Highlights
- •
The spatial irradiance and spectral emission from a LCU may not be homogeneous.
- •
LCUs with an inhomogeneous output can affect localized hardness values of RBCs.
- •
Researchers and manufacturers should be aware of the effects of beam inhomogeneity.
Abstract
Objective
To demonstrate the effect of localized irradiance and spectral distribution inhomogeneities of one LED-based dental light-curing unit (LCU) on the corresponding microhardness values at the top, and bottom surfaces of four dental resin-based composites (RBCs), which contained either camphorquinone (CQ) alone or a combination of CQ and monoacylphosphine oxide (TPO) as photoinitiators.
Methods
Localized irradiance beam profiles from a polywave LED-based LCU were recorded five times using a laser beam analyzer, without and with either a 400 nm or 460 nm narrow bandpass filter placed in front of the camera lens. Five specimens of each of the four RBCs (two containing CQ/TPO and two containing CQ-only) were exposed for 5-, 10-, or 30-s with the light guide directly on the top surface of the RBC. After 24 h, Knoop microhardness values were measured at 45 locations across the top and bottom surfaces of each specimen. Microhardness readings for each RBC surface and exposure time were correlated with localized patterns of the LCU beam profile, measured using the 400 nm and 460 nm bandpass filters. Spearman rank correlation was used to avoid relying on an assumption of a bivariate normal distribution for the KHN and irradiance.
Results
The local irradiance and spectral emission values were not uniformly distributed across the light tip. There was a strong significant positive correlation with the irradiance beam profile values from the LCU taken through bandpass filters and the microhardness maps of the RBC surfaces exposed for 5 and 10 s. The strength of this correlation decreased with increasing exposure time for the RBCs containing CQ only, and increased for the RBCs containing both CQ and TPO.
Conclusions
Localized beam and spectral distributions across the tip end of the light guide strongly correlated with corresponding areas of microhardness in both the top and bottom surfaces among four RBCs with different photoinitiator contents.
Significance
A light-curing unit with a highly inhomogeneous light output can adversely affect localized microhardness of resin-based composites and this may be a contributing factor for premature failure of a restoration.
1
Introduction
More than two hundred sixty million direct resin-based composite (RBC) restorations are placed annually worldwide . Premature failure of these restorations will have significant health and financial implications. Several recent reports indicate that the median longevity of posterior RBCs placed in dental offices worldwide is close to a mere 6 years . A recent Cochrane review reported that posterior RBCs are almost twice as likely to fail as amalgam restorations . The two most common causes of failure are secondary caries and bulk fracture. Both these outcomes may be the result of inadequate photo-polymerization of the RBC, which is known to adversely affect both the RBC properties and the bond strength between the RBC and preparation walls . Thus it is important to examine the likely causes of inadequate photo-polymerization of RBC restorations.
To ensure optimal photo-polymerization of the RBC, the radiant exposure and spectral range requirements of the RBC must be fulfilled by the radiant output from the light-curing unit (LCU), while avoiding damage to the oral tissues caused by excessive temperature increases . The most commonly used photosensitizer in RBCs is camphorquinone (CQ). However, CQ is bright yellow and it only moderately photobleaches upon exposure to a LCU when using clinically relevant times . Alternative photoinitiators that are not as chromogenic as CQ are used by some RBC manufacturers . These alternative photoinitiators, such as monoacylphosphine oxide (TPO) and derivatives of dibenzoyl germanium , have peak absorbance values below 420 nm . Consequently, these photoinitiators will not be efficiently activated by monowave LED-based LCUs that deliver light mostly in the 445 nm to 480 nm spectral range . Although these alternative photoinitiators are more reactive than CQ, fewer of these photons will reach the bottom of the RBC due to the effects of filler particle size and increased Rayleigh scattering of the lower wavelengths of light .
Manufacturers do not commonly list all the photoinitiators or the exact filler particle sizes used in their products making it difficult to predict the performance of narrow band, single-peak blue LED-based LCUs on a specific brand of RBC. To overcome this limitation, third-generation, polywave blue–violet LED-based LCUs have been introduced that claim to polymerize all resin-based restorations. These blue–violet LCUs use a combination of up to three different “colors” of LED chips, with spectral emissions peaking near 440–460 nm (blue) and near 400–410 nm (violet) .
Differences in the light outputs among LCUs are often not readily detectable by visual inspection, nor by a “dental radiometer”. The ISO 6050 standard for calculating irradiance from a LCU assumes that the irradiance and spectral emission profile of the LCU light beam are homogeneous and can be fully characterized by a single irradiance value . Similarly, the ISO 11405 bond strength test and the ISO 4049 depth of cure tests assume that light output from the LCU is uniformly distributed and that the specimen will receive the same irradiance and spectrum of light across its entire surface. It is now well established that the irradiance distribution from many dental LCUs can be very inhomogeneous and this inhomogeneity can cause non-uniform polymerization of the RBC . Additionally, there is a problem with the method used to calculate the irradiance – the physical diameter of the light guide is used instead of the functional diameter of the light beam within the light guide which will produce an erroneously low averaged irradiance from the LCU. These shortcomings in the standards lead to incorrect and misleading irradiance and depth of cure values being stated by both manufacturers and researchers alike.
As early as 1983, researchers were reporting discrepancies in the uniformity of the beam profiles of UV curing lights . Laser beam analyzers are now commonly used to measure the distribution of power across light beams and the dimensions of the functional diameter of that light beam . Use of these more sophisticated beam analyzers has shown that the problem of irradiance inhomogeneity has been further compounded by the introduction of polywave blue–violet LED-based LCUs. For one area of the light tip, the relative contributions of the violet and blue portions of the emitted radiation spectrum to the total irradiance at that point can be dramatically different from another, such that some regions only deliver blue light (∼450–470 nm) and some only violet light (∼400–410 nm) . Additionally, these units deliver a wide range of irradiance values that can vary by more than a factor of 10 across their tip ends.
The combination of polywave LED units and resins using multiple photoinitiators that absorb different wavelengths, and use different filler particle sizes complicates research on dental resins. TPO is highly reactive to wavelengths less than 420 nm and therefore requires a lower irradiance to achieve the same degree of conversion as a CQ containing resin. Therefore, the commonly used averaged irradiance and spectral emission across a LCU cannot fully describe the result the light output has on resin polymerization. Instead, the relative interactions of these inhomogeneous light outputs at the specific locations across resin surfaces that contain different filler particle sizes should be considered.
Surface microhardness using Knoop indenters is a reliable method to determine an important material property of the cured RBC , and a strong positive correlation exists between the degree of monomer conversion and microhardness value . Such microhardness measurements are often taken 1-mm apart across the surface of the specimen . Microtensile bond strength test specimens are commonly sectioned into 1 mm 2 sticks and a previous study showed how the irradiance from a LCU can be reported within specified 1 mm 2 regions across the light guide tip .
The purpose of this study is to use this technique to correlate specific irradiance beam profile values taken through narrow bandpass filters centered at 460 and 400 nm with the microhardness values measured within the same targeted region. The research hypotheses tested were that (1) the localized microhardness maps across the top and bottom surfaces of cured dental RBC specimens will show a positive correlation with the irradiance beam profile values from the LCU taken through bandpass filters, and (2) the strength of the correlation between irradiance and microhardness will decrease with increasing exposure time.
2
Materials and methods
2.1
Characterizing the LCU
One polywave LCU (Bluephase Style LCU, Ivoclar Vivadent, Amherst, NY, USA) that has two LED chips emitting light at 456 nm and one chip emitting at 409 nm, was powered directly from the domestic power line. The light guide used was the original version supplied with the LCU (the manufacturer has subsequently updated their light guide to provide a more homogeneous light output). The irradiance distribution of the light beam from this LCU was measured across the light guide tip using a laser beam profiler . The laser beam profiler used a camera with a 50 mm focal length lens (USB-L070, Ophir-Spiricon, Logan, UT, USA) positioned at a fixed distance from the diffusing surface of a translucent, ground glass target (DG2X2-1500, Thor Laboratories, Newton, NJ, USA). To assess the irradiance distribution at each of the two LED emission wavelengths, the beam profiler was used with the addition of a 10-nm wide interference bandpass filter, centered first at 400 nm and then at 460 nm (Items #65-132 and #88-010, Edmund Industrial Optics, Barrington, NJ, USA), placed in front of the camera lens. The resulting three images were calibrated according to the pixel scale of the camera using software (Beamgage version 5.11 software, Ophir-Spiricon). This calibration enabled precise linear measurements of the images. Prior to each image collection, the UltraCal function in the software was used to subtract ambient light and to normalize all pixels to a similar baseline value. The LCU was then activated, given 5 s to stabilize and then the resulting image was recorded on a personal computer using Beamgage.
The spectral radiant power output from the LCU was recorded five times using a 6-in. integrating sphere (Labsphere, North Sutton, NH, USA) connected to a fiberoptic spectrometer (USB 4000, Ocean Optics, Dunedin, Fla, USA). The overall radiant power, the functional light guide tip end diameter values, and the radiant powers from each individual LED were all entered into the Beamgage software, which then calculated the irradiance at each camera pixel. The scaled numerical data associated with each camera image was exported into a computer graphics software program (Origin Pro v 9.0, OriginLab, Northampton, MA, USA). Using a previously described technique , the ‘functional area’ of the light tip was divided into 45, adjacent and non-overlapping, 1-mm 2 square areas. These 45 areas all fit within the optical area of the LCU tip, but they did not cover the entire tip. The graphics software produced an average local irradiance value for each 1-mm 2 area from the irradiance received by the approximately 4624 individual pixels contained within that area. Scaled irradiance images were also made for the images taken through the 400 nm bandpass filter and the images obtained through the 460 nm bandpass filter. The spectral radiant power from the LCU was measured at three selected locations across the light tip using a 4 mm inner diameter aperture placed in front of the integrating sphere. For one position, the light tip was centered over this aperture, for the others the light tip was moved so the aperture was centered either over the violet or one blue LED chip.
2.2
RBC specimen preparation
To evaluate the effect of the polywave LED light on surface microhardness, four resin-based composites were used: Tetric Evoceram shade A3 [EV-A3]; Empress Direct shade Trans 30 [ED-T30] (Ivoclar Vivadent); Filtek Supreme Plus shade A4B [SP-A4B]; Filtek Supreme Ultra shade CT [SU-CT] (3M ESPE, St. Paul, MN, USA). The two RBCs produced by Ivoclar Vivadent contain both CQ and TPO as co-photoinitiators , while the remaining two RBCs from 3M ESPE contain only CQ and amines (information confirmed by manufacturers). The SU-CT and ED-T30 were translucent shades, while SP-A4B and EV-A3 were less transparent body shades. The RBCs were placed into 1.2-mm thick aluminum rings having an 8.2 mm inner diameter hole. The specimens were intended to be 1.0 mm thick, however at the end of the study, when the thickness of the RBC specimens they was measured with a digital micrometer, they were found to be 1.2 mm thick. The inner diameter hole of the metal ring allowed at least a 1-mm wide circumferential buffer area of RBC between any hardness measurement and the inner wall of the ring, thus reducing the effect the mold might have on resin photo-polymerization . Each ring had a facet to enable precise orientation of the specimens both under the tip end of the LCU, and in the hardness tester. After a Mylar strip was placed over each surface of freshly placed composite, the ring was placed on flat block of previously polymerized Filtek Supreme Ultra shade A2 (3M ESPE) to provide a uniform reflective backing. The curing light was held in a clamp in the same position each time, centered directly over the top surface of the RBC and just out of contact with the Mylar strip. The light guide completely covered the resin specimen. The rings with the RBC specimens could be repeatedly repositioned under the tip end of the LCU and in the hardness tester to a tolerance better than ±0.2 mm.
2.3
Light exposure conditions and microhardness test conditions
The RBC specimens were exposed to light for 5-, 10-, or 30-s from the LCU without the use of any intervening bandpass filters. Once exposed, the specimens were dark-stored for 24 h at room temperature and ambient humidity. The Mylar strips remained on the specimens during storage to minimize the formation of an air-inhibited layer on the RBC surface . After 24 h the Mylar strips were removed. The irradiance grid was projected onto the cured RBC specimen surface and 45 Knoop microhardness measurements were made with a 1 mm pitch at the center of each square in a 7 × 7 grid pattern, to produce a map of the hardness across the top and bottom surfaces. These measurements were made using a Knoop automated hardness tester (HM-123, Mitutoyo Canada Inc., Mississauga, Ont. Canada) that applied a 50 gram (0.49 N) load for 8 s . The Knoop microhardness values were exported into a graphing program (Origin Pro) to produce smoothed two-dimensional color-coded hardness contour maps of the top and bottom of the RBCs.
2.4
Statistical methods
The measurements were repeated five times for each exposure time and RBC in a random order with respect to both conditions. Knoop microhardness values at each of the 45 measurement points on the specimens were analyzed using Repeated Measures Analysis of Variance with a time effect, incorporating a model for the spatial correlation of each KHN point with its neighbors. The top and bottom composite surfaces were analyzed separately. Microhardness values for each of the specimens were also exported into the graphics program (OriginPro V9.0, OriginLab, Northampton, MA, USA). For each set of conditions (RBC, top or bottom surface, and light exposure time), the microhardness values were averaged among the five replications. The mean microhardness values for that square were then used to correlate with the corresponding mean beam profile irradiance values, obtained using the 400 nm and 460 nm bandpass filters for the same square location at the tip of the light guide. The Spearman rank correlation test was used and each observation was replaced with a rank between 1 and 45 from the highest to the lowest KHN value, or from the highest to the lowest irradiance value. This method avoided relying on an assumption of a bivariate normal distribution for the KHN and irradiance. The correlation between microhardness and beam profile, and corresponding statistical significance were determined from these ranks, at a pre-set alpha of 0.05.
2
Materials and methods
2.1
Characterizing the LCU
One polywave LCU (Bluephase Style LCU, Ivoclar Vivadent, Amherst, NY, USA) that has two LED chips emitting light at 456 nm and one chip emitting at 409 nm, was powered directly from the domestic power line. The light guide used was the original version supplied with the LCU (the manufacturer has subsequently updated their light guide to provide a more homogeneous light output). The irradiance distribution of the light beam from this LCU was measured across the light guide tip using a laser beam profiler . The laser beam profiler used a camera with a 50 mm focal length lens (USB-L070, Ophir-Spiricon, Logan, UT, USA) positioned at a fixed distance from the diffusing surface of a translucent, ground glass target (DG2X2-1500, Thor Laboratories, Newton, NJ, USA). To assess the irradiance distribution at each of the two LED emission wavelengths, the beam profiler was used with the addition of a 10-nm wide interference bandpass filter, centered first at 400 nm and then at 460 nm (Items #65-132 and #88-010, Edmund Industrial Optics, Barrington, NJ, USA), placed in front of the camera lens. The resulting three images were calibrated according to the pixel scale of the camera using software (Beamgage version 5.11 software, Ophir-Spiricon). This calibration enabled precise linear measurements of the images. Prior to each image collection, the UltraCal function in the software was used to subtract ambient light and to normalize all pixels to a similar baseline value. The LCU was then activated, given 5 s to stabilize and then the resulting image was recorded on a personal computer using Beamgage.
The spectral radiant power output from the LCU was recorded five times using a 6-in. integrating sphere (Labsphere, North Sutton, NH, USA) connected to a fiberoptic spectrometer (USB 4000, Ocean Optics, Dunedin, Fla, USA). The overall radiant power, the functional light guide tip end diameter values, and the radiant powers from each individual LED were all entered into the Beamgage software, which then calculated the irradiance at each camera pixel. The scaled numerical data associated with each camera image was exported into a computer graphics software program (Origin Pro v 9.0, OriginLab, Northampton, MA, USA). Using a previously described technique , the ‘functional area’ of the light tip was divided into 45, adjacent and non-overlapping, 1-mm 2 square areas. These 45 areas all fit within the optical area of the LCU tip, but they did not cover the entire tip. The graphics software produced an average local irradiance value for each 1-mm 2 area from the irradiance received by the approximately 4624 individual pixels contained within that area. Scaled irradiance images were also made for the images taken through the 400 nm bandpass filter and the images obtained through the 460 nm bandpass filter. The spectral radiant power from the LCU was measured at three selected locations across the light tip using a 4 mm inner diameter aperture placed in front of the integrating sphere. For one position, the light tip was centered over this aperture, for the others the light tip was moved so the aperture was centered either over the violet or one blue LED chip.
2.2
RBC specimen preparation
To evaluate the effect of the polywave LED light on surface microhardness, four resin-based composites were used: Tetric Evoceram shade A3 [EV-A3]; Empress Direct shade Trans 30 [ED-T30] (Ivoclar Vivadent); Filtek Supreme Plus shade A4B [SP-A4B]; Filtek Supreme Ultra shade CT [SU-CT] (3M ESPE, St. Paul, MN, USA). The two RBCs produced by Ivoclar Vivadent contain both CQ and TPO as co-photoinitiators , while the remaining two RBCs from 3M ESPE contain only CQ and amines (information confirmed by manufacturers). The SU-CT and ED-T30 were translucent shades, while SP-A4B and EV-A3 were less transparent body shades. The RBCs were placed into 1.2-mm thick aluminum rings having an 8.2 mm inner diameter hole. The specimens were intended to be 1.0 mm thick, however at the end of the study, when the thickness of the RBC specimens they was measured with a digital micrometer, they were found to be 1.2 mm thick. The inner diameter hole of the metal ring allowed at least a 1-mm wide circumferential buffer area of RBC between any hardness measurement and the inner wall of the ring, thus reducing the effect the mold might have on resin photo-polymerization . Each ring had a facet to enable precise orientation of the specimens both under the tip end of the LCU, and in the hardness tester. After a Mylar strip was placed over each surface of freshly placed composite, the ring was placed on flat block of previously polymerized Filtek Supreme Ultra shade A2 (3M ESPE) to provide a uniform reflective backing. The curing light was held in a clamp in the same position each time, centered directly over the top surface of the RBC and just out of contact with the Mylar strip. The light guide completely covered the resin specimen. The rings with the RBC specimens could be repeatedly repositioned under the tip end of the LCU and in the hardness tester to a tolerance better than ±0.2 mm.
2.3
Light exposure conditions and microhardness test conditions
The RBC specimens were exposed to light for 5-, 10-, or 30-s from the LCU without the use of any intervening bandpass filters. Once exposed, the specimens were dark-stored for 24 h at room temperature and ambient humidity. The Mylar strips remained on the specimens during storage to minimize the formation of an air-inhibited layer on the RBC surface . After 24 h the Mylar strips were removed. The irradiance grid was projected onto the cured RBC specimen surface and 45 Knoop microhardness measurements were made with a 1 mm pitch at the center of each square in a 7 × 7 grid pattern, to produce a map of the hardness across the top and bottom surfaces. These measurements were made using a Knoop automated hardness tester (HM-123, Mitutoyo Canada Inc., Mississauga, Ont. Canada) that applied a 50 gram (0.49 N) load for 8 s . The Knoop microhardness values were exported into a graphing program (Origin Pro) to produce smoothed two-dimensional color-coded hardness contour maps of the top and bottom of the RBCs.
2.4
Statistical methods
The measurements were repeated five times for each exposure time and RBC in a random order with respect to both conditions. Knoop microhardness values at each of the 45 measurement points on the specimens were analyzed using Repeated Measures Analysis of Variance with a time effect, incorporating a model for the spatial correlation of each KHN point with its neighbors. The top and bottom composite surfaces were analyzed separately. Microhardness values for each of the specimens were also exported into the graphics program (OriginPro V9.0, OriginLab, Northampton, MA, USA). For each set of conditions (RBC, top or bottom surface, and light exposure time), the microhardness values were averaged among the five replications. The mean microhardness values for that square were then used to correlate with the corresponding mean beam profile irradiance values, obtained using the 400 nm and 460 nm bandpass filters for the same square location at the tip of the light guide. The Spearman rank correlation test was used and each observation was replaced with a rank between 1 and 45 from the highest to the lowest KHN value, or from the highest to the lowest irradiance value. This method avoided relying on an assumption of a bivariate normal distribution for the KHN and irradiance. The correlation between microhardness and beam profile, and corresponding statistical significance were determined from these ranks, at a pre-set alpha of 0.05.
3
Results
3.1
Irradiance beam profile and spectral emission
An image of the tip end of the Bluephase Style LCU light guide ( Fig. 1 a) shows the three well-separated LED chips visible through the light guide. The spatial separation visible at the tip end is strong evidence of the non-uniformity of the beam profile from this LCU. The irradiance beam profile measured without any bandpass filter ( Fig. 1 b) is a near identical image of the tip end, and three ‘hot spots’ in irradiance can be directly associated with the locations of the three individual LED chips within the LCU head. The mean radiant power (± standard deviation) delivered by the Bluephase Style LCU was 671 ± 5 mW. Using a 9 mm functional diameter of the light guide, an averaged irradiance value across the light tip was calculated to be 1055 ± 8 mW/cm 2 . Thus, the specimens received 5.3 J/cm 2 after 5-s, 10.6 J/cm 2 after 10-s, and 31.7 J/cm 2 after 30-s of light exposure. Fig. 1 c and d show the irradiance beam profiles imaged through the 460 nm and 400 nm bandpass filters respectively. Although the unfiltered irradiance beam profile in Fig. 1 b is non-uniform, the irradiance beam profiles measured through the bandpass filters are strikingly inhomogeneous, displaying ‘hot’ spots and large ‘cold’ regions in irradiance. Using the 460 nm bandpass filter, two irradiance ‘hot spots’ are observed, attributed to the two blue 456 nm LED chips ( Fig. 1 c). Using the 400 nm bandpass filter indicates that there is a single 2.7 mm diameter ‘hot spot’ in irradiance superimposed onto the remaining 9 mm diameter low irradiance ‘cold’ region ( Fig. 1 d). This irradiance ‘hot spot’ is attributed to the violet 409 nm LED chip within the LCU head. For the three locations shown in Fig. 2 , the spectral radiant power measured through the 4 mm diameter aperture varied dramatically. The fraction of the total radiant power emitted below 425 nm ranged from 2.4% to 90% depending on the position of the aperture in front of the tip end of the light guide. The spectra confirmed that the top LED chip mostly radiates near 409 nm, while the bottom two LED chips emit near 456 nm.
Each 1-mm 2 square contained approximately 4624 camera pixels and the mean irradiance values within each of the squares for the two spectral emission ranges are shown in Fig. 3 . When imaged through the 460 nm bandpass filter (Fig 3a), the irradiance was as high as 5424 mW/cm 2 , related to an irradiance ‘hot spot’, and as low as 38 mW/cm 2 , attributed to a ‘cold region’ in the irradiance beam profile. When imaged through the 400 nm bandpass filter (Fig 3b), the local irradiance values were found to be as high as 1878 mW/cm 2 , related to an irradiance ‘hot spot’, and as low as 3 mW/cm 2 .
3.2
Microhardness maps
The microhardness maps of cured RBC samples ( Fig. 4 ) illustrate how the same polywave LCU has different effects on surface hardness of various RBC systems. A visual comparison between the irradiance beam profiles made of the Bluephase Style LCU through the two bandpass filters ( Fig. 3 ) indicates a positive correlation with the localized microhardness maps ( Fig. 4 ) when all the RBCs were exposed for 5- and 10-s. The irradiance values from the two irradiance beam profiles ( Fig. 3 a and b) were then ranked and used to correlate with the ranked values from the microhardness maps.
