d e n t a l m a t e r i a l s 3 1 ( 2 0 1 5 ) 339–350

Available online at www.sciencedirect.com

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

Robust spectrometer-based methods for characterizing radiant exitance of dental LED light curing units Adrian C. Shortall a , Christopher J. Felix b , David C. Watts c,∗ a b c

The Dental School, St. Chad’s Queensway, Birmingham B4 6NN, UK Bluelight Analytics Inc., 24-2625 Joseph Howe Dr, Halifax, NS B3L 4G4, Canada The University of Manchester: School of Dentistry and Photon Science Institute, Manchester M13 9PL, UK

a r t i c l e

i n f o

a b s t r a c t

Article history:

Objectives. Firstly, to assess light output, from a representative range of dental light curing

Received 10 February 2015

units (LCUs), using a new portable spectrometer based instrument (checkMARCTM ) com-

Received in revised form

pared with a “gold standard” method. Secondly, to assess possible inconsistency between

25 February 2015

light output measurements using three different laboratory-grade thermopile instruments.

Accepted 26 February 2015

Methods. The output of four blue-dental LCUs and four polywave blue-and-violet-LCUs was measured with two spectrometer-based systems: a portable spectrometer instrument and a benchtop Integrating Sphere fiber-coupled spectrometer system. Power output was

Keywords:

also recorded with three thermopiles according to ISO 10650-2. Beam profile images were

Dental Light curing unit

recorded of LCU output to assess for spatial and spectral beam uniformity.

Thermopile

Results. Power recorded with the portable spectrometer instrument closely matched the ‘gold

Spectrometer

standard’ Integrating Sphere apparatus calibrated according to International Standards.

ISO 10650-2

Radiant exitance for the eight LCUs differed significantly between the three thermopiles.

Radiant exitance

Light source to thermopile sensor distance influenced recorded power significantly (p < 0.05),

Irradiance

indicating the severe limitations of thermopiles for absolute measurements. Polywave LCU

Spectral radiant power

beam profiles demonstrated output spectral heterogeneity. Significance. Spectrometer-based methods are capable of overcoming the limitations inherent with thermopile-based measurement techniques. Spectrometer based measurements can fulfill the intention of ISO 10650. © 2015 Academy of Dental Materials. Published by Elsevier Ltd. All rights reserved.

1.

Introduction

1.1.

The challenge of photo-polymerization in dentistry

Effective photo-polymerization is a prerequisite for the long term clinical success and quality of light-activated

∗

resin-based composite (RBC) restorations. Amongst the factors which influence the polymerization efficiency of (multi-)methacrylate RBCs, the light curing unit (LCU) is a key extrinsic variable [1,2]. Whilst quartz tungsten halogen (QTH) LCUs have been the mainstay of dental practice for decades [3] blue light emitting diode (LED) units have replaced them

Corresponding author at: University of Manchester School of Dentistry, Oxford Road, Manchester M13 9PL, UK. Tel.: +44 1612756749. E-mail address: [email protected] (D.C. Watts).

http://dx.doi.org/10.1016/j.dental.2015.02.012 0109-5641/© 2015 Academy of Dental Materials. Published by Elsevier Ltd. All rights reserved.

340

d e n t a l m a t e r i a l s 3 1 ( 2 0 1 5 ) 339–350

as the most popular type because they are compact, portable, energy efficient and have potentially superior service life [4,5].

1.2. The challenges of measuring radiant output of dental LED-LCUs Unlike point light sources or collimated laser sources, dental LED-LCUs pose unique challenges. Hand-held dental radiometers commonly used for measuring irradiance are inaccurate and may vary within the same model [6]. Results from such ‘dental radiometers’ are influenced by differences in measurement time, spectral sensitivity and mismatch between the instrument sensor aperture and the LCU area [7–9]. Attention to detail is required for accurate performance of all radiometric measurements. Meaningful reporting of radiometric data requires appropriate nomenclature and precise physical definitions of key terms such as radiant exitance [10]. The diversity of: (i) LCU types (gun or pen style, corded or cordless, fixed or removable lens cap or fiber optic light guide) and (ii) irradiation protocols available (pulse, soft start and modulated cure) complicate the measurement of their power output, spectral and irradiance characteristics. The following major challenges exist: • The diameter of the LCU optic or exit window may range from 10 mm and there may be non-uniformity of the light beam-profile. • The target surface area (diameter) will vary with different clinical scenarios or laboratory test setup requirements. • The LCU optic – “target” distance can vary in clinical practice from zero to 10 mm and the irradiance declines over this distance. • LED-LCUs may incorporate one or more LED-chips each outputting a different wavelength range. • The irradiation time can vary from 1 s to 20 s or more. Power output varies over time with pulsed or modulated cure protocols. The available instrumentation for measuring output includes (i) light collection devices (cosine correctors, integrating spheres), (ii) bandpass filters to isolate specific spectral regions, (iii) thermopiles, (iv) spectrometers. The selection of appropriate combinations is challenging for the reasons outlined above. Integrating sphere light collection devices coupled to spectro-radiometers are considered as the gold standard for the optical characterization of dental LCUs in the laboratory. Total radiant power and spectral power as a function of wavelength may be determined. Recently introduced modified integrating sphere instruments (checkMARCTM from Bluelight analytics and the larger

BTS256-Hi from Gigahertz-Optik GmbH, Türkenfeld, Germany) make it practical to conduct such measurements in dental practice.

1.3. Consideration of the existing ISO Standard for LED-LCUs ISO 4049 [11] is concerned with polymer based restorative materials and ISO 10650-2 [12] exists to standardize the requirements and test methods for light curing units (LCUs). ISO 10650-2 outlines a standardized protocol for measuring the output of dental LED-LCUs. It addresses some of the issues mentioned above but fails to deal with others. Because of the relatively slow response-time inherent with thermopile detectors, LCUs with short irradiation times cannot be measured reliably. When ISO 10650-1 was published in 2004 it identified 400–515 nm as the blue region for LCUs. The standard was published before multi-peak or polywave LED-LCUs were marketed, designed to cure materials containing alternative photo-initiators. The cut-off point between violet and blue light may be identified as 425 nm spectrally. The development of high irradiance (>1500 mW/cm2 ) LED-LCUs with very short irradiation times coupled with the introduction of spatially non-uniform multi-wavelength sources have increased the need to revise these standards for contemporary dental practice. ISO 10650-2 (2007) specifies measurement of radiant exitance for three discrete wavelength regions using a series of four long-pass filters (Table 1). The ISO 10650-1 source document requires that the entrance aperture of the radiometer be greater than the cross-sectional area of the optic tip (light-guide exit window) and that the edge of the optic tip should be at least 2 mm from the edge of the radiometer entrance aperture so that all radiant emission is measured by the radiometer. The protocol involves sequentially measuring radiant exitance with specified filters and calculating and reporting light transmission in the ranges (a) 190–385 nm, (b) 400–515 nm and (c) above 515 nm. Region (a) includes not only the ultraviolet region but also the near-blue wavelength region of around 380 nm. Unfortunately the 385–400 nm region is unaccounted for, although several dental LED-LCU sources emit considerable radiant energy in this region, since the 385–415 nm bandwidth is important for some alternative photo-initiators. In Section 5.2.2, it specifies that the radiant exitance between 190 and 385 nm should be less than 200 mW/cm2 . The region above 515 nm reaches approximately to 1100 nm, which is the detection limit of the measurement device specified in the Standard. According to Section 5.2.3, the upper limit for radiant exitance above 515 nm should be not greater than 100 mW/cm2 . No requirement is given for radiant exitance (or power) in the 400–515 nm wavelength region

Table 1 – Long-pass filters according to ISO 10650-2 and wavelength regions of interest. NB In our work, measurements were also conducted for wavelengths between 190 and 515 nm, using filters a and d. Filter type

Quartz (SQ1)

Schott GG385

Schott GG400

Transmission

a > 190 nm

b > 385 nm

c > 400 nm

[0,1-2]ISO 10650-2 procedures and corresponding wavelength regions (nm)

a–b 190–385

c–d 400–515

d >515

Schott OG515 d > 515 nm

341

d e n t a l m a t e r i a l s 3 1 ( 2 0 1 5 ) 339–350

Table 2 – LED light curing units used in this study and their active optic areas. Each unit was operated and measured using its ‘standard’ output mode, where more than one mode was available, corresponding to the manufacturer-stated irradiances. Light curing unit (LCU) Bluephase G2 Bluephase Style SmartLite maX Valo Cordless Elipar S10 Flash Max P3 460 4W Radii Plus S.P.E.C. 3

Manufacturer

Serial number

Polywave

Ivoclar Vivadent Ivoclar Vivadent Dentsply Ultradent 3M ESPE CMS dental SDI Coltene/Whaledent

223306 1100001225 003512 C01535 9391120138 12FMP0072 56497 121110267

Y Y Y Y N N N N

which is the blue light region for polymerization activation. To fulfill the test requirement, the radiant exitance in the 400–515 nm region must not be less than “the manufacturer’s stated value” when measured in accordance with Section 7.2. A micrometer, with an accuracy of 0.02 mm, is used to measure the diameter of the optic tip, if it has a circular cross-section, although this may not allow for the presence of any optic cladding. For elliptical cross-section tips the major and minor axes are measured to calculate the optical crosssectional area. The Standard does not specify LCU-to-detector distance and acknowledges that the results do not represent true radiant exitance or analogously irradiance. Source distance is defined by the radiometer employed and by the optical filter thickness.

1.4. Assessment of alternative spectrometer-based systems for LED-LCUs The focus of this work was to assess spectrometer-based extra-oral methods for measuring radiant exitance of LEDLCUs. The aim was to see if spectrometer based systems could overcome the limitations inherent in the current ISO 10650-2 test method. A specific aim was to assess how a portable spectrometer instrument (checkMARCTM ) performed compared to a ‘gold standard’ integrating-sphere based spectrometer method. A secondary objective was to assess how thermopile design impacted on power measurement according to ISO 10650-2. Hypotheses: 1. Mean power, as measured with the portable spectrometer instrument (checkMARCTM ), will not vary by more than 5% from the Integrating Sphere/spectrometer ‘gold standard’ reference value for any of the eight representative LEDLCUs. 2. The mean radiant exitance or irradiance values, from each of eight representative LED-LCUs), as determined with the method described in ISO 10650-2 for the wavelength range between 400 nm and 515 nm, using three different laboratory grade thermopiles, will agree between the instruments to within 5%.

2.

Materials and methods

Eight contemporary LED-LCUs, representative of the range of different types of unit currently available, were used in this

Stated irradiance (mW/cm2 ) 1200 ± 10% 1100 ± 10% 1200 1000 1200 ± 10% 5000–6000 1500 1600

Measured active optic area (cm2 ) 0.646 0.622 1.626 0.750 0.622 0.720 0.443 0.407

work (Table 2). As shown in the final column of Table 2, the active or functional optic areas ranged from 0.41 to 1.63 cm2 . These were determined by analysis of captured images of the end-optics and excluded any optic cladding areas, if incorporated in the optic design. Four of the LCUs were single-peak or monowave blue LED sources and the manufacturers of the other four units (Bluephase G2, Bluephase Style, Smartlite maX and Valo cordless) indicated that their units were broad spectrum and thus were multi-peak or polywave with output in violet as well as blue wavelengths. None of the manufacturers of the above LCUs claimed conformity with ISO 10650-2 for their unit. Radiant exitance refers to the magnitude of light emitted from unit area of the source, whereas irradiance refers to the radiant power received by a surface. The latter is more clinically relevant given that it refers to the radiant power that impacts the material surface. Although both quantities have the same units: mW/cm2 , they can only be considered to be numerically equal at zero distance, when the emitted light is received directly from the LCU by the target material or detector. At non-zero distances, the irradiance from most LED devices attenuates with target distance as the light spreads out over a wider area. Some of the units had more than one output (or mode) setting, but the selected (standard) mode and the respective irradiances, stated by the manufacturer of each unit, are tabulated. Wherever possible, LED-LCUs were measured in the highest power mode which allowed 10 s continuous irradiation to comply with ISO 10650-2. In the case of one LCU (Flashmax P3), where the longest continuous irradiation time was 3 s, the output was measured after 3 s activation, where meter response-time allowed.

2.1.

Spectrometer measurements

Two spectrometer systems, each calibrated to international standards, were compared, namely:

(a) a ‘gold standard’ integrating sphere system. This consisted of a laboratory-grade integrating sphere (IS: Labsphere sn: 1229110001), 15 cm in diameter, with a 19 mm entrance port linked to a calibrated spectro-radiometer (Ocean Optics USB 4000 sn: USB4F08680). This was used to register total power, within the 360–540 nm range, and spectra for each LED-LCU (n = 5 measurements per LCU).

342

d e n t a l m a t e r i a l s 3 1 ( 2 0 1 5 ) 339–350

(b) a portable calibrated instrument (checkMARCTM ; Bluelight Analytics, Inc., Halifax, NS B3L 4G4, Canada). This featured a 16 mm diameter active light collection area, incorporating a sealed Teflon diffuser window. This allowed light transmission into the light collection chamber attached via a fiber optic connector to an inbuilt (STS, Ocean Optics) spectrometer assembly. The device was calibrated using a NIST reference light source (HL2000, Ocean Optics) and spectral radiant power measurements were taken using SpectraSuite software (Ocean Optics). The total spectral radiant power output from each LCU was divided by the cross sectional active area (cm2 ) of each light tip to calculate the average irradiance.

The calibrated checkMARCTM radiometer allowed irradiance to be determined for any currently available dental LCU, irrespective of its exit-window dimensions. Irradiance and spectra for each LED-LCU were recorded at zero source-tosensor distance (n = 5 measurements per LCU).

2.2.

Thermopile measurements

Using three types of thermopile detector, power measurements were also undertaken according to ISO 10650-2. These thermopiles were: (a) PM10-19C (sn: 1168344); (b) PM10 (sn: 0789J11R), Coherent Inc., and (c) S310C-SP4 (sn: 1208132; ThorLabs). They had similar sensor diameters (19–20 mm) but had different well depths: 6, 11.5 and 16 mm respectively. The Coherent PM10-19C had a USB connection that housed the signal processing and microelectronics, eliminating the need for a separate meter. Circular ISO-specified Schott long-wavelength pass (LWP) filters of diameter 25 mm (SQ1, GG385, GG400, OG515; thickness 3.0 ± 0.1 mm) were placed sequentially (as per Table 1) above each thermopile, to take power readings as per ISO 10650-2. With these in position, the LCU optic-to-sensor distances were 9 mm for the PM10-19C, 14.5 mm for the PM10 and 19 mm for the S310C-SP4. Power outputs recorded with the LWP filters were used to calculate radiant exitance in the region 190–400 nm, to determine the output below 400 nm for the LCUs compared with the blue light output in the 400–515 nm range, as specified in the Standard. Additional measurements, beyond those specified in ISO 10650-2, were undertaken with the thermopiles to understand: (i) how the presence of the LWP filters and (ii) light source-to-sensor distance influenced the power measured for all LCUs and thermopile models. Thus the measurements were repeated at corresponding light source-to-sensor distances for each thermopile/LCU combination, but without any interposed filters. This allowed the influence of the filter on recorded power output to be determined. The light sources were also measured with their exit windows positioned at 1 mm distance from the thermopile sensors when light unit and thermopile design permitted. For this experiment, one of the thermopiles (Thorlabs S310C-SP4) was modified, before recalibration, to incorporate a 19 mm wide recess in the meter housing to allow access for all the LCU optics to 1 mm from the sensor face. This modified sensor was used in conjunction with a PM-200 meter.

Fig. 1 – Integrating Sphere spectra for all LED LCUs taken in standard output modes.

2.3.

Beam profile measurements

Beam profiling, widely used for characterization of laser beams, has been adapted for dental curing LCUs. The variables considered include performance over clinically relevant distances, spectral inhomogeneity and identification of hot (high irradiance) and cold (low irradiance) areas. Beam profile images were recorded of all the LCUs with the optic rod of each unit placed against the surface of a frosted glass screen target (DG2X2-1500, Thor laboratories, Newark, NJ, USA) and also with the optic at 4 and 10 mm from the target. The method has been described previously [13]. The images allowed an assessment to be made of beam heterogeneity over the active area of the LCU exit window as well as the decrease in beam irradiance with distance onto a central 4 mm diameter target area. Each image was individually calibrated, according to the power recorded for each LCU with the Integrating Sphere, by setting the highest level measured to assume the highest array count in the Beamgage Standard (v.5.6: Ophir Spiricon) software. Spectral irradiance for violet (400–410 nm) and blue (455–465 nm) wavelength regions were recorded and calibrated on separate images with the aid of appropriate narrow (FWHM = 10 nm) bandpass filters. This allowed an assessment to be made of the wavelength distributions for violet and blue wavebands across the exit windows of the LED sources.

3.

Results

Fig. 1 shows the spectral outputs of the eight LED-LCUs used in this investigation, as measured by the integrating sphere equipment. This confirmed the multi-peak or polywave spectral output for four of the eight LCUs. Taking 425 nm as the cut-off point between the violet and blue LED chipsets (Fig. 1) the percentage of violet output relative to blue light for the four polywave LED-LCUs ranged from 15% for the Bluephase G2 to 35% for the Smartlite maX whereas the four monowave blue LED units had very low outputs in the violet region (Table 3). Spectrometer data from both Integrating Sphere (IS) and the checkMARCTM (cM) devices are presented in Table 4 (columns 3–4) expressed both as Power and as Radiant Exitance (or Irradiance), for each LCU at ‘zero’ distance. Fig. 2

LCU Bluephase G2 Bluephase Style SmartLite maX Valo Cordless Elipar S10 FlashMax P3 Radii Plus S.P.E.C. 3

$$$

Measurement (units) Power (W) Irradiance (mW/cm2 ) Power (W) Irradiance (mW/cm2 ) Power (W) Irradiance (mW/cm2 ) Power (W) Irradiance (mW/cm2 ) Power (W) Irradiance (mW/cm2 ) Power (W) Irradiance (mW/cm2 ) Power (W) Irradiance (mW/cm2 ) Power (W) Irradiance (mW/cm2 )

checkMARCTM at 0 mm 0.86 1337 0.69 1104 0.76 465 0.80 1064 0.76 1217 0.97 1344 0.46 1028 0.76 1866

Integrating Sphere at 0 mm 0.85 1319 0.69 1110 0.73 446 0.73 967 0.75 1204 1.01 1396 0.45 1021 0.76 1859

PM10-19C Distance = 9 mm 0.64 993 0.51 813 0.34 209 0.49 658 0.56 903 0.54 753 0.25 562 0.47 1165

PM10 Distance = 14.5 mm 0.63 969 0.45 720 0.22 133 0.45 600 0.49 783 $$$ $$$

0.19 433 0.31 769

S310C Distance = 19.0 mm 0.60 921 0.37 591 0.18 108 0.43 573 0.47 748 0.37 509 0.17 392 0.25 605

343

Where stable and reliable readings could not be obtained for the Flashmax P3 unit because of its short activation time and the relatively slow response of the PM10/Fieldmax TO thermopile sensor/meter combination.

d e n t a l m a t e r i a l s 3 1 ( 2 0 1 5 ) 339–350

Blue (425–515 nm)

84.6% 82.5% 64.6% 79.8% 98.4% 99.1% 99.1% 99.1%

Measurement device = Integrating Sphere

Violet (385–425 nm)

15.4% 17.5% 35.4% 20.2% 1.6% 0.9% 0.9% 0.9%

Table 3 – The percentage of the spectral radiant power (W) from eight LCUs in the violet spectral region (385–425 nm) and the percentage of the spectral radiant power in the blue spectral region (425–515 nm) for the eight LCUs as determined with the Integrating Sphere.

LCU

Bluephase G2 Bluephase Style SmartLite maX Valo Cordless Elipar S10 FlashMax P3 Radii Plus S.P.E.C. 3

For the four polywave LCUs the violet output ranged from 15% to 35% of the total output whereas the monowave units had less than 1% of their output in the violet region apart from the Elipar S10 LCU at 1.6%.

Fig. 2 – Correlation plots for (a) power and (b) radiant exitance for measurements with cM versus IS instruments.

Table 4 – checkMARCTM and Integrating Sphere power and irradiance data, all measured at 0 mm distance. Also, data obtained with three thermopile devices at varied distances, for the wavelength range 400–515 nm, using the filter types as described in Table 1 and in ISO 10650-2.

344

d e n t a l m a t e r i a l s 3 1 ( 2 0 1 5 ) 339–350

shows the high correlations (r2 = 0.99) observed between cM and IS spectrometer measurements for the set of LCUs, both for Power (Fig. 2(a)) and Radiant Exitance (Fig. 2(b)). In Table 4 (columns 5–7) the outputs, according to ISO 10650-2 thermopile measurements, are also listed between 400 nm and 515 nm (‘blue light’ values), using each of the three thermopiles, at the stated LCU optic-to-sensor distances. Table 5 gives the mean power values for the LCUs with each of the three thermopiles: (i) between 400 and 515 nm and (ii) between 190 and 515 nm, (iii) at corresponding LCU to sensor distances as in (i) and (ii), but without any LWP filters and (iv) without filters at only 1 mm LCU to sensor distance, whenever possible. Table 6 shows the mean irradiance for all LCUs, measured according to ISO 10650-2, for the wavelength ranges: (i) 515 nm. The energy transmitted above 515 nm as a percentage of the total radiant exitance was less than 2% for all the LCUs and ranged from 0.7% to 1.9% for the LCU-thermopile combinations. Because the measurements were so repeatable, data scatter was small (coefficients of variation PM10 > S310 C with statistical significance (p < 0.05) between each meter with every LCU. Differences between high and low ranking meters ranged from 7% for the Bluephase G2 to 48% for the Smartlite maX and S.P.E.C. 3 units. The mean irradiance measured in accordance with ISO 10650-2 (between 400 nm and 515 nm) with the three different thermopile instruments fell below the Manufacturer Stated (MS) irradiance value in every case (Tables 2 and 4). Irradiance for the Bluephase G2 LCU was closest to MS value for all thermopiles (77–83% of MS). Three LCUs (Flashmax P3, Radii Plus and Smartlite maX) recorded irradiance values below 50% of MS irradiance, for all thermopile instruments used in accordance with ISO 10650-2, and a further unit (S.P.E.C. 3) did so with two (PM10 and S310C) out of the three meters. The mean irradiance values as a percentage of the MS irradiance (assuming a minimum value of 5000 mW/cm2 ) for the Flashmax P3 with the PM10-19C and Thor S310C thermopiles were 15.1% and 10.2% respectively.

Mean irradiance values recorded with the light sources positioned at 1 mm distance without any intervening filter to the thermopile sensor faces exceeded 90% on average of the Integrating Sphere determined irradiance values for all LCUs. The results (Tables 2 and 5) illustrate that the LCU Manufacturer Stated Irradiance is commonly not indicative of the LCU’s Radiant Exitance as determined by ISO 10650-2 method. The influence was appreciable of the SQ1 LWP filter (which transmits wavelengths >190 nm) on power registered by the different thermopiles for the different LED-LCUs (Table 7). Significantly greater (p < 0.05) mean power was recorded for some LED-thermopile combinations compared to measurements at the same distance, without the SQ1 filter in place. However, the

opposite was the case for other combinations. A fixed model GLM ANOVA of mean power recorded for all LCUs according to meter type (3 levels), and the presence or absence of SQ1 LWP filter (2 levels) revealed that meter type was significant (F = 23.92, p < 0.001) whereas there was no overall significant effect for filter presence (F = 0.00, p = 0.962) and the interaction term also did not reach significance (F = 0.32, p = 0.724). For the PM10-19C sensor (which had the lowest source-to-sensor distance of 9 mm), mean power across all units dropped slightly with the SQ1 LWP filter present (versus no filter). But for the other two meters, the opposite trend was observed. As light source-to-sensor distance increased, using different thermopiles, this impacted on recorded power values. So a complex interaction was observed between LCU type, tip design and filter interposition for different source-to-sensor distances. Significant (p < 0.05) decreases in measured power, from values at 1 mm, were observed (Table 5 column -6) in comparison to the distance from the sensor dictated by quartz (Schott SQ1) LWP filter placement (Table 7 column 3). Whilst statistically significant (p < 0.05) differences were found between individual thermopiles, at the 1 mm source-to-sensor distance (Table 5), these were insignificant in comparison to the over-riding effect of source-to-sensor distance when the LWP filters were interposed, as required according to the ISO 10650-2 method.

3.1.

Beam profile data

The beam profile images, shown in Figs. 3a–h and 4a, b respectively, demonstrate the detailed spectral distribution patterns for the non-uniform irradiance output for all 8 LED-LCUs and account for the disparities recorded in the mean irradiance for the lights when measured either at full light source exit window aperture or through a 4 mm diameter sensor. The active areas of the LED light source exit windows are outlined in pink on the images. Clinically, the irradiance from the LCU is most often assumed to be uniform but clearly, depending on the LCU optical configuration, the spectral distribution and the positioning of the light tip, there are hot and cold areas within the active area of the light tip that will cause a significant variation in light delivery to the material. This reflects the possible variation in energy delivery to an ISO depth-of-cure specimen when the light source has a large exit area and no alignment ring is used to center the light tip over the specimen. This is most likely to occur with those units (Bluephase

346

d e n t a l m a t e r i a l s 3 1 ( 2 0 1 5 ) 339–350

Table 7 – Power (W) data for each LCU, recorded with each of the three thermopiles, with and without an overlying SQ1 Quartz filter at corresponding distances as dictated by the ISO standard and meter design. LCU

Thermopile model

Power (W) measured with SQ1 filter

Power (W) measured without SQ1 filter

Bluephase G2

PM10-19C PM10 S310C

0.730 (0.002) 0.712 (0.005) 0.678 (0.002)

0.776 (0.002) 0.713 (0.001) 0.657 (0.002)

Bluephase Style

PM10-19C PM10 S310C

0.581 (0.002) 0.507 (0.002) 0.425 (0.003)

0.618 (0.001) 0.541 (0.003) 0.388 (0.001)

SmartLite maX

PM10-19C PM10 S310C

0.478 (0.007) 0.307 (0.002) 0.246 (0.001)

0.472 (0.004) 0.250 (0.001) 0.225 (0.001)

Valo Cordless

PM10-19C PM10 S310C

0.587 (0.001) 0.533 (0.001) 0.503 (0.001)

0.615 (0.001) 0.521 (0.002) 0.505 (0.001)

Elipar S10

PM10-19C PM10 S310C

0.614 (0.001) 0.528 (0.001) 0.511 (0.001)

0.642 (0.000) 0.557 (0.001) 0.495 (0.001)

FlashMax P3

PM10-19C PM10 S310C

0.591 (0.005) – 0.394 (0.002)

0.616 (0.002) – 0.410 (0.002)

Radii Plus

PM10-19C PM10 S310C

0.269 (0.001) 0.212 (0.001) 0.190 (0.002)

0.279 (0.001) 0.218 (0.001) 0.191 (0.004)

S.P.E.C. 3

PM10-19C PM10 S310C

0.520 (0.001) 0.347 (0.008) 0.274 (0.001)

0.529 (0.001) 0.318 (0.001) 0.227 (0.001)

Reliable readings could not be obtained for the Flashmax P3 unit with the PM10 meter because of its short activation time and the relatively slow response for the PM10/Fieldmax TO thermopile sensor/meter combination.

Style, Smartlite maX, FlashMax P3, Radii and S.P.E.C. 3) which had highly non-uniform irradiance output as shown by the beam profile images (Figs. 3b, c, f–h and 4b). The beam profile image of the FlashMax P3 unit at 0 mm target to source distance shows how most of the light is concentrated within the 4 mm frustum shaped tip but light output was significantly reduced from source because of the tip (Fig. 3f). Whilst some light was also transmitted through the peripheral tip surround, on which the irradiance calculation has been based, it is of such a low level as to be clinically irrelevant.

4.

Discussion

The checkMARCTM spectrometer-based measurements corresponded very closely to the Integrating Sphere (IS) measurements (Pearson correlation > 0.99). With checkMARCTM , mean power values for the LCUs were typically within 2% of the corresponding Integrating Sphere values (Table 4). An exception was the power of the Valo LCU measured on checkMARCTM that was significantly greater (p < 0.05) by 10% than with the Integrating Sphere. This is because the Valo has a convex glass tip and so, when the leading edge of the exit window was placed level with the central portion of the IS port, and not within the sphere, not all the light output was recorded. With

checkMARCTM , light falls directly onto its recessed diffuser surface. Thus the checkMARCTM was the more accurate device in this case. Additional work revealed that when a Valo LCU had its convex glass tip placed completely within the IS port entrance, the mean irradiance increased by 11% compared to the value recorded with the front edge of the convex lens level with the IS entrance. The large disparity between the irradiance values reported here for the FlashMax P3 unit and the manufacturer-stated irradiance limits lies in the way irradiance was determined for this unit. We evaluated irradiance for this LCU based on the full area of the base of the translucent plastic disposable ‘4 mm’ blunt light tip. In addition to the results already mentioned, we also measured power output from this unit, without the disposable sleeve/4 mm light tip in place, both on the IS and the power meters. The IS method was the only way we were able to obtain power outputs that yielded a mean irradiance value (5976 mW/cm2 ) in line with the manufacturer’s claims of irradiance up to 6000 mW/cm2 . The manufacturerclaimed irradiance is impossible to obtain when the unit is used according to recommendations for clinical use that specifically state that the LCU should be used with the light tip and disposable protective barrier. A similar situation existed for the Smartlite maX unit where the manufacturer-stated irradiance relates to only the inner part of the exit window area.

d e n t a l m a t e r i a l s 3 1 ( 2 0 1 5 ) 339–350

347

Fig. 3 – Beam profile images of LED-LCUs at specified distances from the target screen, with normalized irradiance specified in each case, using (i) 460 nm and (ii) 405 nm filters. The active areas of the LCU exit windows are outlined in pink on the images. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

In this work only one example of each type of light source was tested. Previous authors have reported significant differences in the performance of supposedly identical units [13,14].

The results demonstrate spectrometer-based methods are capable of overcoming the limitations inherent thermopile based measurement techniques. with Namely:

348

d e n t a l m a t e r i a l s 3 1 ( 2 0 1 5 ) 339–350

Fig. 4 – Beam Profile images for blue light (460 nm filter) at 0 and 10 mm from the target screen, with normalized irradiance specified in each case. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

• When measured according to ISO 10650-2, the meter design – that dictated the sensor surface to light source distance with the LWP filters – affected the power recorded and the ranking of individual LED-LCUs. • The relatively slow response of thermopile recorders means that they cannot compete with spectrometer based power measuring systems for recording power over short radiation times. • The ISO 10650-2 methodology records data only at a single moment and is unable to assess radiant exposure (irradiance × time) accurately. Radiant exposure may be determined by spectrometer based instruments. • Whilst ISO 10650-2 measures power in four spectral bandwidths it offers little useful information about the spectral characteristics of contemporary LED-LCUs. Thermopiles provide a power value capable of being cross referenced with measurements from a spectrometer to verify calibration but because dental materials have varying spectral requirements, it is important to provide spectral data when assessing a dental LCU. Spectral output is a critical variable in regard to light unit efficacy. The violet light output of polywave LED-LCUs is important for some alternative photo-initiators. Fig. 1 shows the differing spectral outputs for all the eight LED-LCUs measured in this study in standard output mode. The difference in spectral outputs between the single-peak or monowave and multi-peak or polywave units is evident. It is important to be

able to record light output in the 385–515 nm range accurately for polywave LED-LCUs. This information is necessary to prevent improper selection of LCUs for dental materials requiring spectral emission capable of activating ‘short-wavelength’ photo-initiators. The ability of any dental LCU to polymerize resin-based material is thus dependent upon the quality as well as the quantity (radiant exposure = irradiance × radiation time) of light. Light quality relates to the concordance between the photo-initiator absorption spectrum of the restorative material and the spectral output of the light. The ISO 10650-1 and -2 standard assumes that the irradiance and the spatial spectral emission profile of a LCU are homogeneous and can be characterized adequately by an overall irradiance value. Instruments which allow the spectral as well as the power output and irradiance uniformity of light sources to be measured (spectroradiometers used in conjunction with beam profile imaging cameras) allow the quality (spectral matching) and uniformity of the light beam to be determined. Beam profiling allows irradiance “hot spots” and “cold spots” to be identified as well as spectral differences across the face of the light source exit window. Michaud et al. [15] assessed the spectral radiant output of a PAC unit, a single peak blue LED (Elipar S10) and two multi peak or polywave LED-LCUs with a beam profiling camera used in conjunction with an Integrating Sphere. Irradiance ‘hot spots’ were reported to correspond with the locations of the blue LED chips for the polywave Smartlite maX and Bluephase Style units. For every LCU there was considerable variability in the irradiance across the light source tip or exit window. Irradiance was more uniformly distributed for the PAC and single peak LED-LCUs. The Bluephase Style delivered inhomogeneous irradiance compared to the monowave Elipar S10 LED-LCU. Localized variations in specimen polymerization may be directly related to variations in light source irradiance and spectral output [16–21]. Beam profiling, however, is limited by two factors: firstly, the lack of spectral flatness in the CCD camera array. Thus the longer wavelength blue output is over-represented relative to the violet output if single images are taken showing all wavelengths together. Secondly, discrete band-pass filters have to be used sequentially to determine the different wavelength outputs. In this work, the narrow (FWHM = 10 nm) band-pass filters, used to image the violet and blue portions of the LEDLCUs, may only reveal part of the picture. Whilst the output around 400–410 nm of all the polywave LED-LCUs (Table 1 and Fig. 1) is seen in the beam profile images (Figs. 3a–d and 4a), for the Valo unit (Fig. 3d) the output was missing from the 440 nm chip (which is located diagonally opposite to the 405 nm chip in the rectangular 4 chip array). Only the Valo outputs from two 460 nm and the single 405 nm chip were imaged. In contrast to broadband QTH and plasma arc (PAC) sources polywave LED-LCUs have previously been identified as having non-uniform irradiance distribution across their light exit windows at the important violet (405 nm) and blue (460 nm) emission wavelengths [13]. Local differences in irradiance distribution and spectral inhomogeneity have been identified as affecting the extent or quality of cure across RBC specimen surfaces and in depth [16–22]. The significance of these findings has yet to be widely appreciated but the general impact of

d e n t a l m a t e r i a l s 3 1 ( 2 0 1 5 ) 339–350

wide differences in light source performance and clinical factors on restoration longevity are considered to be significant [23]. The characteristics of LED and plasma arc (PAC) lights in comparison to QTH units have been reported using irradiance and depth of cure determinations according to ISO standards [24]. The authors concluded that both LED and PAC units required longer radiation times than those recommended by manufacturers and reported potential thermal and UV-A hazards from the PAC units. ISO 10650 has been used to compare the irradiance of seven LED units with a QTH control unit [25]. Irradiance of these LCUs ranged from 119 to 447 mW/cm2 whereas corresponding irradiances using two commercial dental radiometers were nearly twice the ISO values and the unit rankings differed with the three methods. As already mentioned, ISO 10650-2 was published before multi-peak wavelength LED-LCUs were introduced and as irradiance is estimated only for the 400–515 nm “blue” light region, and energy in the 385–400 nm region is not assessed, then the current standard is not suited for polywave LED-LCUs which have violet as well as blue light peaks. Spectrometer based assessments allow irradiance to be measured against time and thereby allow radiant exposure (irradiance × time) to be calculated irrespective of any variations in light output over time. Techniques that record a single snapshot of power or irradiance at a given moment lack this capability. Work has also been reported toward developing a test to predict the setting time of any light-activated restorative, irrespective of material formulation and light source variables [26,27]. Irradiance reporting should accord with specimen sample diameter to enhance the validity of any conclusions in regard to radiant exposure received. Thus when determining ISO-specification properties of 4 mm diameter specimens, the appropriate incident irradiance reported should be the average for the central 4 mm diameter area of the light tip. This is readily accomplished either with the MARC-RCTM resin calibrator unit or with the checkMARCTM unit using a thin opaque mask with a 4 mm diameter opening.

4.1.

Conclusion

Spectrometer-based methods are capable of overcoming the limitations inherent with thermopile based measurement techniques. ISO 10650-2 only registers power as influenced by radiometer design and band-pass filter constraints. Spectrometer-based measurements can fulfill the intention of ISO 10650. The checkMARCTM device is one type of calibrated portable radiometer. This, and any comparable design, can provide ease of use and validity of power measurement plus spectral data formerly only available with a laboratory based Integrating Sphere setup. Practical devices for routine measurement of dental curing light sources in the dental clinic and the research laboratory enable the clinician and scientist to know the key characteristics of light irradiation applied both intraorally and to standard research specimens. Additional information, as described here, may strengthen future ISO reporting requirements, better informing scientists and clinicians of the actual extra-oral and intra-oral performance of their LCU.

349

references

[1] Stansbury J. Dimethacrylate network formation and polymer property evolution as determined by the selection of monomers and curing conditions. Dent Mater 2012;28:13–22. [2] Leprince JG, Palin WM, Hadis MA, Devaux J, Leloup G. progress in dimethacrylate-based dental composite technology and curing efficiency. Dent Mater 2013;29:139–56. [3] Shortall A, Harrington E. Guidelines for the selection, use and maintenance of visible light activation units. Brit Dent J 1996;180:383–7. [4] Jandt KD, Mills RW. A brief history of LED photopolymerization. Dent Mater 2013;29:605–17. [5] Rueggeberg FA. State-of-the-art. Dental photocuring – a review. Dent Mater 2011;27:39–52. [6] Price RB, Labrie D, Kazmi S, Fahey J, Felix CM. Inter and intra-brand accuracy of four dental radiometers. Clin Oral Investig 2012;16:707–17. [7] Marovic´ D, Matic´ S, Kelic´ K, Klaric´ E, Racic´ M, Tarle Z. Time dependent accuracy of dental radiometers. Acta Clin Croat 2013;52:173–80. [8] Kakeyama A, Haruyama A, Asami M, Takahashi T. Effect of emitted wavelength and light guide type on irradiance discrepancies in hand-held dental radiometers. Sci World J 2013, article ID: 647941. [9] Shortall AC, Harrington E, Wilson HJ. Light curing unit effectiveness assessed by dental radiometers. J Dent 1995;23:227–32. [10] Kirkpatrick SJ. A primer on radiometry. Dent Mater 2005;21:21–6. [11] ISO 4049. 1988 (2) Dentistry – resin based filling materials. Geneva, Switzerland: International Organization for Standardization; 1988. [12] ISO 10650-2. Dentistry – powered polymerization activators – Part 2: light-emitting diode (LED) lamps. Geneva, Switzerland: International Organization for Standardization; 2007. p. 1–7. [13] Price RBT, Labrie D, Rueggeberg FA, Felix CM. Irradiance differences in the violet (405 nm) and blue (460 nm) spectral ranges among dental light-curing units. J Esthet Restor Dent 2010;22:363–77. [14] Lambert RL, Passon JC. Inconsistent depth of cure produced by identical visible light generators. Gen Dent 1988;36:124–5. [15] Michaud PL, Price RBT, Labrie D, Rueggeberg FA, Sullivan S. Localized irradiance distribution found in dental light curing units. J Dent 2014;42:129–39. [16] Balland D, Guillard R, Andre JC. Industrial photochemistry – Vi. Light distribution in photopolymerizable composite materials and heterogeneities observed in light sources used for the photoreaction. Polymer Photochem 1984;4:111–33. [17] Arikawa H, Kanie T, Fujii K, Takahashi H, Ban S. Effect of inhomogeneity of light curing units on the surface hardness of composite resins. Dent Mater J 2008;27:21–8. [18] Vandewalle K, Roberts HW, Rueggeberg FA. Power distribution across the face of different light guides and its effect on composite surface microhardness. J Esthet Restor Dent 2008;20:108–17. [19] Palin WM, Senyilmaz DP, Marquis PM, Shortall AC. Cure width potential for MOD resin composite molar restorations. Dent Mater 2008;24:1083–94. [20] Price RB, Felix CA. Effect of delivering light in specific narrow bandwidths from 394 to 515 nm on the micro-hardness of resin composites. Dent Mater 2009;25:899–908. [21] Price RBT, Labrie D, Rueggeberg FA, Sullivan B, Kostylev I, Fahey J. Correlation between the beam profile from a curing light and the microhardness of four resins. Dent Mater 2014;30:1345–57.

350

d e n t a l m a t e r i a l s 3 1 ( 2 0 1 5 ) 339–350

[22] Haenel T, Hausnerová B, Steinhaus J, Price RBT, Sullivan B, Moeginger B. Effect of the irradiance distribution from light curing units on the local micro-hardness of the surface of dental resins. Dent Mater 2015;31:93–104. [23] Bayne SC. Correlation of clinical performance with ‘in vitro tests’ of restorative dental materials that use polymer matrices. Dent Mater 2012;28:52–71. [24] Nomoto R, McCabe JF, Hirano S. Comparison of halogen, plasma and LED curing units. Oper Dent 2004;29: 287–94.

[25] Aravamudhan K, Floyd CJE, Rakowski D, Flaim G, Dickens SH, Eichmiller FC. Light-emitting diode curing light irradiance and polymerization of resin-based composite. J Am Dent Assoc 2006;137:213–23. [26] Harrington E, Wilson HJ, Shortall AC. Light-activated restorative materials. A method for determining effective radiation times. J Oral Rehab 1996;23:210–8. [27] Shortall AC, Palin WM, Burtscher P. Refractive index mismatch and monomer reactivity influence cure depth. J Dent Res 2008;87:84–8.