Medical Campus, Aurora, CO, USA

a r t i c l e

i n f o

a b s t r a c t

Article history:

Objectives. This study provides measurement of the volatility of selected photoinitiators

Received 2 June 2015

and monomers used in dental adhesive resins. A detailed determination of the spatial and

Accepted 8 June 2015

temporal character of camphorquinone (CQ) volatilization with respect to air flow condi-

Available online xxx

tions as well as media viscosity is assessed to gauge the effect of evaporative loss on the photopolymerization process and the photopolymers formed.

Keywords:

Methods. Vapor pressures of materials are measured by thermogravimetric analysis. A quan-

Evaporation

titative model assuming one-dimensional Fickian diffusion with surface evaporation is

Sublimation

presented and compared with measured photoinitiator volatilization from viscous and

Volatility

non-viscous resin samples, obtained by spectrophotometry and confocal microscopy. Model

Film

resins are prepared and subject to airthinning followed by photocuring, monitored in real-

Surface

time by Fourier transform infrared spectrometry.

Photoinitiator

Results. Vapor pressure measurements of the individual components of the adhesive resin

Camphorquinone

span nearly four orders of magnitude, with the photoinitiator CQ near the middle (0.6 Pa)

CQ

and the monomer HEMA at the upper end (10 Pa). We see depth-averaged CQ loss from non-

HEMA

viscous open films, while depthresolved measurements of viscous droplets show strong surface-localized CQ depletion. Good agreement is observed between measurements and the model. Finally, air-thinning of samples prepared with more-volatile photoinitiator and monomer is shown to cause longer induction times, slower early-stage polymerization rates and lower late-stage degree of conversion. Significance. Widely used compounds with vapor pressures as low as 0.6 Pa (0.001 Torr) undergo significant volatilization from samples ventilated under conditions generally representative to clinically used air-thinning procedures, with the potential to adversely affect the photopolymerization of both viscous and non-viscous resins. The inverse relationship between air-thinning and adhesive bond strength, observed elsewhere, may be partially caused by this same effect. © 2015 Academy of Dental Materials. Published by Elsevier Ltd. All rights reserved.

∗ Corresponding author at: Department of Craniofacial Biology, University of Colorado Denver, Mail Stop 8120, 13065, E. 17th Avenue Aurora, CO 80045, USA. Tel.: +1 303 724 1044; fax: +1 303 724 1945. E-mail address: [email protected] (J.W. Stansbury). http://dx.doi.org/10.1016/j.dental.2015.06.008 0109-5641/© 2015 Academy of Dental Materials. Published by Elsevier Ltd. All rights reserved.

Please cite this article in press as: Forman DL, et al. Evaporation of low-volatility components in polymeric dental resins. Dent Mater (2015), http://dx.doi.org/10.1016/j.dental.2015.06.008


2


1.

Introduction

Apart from polymerizable solvated adhesives and coatings for which solvents are meant to evaporate from solution after application to the target surface, the remaining, less-volatile components to coatings and adhesives are often assumed to evaporate so slowly that their volatility is essentially zero or at least negligible. This assumption is justifiable in many instances and applications. However, for situations involving thin films or highly ventilated environments or both, additional study is required to measure the potential for and effects of volatilization of the primary formulation components. Azeotropes aside, the formulation will change as some components volatilize more quickly than others and thus alter the formulation composition and properties. While pursuing other research [1,2], the authors determined that camphorquinone (CQ), a visible-light photoinitiator widely used in dental materials [3], was preferentially being depleted from thin-films during the course of experiments. Although measurable photoinitiator volatility is not unknown [4], no mention of CQ volatility could be found in the literature (apart from mention of purification by vacuum sublimation [5]), and the questions arose of whether CQ evaporation could significantly affect the performance of dental materials and how the CQ volatilization rate compares to that of other compounds. Active air thinning is used extensively in the application of dental adhesive systems. Solvents are included in dental adhesives to control initial viscosity, displace water and promote infiltration into demineralized dentin and to moderate the moisture-induced phase separation potential between the relatively hydrophilic and hydrophobic comonomers typically combined in these formulations [6]. In addition to solvent reduction, air thinning also aids placement of uniform adhesive layers. Although complete solvent elimination is not practically achievable under the constraints imposed in clinical dentistry, it is essential to minimize residual solvent in the adhesive at the time of polymerization to avoid reduced performance due to dilution of the monomers, which slows the rate of polymerization and decreases the resulting polymer network density [7,8]. Residual solvent during polymerization of the adhesive layer has the potential to affect both initial and long-term dentin bond strength and integrity. There are many conflicting reports regarding the effect of air thinning on the performance of the dental adhesives. Small amounts of air-drying are reported to increase bond strength while extended air-drying has been reported to actually decrease the bond strength, possibly due to thinning of the adhesive layer that also leads to increased oxygen inhibition and subsequent lowering of monomer conversion [9,10]. The intensity of application of the adhesive, the number of coats applied and also temperature and speed of the air can also result in significant differences in bonding properties. While these already constitute a large number of factors that affect the bond strength in dental adhesive applications, our results suggest that evaporation of the initiator from the formulation is another important factor that has been overlooked. This paper seeks to answer the questions related to volatility by quantifying the vapor pressures of selected compounds

other than solvents that are commonly found in dental resins, and by exploring via experiment and theory the characteristics and effects of volatilization from polymeric resins.

2.

Materials and methods

2.1.

Materials

Triethylene glycol dimethacrylate (TEGDMA), bisphenol A glycerolate dimethacrylate (BisGMA), 2-hydroxyethyl methacrylate (HEMA), camphorquinine (CQ), ethyl N,Ndimethylaminobenzoate (EDAB), camphor and dibutyl phthalate DBP were all obtained from Sigma-Aldrich. Urethane dimethacrylate (UDMA) was obtained from Esstech and 2,4,6-trimethylbenzoyldiphenylphosphine oxide (TPO) was obtained from BASF. All materials were used without further purification unless otherwise noted.

2.2.

Thermogravimetic analysis

A thermogravimetric analyzer (TGA; Perkin-Elmer Pyris 1) was used with 50 mg platinum sample pans, under a constant nitrogen purge flowrate of 20 mL min−1 . Evaporation rates were determined by holding each individual component sample at a series of constant temperatures for an interval of 5 to 10 min after temperature stabilization had occurred. Before data collection, moisture was removed by holding the samples under nitrogen at an elevated temperature for approximately 20 min, or until the evaporation rate became completely linear. Sample pans were washed thoroughly and sonicated before use, and were filled with only enough sample material to completely cover the floor of the pan (typically ∼10 mg). Vapor pressure is determined by a method that follows the work of Price [11] and others (see, for example [4,12–17]). Mass loss rate is given as

−dm = p˛ dt

M , 2RT

(1)

where p is pressure (Pa), ˛ is the vaporization coefficient, M is molar mass (Da), R is the ideal gas constant (m3 Pa K−1 mol−1 ) and T is temperature (K). For convenience, we write this as p = k,

(2)

√ where k = 2R/˛ and = dm/dt T/M. For a given setup, k may be experimentally determined so that mass loss rate, temperature and molecular weight can be converted to a vapor pressure p. As in [12], we use glycerol as a reference material using vapor pressure data from [18] to obtain the linear best-fit of log(p) vs. log() (R2 = 0.998). 2.3.

Low-speed forced convection

For low viscosity standing droplet tests, resin was prepared by adding 0.4 wt% camphorquinone (CQ) into triethylene glycol dimethacrylate (TEGDMA). When the CQ was fully dissolved, 0.6 g of the resin was spread evenly over a 76 × 50 mm microscope slide. The slides had been previously cleaned in Nano



ARTICLE IN PRESS 3

d e n t a l m a t e r i a l s x x x ( 2 0 1 5 ) xxx–xxx

Strip (Cyantek), rinsed in de-ionized (DI) water and blown dry with nitrogen. Samples were then stored on a level platform 10 cm inside a fume-hood sash, using shims and visual inspection to maintain a uniform-height puddle of resin across each slide. Air velocity near the platform was 0.5 m s−1 , measured with a hot wire anemometer (Extech). At the time specified, samples were removed and the standing liquid was drained into 1 mm-path plastic cuvettes. For high viscosity standing droplet tests, resin was prepared by adding 1.0 wt% CQ and 0.5 wt% EDAB to a blend of BisGMA/TEGDMA (4:1 by mass). The mixture was stirred at 60 ◦ C until solid components were dissolved and the resin appeared completely uniform and clear, with a yellow tint associated with CQ. A small droplet was then spread across a number 1.5 microscope coverslip and stored on the previously described fume-hood platform. After 2.5 h, the sample was removed and immediately sealed in a 1 cm3 chamber purged with ultra-high purity nitrogen at 0.25 L min−1 . After 30 s of pure nitrogen purge, the sample was illuminated with a blue-LED dental lamp at 22 mW cm−2 for 20 s. Light from the cure-lamp struck at the resin/coverslip interface first and then propagated through the droplet to the open top surface. The control sample was processed similarly, but was cured immediately after application to the coverslip and not stored in the fume-hood. For the vapor-phase replenishment test, a sample was prepared and stored in the fume-hood as above. After 2.5 h, the sample was removed and placed in the 1 cm3 purge chamber. The 0.25 L min−1 nitrogen purge stream had been modified to first pass in and out of a 50 mL Erlenmeyer flask containing 0.3 g CQ powder, with an outlet filter to prevent any particulates from reaching the resin. After 60 s of purge with nitrogen and CQ vapor, the sample was photo-cured as above.

2.4.

Analysis of coupled evaporation and diffusion

The following is a quantitative model for the coupled processes of evaporation at a surface and diffusive transport beneath, as presented in the text by Crank [19]. The boundary condition for evaporation across a surface can be represented as −D

∂C = a (C0 − CS ) ∂z

(3)

at z = 0, where C0 is the local concentration just inside the boundary, Cs is the equilibrium concentration with the ambient environment and a is the evaporation coefficient. Given an infinite sheet with thickness −l < z < l and applying the boundary condition at each surface, concentration C is solved below given initial uniform concentration C2 and ambient equilibrium concentration C0 :

∞ 2L cos (ˇn z/l) exp −ˇn2 Dt/l2 C − C2 2 =1− .

C0 − C2

n=1

ˇn + L2 + L cos (ˇn )

(4)

here l is sheet thickness, a is evaporation rate and D is diffusivity. ˇn are the positive roots of ˇ tan ˇ = L

(5)

and L is the dimensionless quantity L=

la . D

(6)

Note that while this solution is apparently given for evaporation from both sides of a sheet with thickness 2l, the solution is identical for a sheet of thickness l bounded by an impermeable substrate on one side; the zero-flux condition at z = 0 is met in both cases. Integrating Eq. (4) to obtain total mass contained in the sheet gives

2L2 exp −ˇn2 Dt/l2 Mt . =1− M∞ ˇn2 ˇn2 + L2 + L ∞

(7)

n=1

when L 1, mass change in the sheet is limited by diffusive transport to and from the surface; when L 1, surface adsorption/evaporation is the limiting factor. In order to evaluate the depth-resolved concentration model, we make some assumptions about material properties. CQ diffusivity of 0.25 ␮m2 s−1 has been measured in a resin with viscosity of 1 Pa s [2]. For the sake of simplicity, we assume here the validity of Stokes–Einstein reciprocity in viscosity/diffusivity to get an estimate of diffusivity. We calculate CQ diffusivity D = 5 ␮m2 s−1 in = 0.05 Pa s TEGDMA [20] and D = 0.03 ␮m2 s−1 in = 9 Pa s BisGMA/TEGDMA (4:1 by mass) [21].

2.5.

High-speed forced convection

Resins were formulated from BisGMA/TEGDMA (1:1 by mass) and BisGMA/HEMA (1:1 by mass), with either CQ/EDAB or TPO as photoinitiators. An automatic film applicator (ZAA 2300, Zehntner Instruments) was used to spread resin samples over a glass plate into uniform films approximately 5 cm wide and 25 cm long. The average film thickness delivered by the applicator was approximately 12 ␮m, determined by spreading a known mass of glycerol over the plate, measuring the coverage area and then using the known density of glycerol ( = 1.26 g mL−1 ) to calculate the film thickness. The same applicator blade-height was used throughout. After deposition, air was blown over the length of the film for a predetermined time interval. Air was supplied from a nozzle (4 cm wide and 1 cm tall) connected by hose to a shop-vac (Ridgid, 5.0 hp) running in reverse. The air velocity over the film was approximately 32 m s−1 as measured by a hot wire anemometer (Extech). This rate is reasonably comparable to the output of dental air-syringe nozzles (∼20 m s−1 ) [22]. After blowing air, the liquid was scraped together from across the glass plate and deposited onto a salt plate for FT-IR characterization.

2.6.

Optical characterization

Visible-light absorption of CQ/TEGDMA solutions was measured against neat TEGDMA in a UV–vis spectrophotometer (Thermo-Nicolet Evolution 300), with the samples in plastic cuvettes. Polymer conversion was measured by monitoring the C C absorption in the 1680–1620 cm−1 region with an FT-IR spectrometer (Thermo-Nicolet Nexus 6700), the resin sample sandwiched between salt plates and illuminated by



ARTICLE IN PRESS

4

d e n t a l m a t e r i a l s x x x ( 2 0 1 5 ) xxx–xxx

Table 1 – Vapor pressures (VP) and Antoine coefficients (A, B and C) for the equation log p = A − B/ (C + T). Following convention, coefficients are computed for p in Torr and T in ◦ C. The vapor pressures shown are obtained by first evaluating the equation in Torr then converting to Pa. Literature values are given where available. Compound

Function

A

B

C

Camphor DBP

Reference Reference

9.133 7.652

3061 2477

291.1 175.3

HEMA

Monomer

6.618

1462

163.4

TEGDMA CQ EDAB TPO PPD

Monomer Photoinitiator Coinitiator Photoinitiator Photoinitiator

7251 1109 4085 1900 254.8

429.5 121.8 306 184.3 24.66

12.64 5.214 9.564 4.226 3.234

320–500 nm filtered light output from a mercury-arc curing lamp (Acticure, Exfo). Z-stack images of CQ fluorescence were acquired on a Nikon A1R confocal microscope using a 1.45 NA, 100 Nikon Plan Apo ␭-series oil-immersion objective. A = 488 nm laser was used for excitation, and a 525 nm-pass filter to collect emission. Z-stack scans were begun inside the coverslip glass and proceeded upwards throughout the cured resin droplet at a resolution of 0.25 ␮m in x, y and 2.0 ␮m in z. All excitation, collection and scan parameters were fixed across all samples tested.

3.

Results

3.1.

Vapor pressures

Table 1 contains the vapor pressures and Antoine coefficients we obtain for selected monomers and photoinitiators. The room-temperature evaporation rate of the monomer HEMA is the highest, excluding the reference material camphor from consideration as it is not used in resins. Next most volatile is the photoinitiator 1-phenyl-1,2-propanedione (PPD), which is unsurprising given the strong odor of that material. CQ has approximately 1/15th the room-temperature vapor pressure of HEMA, and the coinitiator EDAB has approximately 1/3rd the vapor pressure of CQ. For comparison, unpublished results related to Forman et al. [2] suggest that EDAB volatilizes from thin films ∼10 more slowly than CQ. TEGDMA has a vapor pressure 1/10th that of CQ, making it one of the lower-volatility substances included in the study. The least volatile measureable substance shown here is the photoinitiator TPO, which was comparable to DBP.

3.2.

Initiator volatilization versus viscosity

To measure CQ volatility from resins and explore the effect of viscosity on photoinitiator volatilization, we characterize volatilization from low and high viscosity resins using the optical properties of CQ. To study volatilization from low viscosity media, we take advantage of the characteristic visible-band absorption profile of CQ. Samples are first prepared by completely dissolving CQ into the monomer TEGDMA and then spreading it evenly over glass plates and storing the samples in the opening of a fume-hood. After 12 h of storage in fume-hood airflow, the solution is a visibly paler

T (◦ C) 70 25 147 20 25 25 25 25 25 25

Meas. VP (Pa)

Lit. VP (Pa)

Ref.

600 0.0026 120 5.9 9.6 0.065 0.61 0.22 0.0019 1.7

530 0.0027 130 8 1.3

[28] [13] [29] [30] [31]

shade of yellow than an initially identical solution stored in a closed vial. Photo-bleaching is not a concern here, as the samples were prepared and stored under yellow LED task lights in a room lit with amber-passband fluorescent lights that decrease ambient irradiance at the mercury h-line (404.7 nm) from 320 nW cm−2 to 0.01 nW cm−2 . Visible-band absorption measurements indicate that CQ absorbance is decreased by half from the test sample during this interval, as shown in Fig. 1. A longer storage interval results in a further absorbance decrease, as shown for the sample collected after 72 h. To study volatilization from high viscosity media, we use the depth-resolved imaging capability of confocal microscopy with the blue-excited green fluorescence of CQ. As before, test samples were stored in a ventilated fume-hood for a controlled amount of time prior to characterization. To prevent the spatial relaxation of CQ concentration gradients in the time interval between airflow exposure and confocal microscope characterization, test samples were photo-cured in an oxygen-free environment immediately after removal from ventilation. Confocal z-stack images of these samples are shown in Figs. 2 and 3. Compared to the control (a), the first test case (b) has a much lower fluorescent intensity near the resin/air interface, so much so that the top surface is difficult to resolve. The second test case (c) was processed identically to the first apart from exposure to CQ vapor just before and during photocuring. In scans of that sample, fluorescent intensity diminishes from the bottom surface to about 3/4 the droplet height and then increases again before abruptly stopping at the top surface. Given the time-series plots of the resin absorption spectrum in Fig. 1, we assume the Beer-Lambert law and take peak absorbance to be proportional to CQ concentration. Next, we use the computed diffusion coefficient of CQ in TEGDMA to evaluate Eq. (7) and obtain an environmentally-dependent fitted value for the evaporation rate a = 3.8 nm s−1 . As shown in Fig. 1(b), nearly uniform depletion of CQ through time is therefore predicted for the low viscosity system, consistent with mass-loss limited by the surface evaporation rate (L = 0.15). For the viscous system, we instead predict a large gradient of CQ depletion towards the air/resin interface (L = 11). Since the evaporation rate coefficient a should remain the same for both low and high viscosity systems (the evaporating species and the ventilated environment are the same), we may evaluate Eq. (4) to predict the depth-resolved CQ concentration profile, which agrees well with the experiment as shown in Fig. 4.




5

Fig. 1 – Photoinitiator volatilization from a resin with low relative viscosity. (a) Visible absorption spectrum of TEGDMA resin samples, initially containing 0.4 wt% CQ. Average resin thickness during storage is 200 ␮m. Notation indicates length of time stored in airflow. (b) Predicted CQ concentration as a function of depth, obtained from evaluating Eq. (4) with D = 5 ␮m2 s−1 and a = 3.8 nm s−1 .

Fig. 2 – Photoinitiator volatilization from a resin with high viscosity. Shown is a volume rendering of raw CQ fluorescence, collected by confocal microscope from an unfilled BisGMA/TEGDMA (4:1) resin. A control sample (a) was photocured immediately after preparation; a test sample (b) was stored under airflow prior to curing, and shows significant CQ depletion near the top surface (white arrow); another test sample (c) was prepared and stored under ventilation identical to (b), but was subsequently exposed to CQ vapor causing enrichment of the top surface layer (white arrow). Maps are rendered to scale in x, y and z, with the x and y edges of length 127 ␮m.

3.3.

Cure profile versus air-thinning duration

After the samples were prepared and treated with airthinning, they were collected from the surface and immediately sandwiched between polished salt-plates for FT-IR characterization. The results are shown in Figs. 5–7. The resin formulated with TEGDMA and TPO did not exhibit sensitivity to ventilation prior to curing as shown in Fig. 5. When TPO is replaced by the more volatile CQ, however, a ventilationtime dependence is apparent. Fig. 6 shows that CQ volatility results in increased induction time, slower early-stage polymerization and lower late-stage conversion. When TEGDMA is additionally replaced by HEMA, the ventilation sensitivity is greater and also of a different character. As shown in Fig. 7, extended ventilation the HEMA/BisGMA sample results in lower late-stage conversion. In the case of the 2 min airthinning treatment, early stage polymerization is faster than both the 5 min treatment and no treatment.

4.

Discussion

4.1.

Differences in vapor pressure

Our vapor pressure measurements span more than 6 orders of magnitude when reference materials are included, a range that is large but consistent with similar studies [11,12]. Over 1000 variation is observed amongst the compounds that are used in dental resins. This factor is large enough to allow some formulation components to evaporate completely while others are largely unaffected. In the case of the CQ/TEGDMA resin trial, this is exactly what we observe. While CQ was the initial focus of this study due to its unexpected room-temperature volatilization, we see that other resin components like HEMA have even higher volatilities. As known reference materials, dibutyl phthalate (DBP) was chosen for its very low vapor pressure, and camphor



6


Fig. 3 – Cross-sections of data from Fig. 2. In the raw fluorescence shown at top, even the control sample shows signal diminishing with distance from the glass/resin interface (left dotted line, located near z = 20 ␮m); however, this gradient correlates with +ˆz confocal scan direction, and is attributed to the excitation-induced bleaching of resin beyond the focus. To correct for this, a cubic polynomial f(z) is fitted to raw data from the control sample (a) and used to produce the plots (a ), (b ) and (c ) at bottom. Right dotted line indicates resin/air interface.

Fig. 4 – Predicted and measured CQ concentration as a function of depth in a viscous BisGMA/TEGDMA (4:1) resin. The predictions are obtained from evaluating Eq. (4) with D = 0.03 ␮m2 s−1 and a = 3.8 nm s−1 for various amounts of time. The measurement is from Fig. 4(b ).

was chosen for its physical (powder at room-temperature, melting point: 175 ◦ C) and structural similarity to CQ. Two literature values for the monomer 2-hydroxyethyl methacrylate (HEMA) are also given, with the poor agreement and apparently anomalous temperature dependence probably due to a combination of factors such as differences in experimental method and whether higher-volatility impurities were removed prior to measurement. Attempts to measure the volatility of the monomers bisphenol-A glycidyl methacrylate (BisGMA) and urethane dimethacrylate (UDMA) were unsuccessful. For these relatively high molecular weight monomers,

evaporation rates even in the 60–100 ◦ C range were low enough to cause poor signal-to-noise in the measurements, and higher temperatures caused polymerization under the oxygen-free nitrogen purge. However, it can safely be said that UDMA and BisGMA display negligible volatility under clinically relevant air-thinning conditions. We believe the best use of Table 1, or other data collected in a similar fashion, is for the relative ranking of compound volatility. We have not attempted to account for some potential sources of error such as meniscus (in the case of liquids), the difficult-to-measure surface area of powders (as in the case of camphor, CQ and EDAB) or diffusion-limited transport in the gas phase. Nevertheless, the vapor pressures we obtain for reference materials compare reasonably well with those available in the literature.

4.2. Viscosity and the character of initiator volatilization Species evaporation from a film has a spatial character that is a strong function of media viscosity, film thickness and evaporation rate as described by the term L in Eq. (6). For a given evaporation rate a, the volatile component concentration stays uniform through film depth when the product of viscosity and thickness is low enough to satisfy L 1 conditions. That is, when initiator selectively evaporates from the surface of a low viscosity resin applied as a thin film, diffusion should even-out the remaining initiator so that the region near the surface contains the same concentration as elsewhere in the film at any given moment. The total mass of initiator in the film will decay exponentially to zero, consistent with our measurements of CQ evaporation from TEGDMA under slow-moving air.




7

Fig. 5 – Photopolymerization of relatively non-volatile model adhesive resin. Degree of conversion measured in the mid-IR for samples prepared with and without ventilation prior to curing (10 mW cm−2 illumination begins at 0.5 min). Average of three runs is shown for each, 0 = 5 = 1.1 (% conversion). Inset shows zoom-in of early-stage polymerization.

Fig. 6 – Photopolymerization of model adhesive resin with moderately volatile photoinitiator (CQ). Degree of conversion measured in the mid-IR for samples prepared with various amounts of ventilation prior to curing (10 mW cm−2 illumination begins at 0.5 min). Average of three runs is shown for each, 0 = 1.1, 2 = 1.3, 5 = 3.0 (% conversion). Inset shows zoom-in of early-stage polymerization.

Conversely, when the product of viscosity and thickness is high enough to give L 1, steep concentration gradients arise. For example, when initiator selectively evaporates from the surface of a thickly-applied viscous resin, the region near the surface will eventually lose all initiator even as the remainder of the volume retains the original initiator concentration. Over

time, the sharply-defined depleted region will penetrate farther and farther into film until no initiator remains. The initial stage of this process is confirmed by our z-sectioned images of CQ fluorescence from a viscous BisGMA/TEGDMA droplet. There is one unavoidable consequence of our z-sectioned CQ conversion measurement that needs to be discussed:



8


Fig. 7 – Photopolymerization of model adhesive resin with moderately volatile photoinitiator (CQ) and monomer (HEMA). Degree of conversion measured in the mid-IR for samples prepared with various amounts of ventilation prior to curing (10 mW cm−2 illumination begins at 0.5 min). Average of three runs is shown for each, 0 = 3.5, 2 = 1.3 = 1.9, 5 = 3.0 = 3.7 (% conversion). Inset shows zoom-in of early-stage polymerization.

photo-bleaching. Since absorption of the excitation beam is small, light propagates relatively unattenuated up to and beyond the focal spot. As a result, the sectioned volume receives approximately the same optical exposure throughout. However, since the accumulated exposure at the end of the scan is much larger than at the beginning, there is the reasonable possibility that the last region to be measured will have already undergone substantial photo-bleaching and will thus appear much dimmer than if it had been measured first. Examination of the control (unventilated) raw fluorescence cross-section indicates that this reaction-based CQ depletion has occurred. Since the accumulated optical exposure as a function of depth is identical for all measurements (scan parameters were fixed), a function f(z) is empirically fitted to the control and used to normalize or calibrate data from the test samples. These normalized results are in good agreement with the predicted depth profile output of the model. There is another, earlier exposure worth mentioning. Immediately after preparation but before measurement, samples are ‘frozen’ with a 20 s blue-light cure at 22 mJ cm−2 . This exposure is sufficient to render the droplets glassy to the touch, but actually bleaches only ∼1% of available CQ [23]. To further verify that these samples are optically thin and that the ‘freezing’ exposure does not generate significant gradients of bleached material within the sample, we calculate transmission through the resin with the Beer–Lambert law: T = exp (−εlc)

(8)

where ε is the molar absorption coefficient, l is optical path length and c is concentration. In this, we use ε = 46 M cm−1

[23], l = 0.015 cm and c = 0.061 M, the transmission coefficient T = 0.96. To summarize, evaporative loss of low volatility components such as initiator may be observed in high and low viscosity resins. Evaporation from non-viscous resin films will tend to affect the entire film uniformly. A large increase in film viscosity will tend to lock-in volatile species concentrations in the interior of the film at the expense of the surface regions, which become significantly depleted. Whether global or local, effects of initiator loss would include increased induction time, decreased ability to overcome in-diffusing atmospheric oxygen, decreased final degree of conversion, decreased crosslinking, lower modulus and decreased mechanical strength. It is interesting to note, however, the possibility of mitigating or reversing these effects by vapor-phase component delivery.

4.3. Effect of initiator and monomer volatilization on photopolymerization kinetics As expected, the resin formulated with relatively non-volatile TEGDMA and TPO did not exhibit sensitivity to ventilation prior to curing. When TPO is replaced by the more volatile CQ, however, a ventilation-time dependence is apparent. The photoinitiator evaporation results in increased induction time, slower early-stage polymerization and lower late-stage conversion. If TEGDMA is additionally replaced by HEMA (commonly used for its compatibility with water), the ventilation sensitivity is greater and also of a different character. This is caused by dramatically increased viscosity and decreased mobility as HEMA evaporates and leaves behind the




essentially non-volatile and highly viscous BisGMA. The change in viscosity after ventilation is large enough to be immediately apparent as samples are handled during collection for FT-IR characterization. Interestingly, the 2 min air-purge causes faster initial polymerization than both the 5 min and non-ventilated resin samples. This can be ascribed to that resin having a viscosity sufficiently high to start polymerization under the Trommsdorf–Norrish effect, but not so high that early-stage polymerization is under diffusion control [24]. While this study does not establish whether component volatility is a significant factor in the quality of clinical restorations, it does raise the question. Many studies have shown that the strength of adhesive bonds is dependent upon the details of the application technique, especially in regard to the air-thinning step. Several of them, [9,10] find that in some cases, prolonged air-thinning (up to 40 s) of the adhesive resin causes weaker bonding. In Hiraishi et al. [25], higher-velocity air causes stronger bonds when the adhesive is HEMA-free. Warmer air temperatures are found to increase bond strength in [26], and agitation is shown to be beneficial in [27]. The effects shown in Figs. 6 and 7 were caused by longer airthinning durations than are typically used in clinical settings. However, the conditions used there are intended to roughly approximate clinical conditions. Halving the film thickness, for example, would increase the relative evaporative loss rate by at least a factor of two. As noted by the various authors of these studies, bond-strength variations are likely dependent on a large number of factors, including phase separation, oxygen inhibition (influenced by adhesive layer thickness), air voids, inhomogeneities in adhesive infiltration into the dentin and viscosity-dependent photopolymerization kinetics (influenced by temperature). Component volatility is an additional factor that may have contributed to these results and could affect the clinical efficacy of existing commercial resin-based formulations.

[5]

[6]

[7]

[8]

[9]

[10]

[11]

[12]

[13]

[14]

[15]

[16]

Acknowledgements [17]

The authors acknowledge funding from the following sources: NSF CAREER Award ECCS 0954202, a DoD/NDSEG Fellowship to DF, a gift from Oracle Corp., NSF IGERT grant 0801680 and the NSF Industry/University Cooperative Research Center for Fundamentals and Applications of Photopolymerization.

[18]

[19]

references

[1] Scott TF, Kowalski BA, Sullivan AC, Bowman CN, McLeod RR. Two-color single-photon photoinitiation and photoinhibition for subdiffraction photolithography. Science 2009;324:913–7. [2] Forman DL, Cole MC, McLeod RR. Radical diffusion limits to photoinhibited superresolution lithography. Phys Chem Chem Phys 2013;15:14862–7. [3] Jakubiak J, Allonas X, Fouassier JP, Sionkowska A, Andrzejewska E, Linden LA, et al. Camphorquinone–amines photoinitating systems for the initiation of free radical polymerization. Polymer 2003;44:5219–26. [4] Ciba Specialty Chemicals. Photoinitiators for UV curing/key products selection guide; 2003. Accessed at

[20]

[21]

[22] [23]

[24]

9

http://people.rit.edu/deeemc/reference 13/Imprint/ Photoinitiators%20for%20UV%20curing.pdf in February 2015 . Luk CK, Richardson F. Circularly polarized luminescence spectrum of camphorquinone. J Am Chem Soc 1974;96:2006–9. Pashley DH, Tay FR, Breschi L, Tjäderhane L, Carvalho RM, Carrilho M, et al. State of the art etch-and-rinse adhesives. Dent Mater 2011;27:1–16. Cadenaro M, Breschi L, Rueggeberg FA, Suchko M, Grodin E, Agee K, et al. Effects of residual ethanol on the rate and degree of conversion of five experimental resins. Dent Mater 2009;25:621–8. Ye Q, Spencer P, Wang Y, Misra A. Relationship of solvent to the photopolymerization process, properties, and structure in model dentin adhesives. J Biomed Mater Res, A 2007;80:342–50. Spreafico D, Semeraro S, Mezzanzanica D, Re D, Gagliani M, Tanaka T, et al. The effect of the air-blowing step on the technique sensitivity of four different adhesive systems. J Dent 2006;34:237–44. Daneshmehr L, Ozer F, Mante FK, Blatz MB. Relationship between air-blowing duration and bond strengths of three adhesive systems to dentin after thermal aging. Dent Mater J 2013;32:767–74. Price DM, Hawkins M. Vapour pressures of hydroxybenzophenone UV absorbers. Thermochim Acta 1999;329:73–6. Aschenbrenner O, Supasitmongkol S, Taylor M, Styring P. Measurement of vapour pressures of ionic liquids and other low vapour pressure solvents. Green Chem 2009;11:1217–21. Donovan SF. New method for estimating vapor pressure by the use of gas chromatography. J Chromatogr, A 1996;749:123–9. Gückel W, Synnatschke G, Rittig R. A method for determining the volatility of active ingredients used in plant protection. Pestic Sci 1973;4:137–47. Gückel W, Kästel R, Kröhl T, Parg A. Methods for determining the vapour pressure of active ingredients used in crop protection Part IV. An improved thermogravimetric determination based on evaporation rate. Pestic Sci 1995;45:27–31. Price DM, Hawkins M. Calorimetry of two disperse dyes using thermogravimetry. Thermochim Acta 1998;315:19–24. Price DM. Vapor pressure determination by thermogravimetry. Thermochim Acta 2001;367:253–62. Vapor pressure of glycerol, Korean Chemical Engineering and Materials Research Information Center, Thermophysical Properties Databank. URL: https://www.cheric.org/ research/kdb/hcprop/showprop.php?cmpid=912 (accessed November 2014). Crank J. The mathematics of diffusion. 2nd ed. Oxford: Clarendon Press; 1975. Goncalves F, Kawano Y, Pfeifer C, Stansbury JW, Braga RR. Influence of BisGMA, TEGDMA, and BisEMA contents on viscosity, conversion, and flexural strength of experimental resins and composites. Eur J Oral Sci 2009;117:442–6. Beun S, Bailly C, Dabin A, Vreven J, Devaux J, Leloup G. Rheological properties of experimental Bis-GMA/TEGDMA flowable resin composites with various macrofiller/microfiller ratio. Dent Mater 2009;25:198–205. Thorp C and Segal AJ. “Dental nozzle.” U.S. Patent Application 13/703,113; 2011. Chen YC, Ferracane JL, Prahl SA. Quantum yield of conversion of the photoinitiator camphorquinone. Dent Mater 2007;23:655–64. Dickens SH, Stansbury J, Choi K, Floyd C. Photopolymerization kinetics of methacrylate dental resins. Macromolecules 2003;36:6043–53.



10


[25] Hiraishi N, Breschi L, Prati C, Ferrari M, Tagami J, King N. Technique sensitivity associated with air-drying of HEMA-free, single-bottle, one-step self-etch adhesives. Dent Mater 2007;23:498–505. [26] Garcia FC, Almeida JC, Osorio R, Carvalho RM, Toledano M. Influence of drying time and temperature on bond strength of contemporary adhesives to dentine. J Dent 2009;37:315–20. [27] Amaral R, Stanislawczuk R, Zander-Grande C, Gagler D, Reis A, Loguercio A. Bond strength and quality of the hybrid layer of one-step self-etch adhesives applied with agitation on dentin. Oper Dent 2010;35:211–9.

[28] Sigma-Aldrich. Camphor MSDS; 2014. Accessed on http://www.sigmaaldrich.com December 2014 . [29] Sigma-Aldrich. Dibutyl phthalate MSDS; 2014. Accessed on http://www.sigmaaldrich.com December 2014 . [30] Evonik. 2-Hydroxyethyl methacrylate MSDS; 2014. Accessed on http://corporate.evonik.com December 2014 . [31] Sigma-Aldrich. 2-Hydroxyethyl methacrylate MSDS; 2014. Accessed on http://www.sigmaaldrich.com December 2014 .


Formation of functionalized nanoclusters by solvent evaporation and their effect on the physicochemical properties of dental composite resins.

Novel acrylic resins for dental applications.

Evaluation of methylene lactone monomers in dental resins.

Factors influencing mercury evaporation rate from dental amalgam fillings.

Synthesis of a large-sized mesoporous phosphosilicate thin film through evaporation-induced polymeric micelle assembly.

solvent evaporation method.

Characterization of inorganic fillers in visible-light-cured dental composite resins.

Preparation of Dental Resins Resistant to Enzymatic and Hydrolytic Degradation in Oral Environments.

Structure-property relationships in hybrid dental nanocomposite resins containing monofunctional and multifunctional polyhedral oligomeric silsesquioxanes.

Bisphenol A Release: Survey of the Composition of Dental Composite Resins.

Effects of prepolymerized particle size and polymerization kinetics on volumetric shrinkage of dental modeling resins.

NIR-spectroscopic investigation of water sorption characteristics of dental resins and composites.

Reduction of bacterial adhesion on dental composite resins by silicon-oxygen thin film coatings.

Permeation of water through the polymeric components of boil-in-bag laminates.

The crystalline components of dental calculi: human vs. dog.

Long-Term Fluoride Release from Dental Resins Affects STRO-1+ Cell Behavior.

Syndromes with unusual dental findings or gingival components.

Effect of the irradiance distribution from light curing units on the local micro-hardness of the surface of dental resins.

Diffusion of residual monomer in polymer resins.

Immuno-electron microscopic study of antigenic surface components of Actinomyces naeslundii in human dental plaque.

Evaporation from open microchannel grooves.

Evaluation of light curing distance and mylar strips color on surface hardness of two different dental composite resins.

Polymeric vs hydroxyapatite-based scaffolds on dental pulp stem cell proliferation and differentiation.

Preparation and properties of calcium-silicate filled resins for dental restoration. Part I: chemical-physical characterization and apatite-forming ability.