Surface roughness values for resin based materials Surface roughness values fo r resin based materials SADJ August 2004 Vol 59 No 7 pp 274 - 279 Corresponding author: ES Grossman MRC/Wits Dental Research Institute School of Oral Health Sciences Faculty of Health Sciences University of the Witwatersrand Private Bag 3 WITS 2050 Tel: +27 11 7172137 Fax: +27 11 7172121 E mail: grossmane@dentistry.wits.ac.za ES Grossman PhD. Senior Specialist Scientist M Rosen LDS MSc(Dent). Honorary Researcher PE Cleaton-Jones BDS MB BCh PhD DSc(Dent) A Volchansky BDS, PhD Honorary Researcher KEYWORDS: surface roughness, scanning elec­ tron microscopy, profilometer, resin based restora­ tive materials, SUMMARY Introduction Surface roughness of dental restorative materi­ als is most often established with the Ra value obtained using profilometry or by assessing surface topography with the scanning electron microscope (SEM). Both methods should vali­ date each other in confirming surface rough­ ness. Aims and objectives The purpose of this study was to compare sur­ face roughness values obtained with a pro­ filometer to the SEM appearance of 6 resin- based restorative materials and assess whether Ra was appropriate as a sole surface roughness measure. Methods Six 5mm diameter specimen discs of Prodigy® (Pr); Z100® (Z); Compoglass F® (C); Hytac Aplitip® (H); Photac-Fil® (Pf) and Vitremer® (V) were prepared against Mylar strips and stored in distilled water for 14 days. One side of each disc was sequentially polished with Soflex discs to super fine state, the other side remained unpolished. Three surface rough­ ness measurements were made on each sur­ face (n=18) recording Ra, Rv, Rp and Rt val­ ues, this data was subjected to a four way ANOVA and Tukey’s Studentised Range Test (p=0.05). Two unpolished and two polished discs per material were prepared for SEM, evaluated and visually grouped for surface roughness. Results Approximate ascending order of roughness was Z, Pr, H, C, V, Pf for Ra, Rv, Rp and Rt and un/polished treatment. Polishing increases sur­ face roughness through raised Rp, Rv or both. SEM evaluation grouped the unpolished spec­ imens into a “bland” (Pr, FI, Z, C) and ‘textured” group (Pf and V). The polished specimens gave four groups: (Pr), (Z and C), (FI) and (V and Pf) of increasing surface complexity. Polishing caused surface scratching, removed the matrix, reduced or removed filler particles and exposed voids within the material. Conclusions This study emphasises the importance of using more than one technique to assess surface roughness. Rv and Rp values should be utilised to better understand polish induced sur­ face feature changes. Rv maximum is a better measure to identify surface defects which could affect restoration longevity. INTRODUCTION Polishing o f resin based restorative materials is necessary to remove excess material and establish aesthetic contour, form and anatomy. During this procedure a smooth restoration sur­ face is obtained that will reflect light in a similar manner to adjacent tooth structure, feel natural to the tongue and produce a surface that will minimise food debris and plaque accumulation. If air bubbles are entrapped within the material during placement, polishing will expose these, creating hollows, holes and cavities at the pol­ ished surface. While an early, in vitro study indi­ cated that bacteria do not accumulate in these voids1, the in vivo situation with modern resin based restorative materials is unclear. Porosities have also been shown to adversely affect the compressive strength of glass ionomer cements2 while others3 have demon­ strated that low velocity cracks are propagated through pores within dental resin composites. Both cracks and holes have major implications for restoration longevity. For close to 30 years a profilometer has been used to measure surface roughness of restora­ tive materials4. The instrument gives four meas­ urements of the topography of a specimen along its reading track, based on a mean line drawn between the peaks and valleys of the roughness profile. Rv is the maximum depth of the profile below the mean line within the sampling range ie. the valley, while Rp is the maximum height of the profile above the mean line within the sam­ pling range ie. the peak. These two values are used to calculate Rt or Rmax which is the maxi­ mum peak to valley distance of the profile in the assessment length (Rt = Rv + Rp). Ra or aver­ age roughness reading is the arithmetic mean of the absolute departures of the roughness profile from the mean line. Ra is often reported as the only measure o f roughness and has been used to express the surface roughness or finish in most studies4"10. In the few cases where the Rt of a surface has been reported this value has been used to monitor major surface flaws and has not formed part of the overall surface roughness assessment4 511 12. Rv and Rp values are sel­ dom reported. Relying on a single method to assess surface roughness may lead to misleading results and conclusions. For this reason it is recommend­ ed13 that a second surface roughness evalua­ tion be used to validate the results of the first. Scanning electron microscopy (SEM) has often been selected because o f its resolution, ease of use and depth of field4 6101214’16. A previously reported study17 compared the Ra surface roughness values of six aesthetic resin- based restorative materials obtained using a profilometer. This investigation examines the SEM appearance of these six restorative mate­ rials17 and utilises Rv, Rp and Rt to assess whether the commonly used Ra value is suffi­ ciently discriminating to identify surface features which could have major implications for restora­ tion longevity in modem restorative materials. MATERIALS AND METHODS Specimens were prepared from two brands of hybrid composites: Prodigy® (Kerr, Orange USA) and Z100® (3M, St Paul, USA); two brands of compomers: Compoglass F® (Vivadent Schaan, Lichtenstein) and Hytac Aplitip® (ESPE, Norristown, USA) and two brands of resin modified glass ionomers: Photac- Fil® (ESPE, Norristown, USA) and Vitremer® (3M, St Paul, USA) according to manufacturer’s recommendations. For each brand six discs 5mm in diameter and 1.5mm thick were pre­ pared in clear perspex moulds. Following place­ ment of the material both upper and lower sur­ faces were covered with Mylar strips (Buffalo Dental Mfg. Co., Brooklyn USA) which in turn were covered with glass microscope slides and compressed with finger pressure to express sur­ plus material. Each disc was light cured for 40 seconds per side with a Dentsply CHL 75 curing light (Dentsply Caulk, Milford, USA) to give a total of 12 surfaces for each brand. After storage in distilled water for 14 days one surface of each specimen disc was polished. One individual car­ ried out all polishing after standardising the tech­ nique on practise specimens. Medium, fine and super fine Soflex discs (3M, St Paul, USA) were used sequentially for 30 seconds each in a slow hand piece with light intermittent pressure. New discs were used for each specimen, which was polished in a north-south direction for 15 seconds and an east-west for a further 15 seconds. This method was used to simulate clinical polishing in www. sadanet.co.za August 2004 Vol. 59 No. 7 SADJ 274 mailto:grossmane@dentistry.wits.ac.za sc i en t i f i c an incisogingival and mesiodistal direction. The other surface of the specimen disc was untouched and served as a control. The speci­ mens were stored for a further seven days in dis­ tilled water until surface profiles were recorded with a Rank Taylor Hobson Ltd., Form Talysurf Series 2 instrument (Taylor Hobson, Leicester, UK) with a 2mm tip, a standard cut-off of 0.8mm and a traverse length of 3.2mm. The specimens were dabbed dry, but not desiccated and placed randomly in the Talysurf. Once seated, three replicate measurements were made across the surface of each specimen disc; all traverses were in the same direction and parallel to each other. The stylus tracks were approximately equidistant on all specimens. The three replicate measure­ ments gave a total of 18 tracks per treatment sur­ face. The data from these 18 tracks yielded the Ra, Rv, Rp and Rt values recorded for each treat­ ment. The data were subjected to a four way analysis of variance ANOVA and Tukey’s Studentised Range test at p=0.05 using SAS (SAS for Windows Version 8.2, SAS Institute Inc., Cary, NC: USA). Specimens were returned to distilled water immediately after surface profile assessment for a further week prior to SEM preparation. A total of 24 discs were removed from the per­ spex strips, four for each material, and prepared for SEM examination. Each disc was air dried and mounted on aluminium specimen stubs using colloidal graphite (DAG 580 Colloidal Graphite in denatured Alcohol. Acheson Colloids Company. Prince Rock, Plymouth PL4 OSP, UK) so that the polished surface was uppermost in two specimens and the unpolished surface uppermost for the other two specimens for each material. All specimens were coated with a thin layer of carbon (25nm) and viewed in a JSM-840 SEM (JEOL Ltd, 1-2 Musashino 3-chome, Akishimo, Tokyo 196, JAPAN) at 5-1 OkV and suitable tilt and magnifications to emphasize sur­ face profile. Viewing was done by an electron microscopist, blind to the surface roughness pro­ file measurements. Representative surfaces of the material were photographed at x200 and x1000 magnification printed and assessed. Surface roughness was judged on the absence or presence of scratches, grooves, pitting, matrix loss, filler particle loss, surface fragmentation, cracks and incorporated voids. Filler particle sizes for all six materials were sought from the manufacturer. This was not provided in all cases therefore exposed filler particles were measured from electron micrographs (Table I). RESULTS Profilometer measurements Tables II and III show the mean values and stan­ dard deviations of the four surface profile read­ Table I. Particle sizes of resin-based restorative materials used in this study Material Manufacturers Manufacturers Measured average average particle size particle size particle size Glass ionomers Photac-Fil Not available 50% = 5.5-7.0;im 100% = <42.0um 7.5 - 50.0^m Vitremer Not available Not available 5.0 - 50.0rrm Compomers Compoqlass F Not available 0.2 - 3.0 Hytac Aplitip 5.0^m 50% = <5.2^m 90% = <16.9um 7.5 - 25.0^/m Hybrid composite Prodiav 0.5iim Not available 2.5i/m > Z100 0.6^m 0.01 - 3.5^m 2.0^m > Table II. Means and standard deviations in pm of Ra, Rv, Rp and Rt for unpolished sur­ faces of the six resin-based restorative materials studied. Means with the same letter in each column indicate that the materials are not significantly different. (Ra values17) Material Ra Rv Rp Rt Photac-Fil 0.71 +0 .83 ab 3.34 + 2.17 b 3.06 + 4.24 a 6.41 + 7.27 ab Vitremer 0.42 ± 0.25 b 6.62 + 3.93 a 1.29 ±0 .85 ab 7.92 ±4 .42 a Compoqlass F 1.15 ± 1.44 a 2.17 ±2 .46 be 2.03 ±3 .02 ab 3.91 ±4 .65 be Hvtac Aplitip 0.15 + 0.19 b 0.45 + 0.40 c 0.95 + 1.67 ab 1.41 + 1.87 c Prodiqy 0.14 + 0.10 b 0.31 + 0.18 c 0.46 ± 0.47 b 0.77 + 0.52 c Z100 0.22 ± 0.36 b 0.30 + 0.32 c 0.35 ± 0.46 b 0.51 ± 0.36 c Table III. Means and standard deviations in pm of Ra, Rv, Rp and Rt for polished sur­ faces of the six resin-based restorative materials studied. Means with the same letter in each column indicate that the materials are not significantly different. (Ra values17) Material Ra Rv Rp Rt Photac-Fil 1.51 ± 1.44 a 12.18 ± 6 .84 a 3.99 ±3 .37 a 16 18 ± 9 .55 a Vitremer 0.65 ± 0.48 ab 5.77 ±2 .88 b 1.67 ± 1.75 b 7.54 + 4.14 b Compoglass F 1.33 ± 1 .93 ab 2.21 ± 2.25 c 1.79 ±2 .34 b 3.98 ±3.81 be Hytac Aplitip 0.60 ± 1.10 ab 2.30 ± 2.49 c 1.18 ±0 .78 b 3.48 ±2 .82 be Prodigy 0.47 ± 0.33 ab 2.16 ± 2.38 c 1.14 ± 1.54 b 3.30 ± 3.84 be Z100 0.35 ± 0.37 b 0.88 ± 0.75 c 0.71 ±0 .72 b 1.60 ± 1.23 c ings for each of the six materials examined in the unpolished and polished condition. While the means for the different surface roughness parameters vary, the relative order of roughness for each material remains largely similar within each test. The maximum and minimum range, mean and median in polished and unpolished surface roughness values are shown for Ra (Fig. 1) arranged from least to greatest maximum value. In this case the relative order of the mate­ rials varies. This was also true for Rp, Rv and Rt - these values are not shown because of space constraints. Polishing tended to increase all max­ imum R values, but in two cases (Photac-Fil and Hytac Aplitip) the maximum Rp values decreased markedly. Overall, polishing caused an increase in Rt values which could be due to increases in the peaks, valleys or both as shown in Fig. 2. Compoglass F, Prodigy and Z100 expe­ rienced increases in both Rv and Rp maximums when polished; Rv maximum increased and Rp maximum decreased in Photac-Fil and Hytac Aplitip; Vitremer experienced an increase in Rp maximum and a decrease in Rv maximum. A similar pattern was apparent if mean, as opposed to maximum Rv and Rp values were used. Figure 1 shows that means and medians do not coincide indicating the skewness in the data. This skewness is best illustrated when the means and medians are plotted as cumulative percentages of specimens (Fig. 3). In this figure the polished and unpolished surface roughness values for Ra, Rp, Rv and Rt are reflected as cumulative percentages for each of the restorative materials. The median is given as a straight line at 50% of the specimen group. The position of the means indicates the cumulative aggregate of specimens incorporated within that rough­ ness value as a percentage of the entire spec­ imen group. Percentage departure from the median ranges from a minimum of 0% where mean and median coincide (as in the case of Rv for unpolished Vitremer) to 39% (Ra for polished Hytac Aplitip). The magnitude of the www. sadanet.co.za August 2004 Vol. 59 No. 7 SADJ 276 sc i en t i f i c percentage departure between median and mean values indicate the presence of significant outliers within the data set which influence the mean value. For example the Ra maximum value of polished Hytac Aplitip is 5.05mm with the next greatest being 0.74mm. The Ra medi­ an value is 0.34mm and Ra mean = 0.60mm. While the plotted graph­ ic distance between mean and median for polished Flytac Aplitip appears minor in terms of the maximum and minimum values (Fig. 1), the notable maximum value outlier has effectively doubled the mean value relative to the median. In addition the mean value falls at the 90% cumulative level of the specimen group. Scanning electron microscopy All unpolished specimen surfaces had a smooth, matrix rich surface layer indicating that the flowable resin component was forced up against the Mylar strip during specimen placement. Shallow scratches or grooves were visible on all specimen surfaces (Fig. 4). Placement defects or voids were present in Photac-Fil and Vitremer from 150mm> in diameter (the large voids are not illustrated due to space constraints), as well as surface cracks (Fig. 5). SEM suggested that unpolished specimens could be divided into two based on the additional presence of surface cracks and voids: the “bland” (Pr, H, Z, C) and “textured” (Pf and V) groups. Polishing caused scratching of the surface and resulted in four surface appearances: Prodigy was the smoothest with scratches of varying depth and small pits irregularly scattered on the surface (Fig. 6). Compoglass F and Z100 (Fig. 7) had similar clusters of pits but were more heavily scratched, with some small voids present in Compoglass F. Flytac Aplitip (Fig. 8) was heavily scratched and had numerous pits or areas of gouged out material. Polishing of Vitremer and Photac-Fil showed a distinct removal of the matrix rich surface layer exposing the filler particles and particle-matrix interface; in addition scratches, cracks and exposed voids were present (Fig. 9). Most voids were present in polished specimens. They were mainly 50mm> in diameter, although larger bubbles up to a maximum of 450mm were pres­ ent. Polished specimens gave four groups: (Pr), (Z and C), (FI) and (V and Pf) based on the complexity and variability of surface topography. The order of surface roughness as determined by R values and the group­ ings of similar surfaces by SEM showed an arrangement closest to the Rv maximum value, more so for the unpolished (from greatest to least V, FT, C, Z, H, Pr) than polished (from greatest to least FT, H, Pr, V, C, Z) treatment. The materials with the largest particle sizes had the most complex surface topography at SEM level. DISCUSSION SEM showed that surface defects were almost exclusively confined to flaws extending below the surface rather than those which protruded above the restorative material surface. While surface scratches and areas of gouged out material were evident in mainly polished specimens (and could be the result of poor polishing technique), voids of various sizes formed the most visible part of surface defects present in both specimen treatments. Porosities in resin composite restorative materials have been noted for many years9 1012 and are ascribed to air bubbles incorporated at various stages within the material18. Such voids are implicated in decreased compressive strength2; crack propagation319; surface roughness20 and microleakage1. While the increased surface area represented by voids and scratches may augment plaque accumulation it is well known that the relationship between surface roughness and enhanced retention is not necessarily parallel2122. Opinion differs as to the acceptable clinical level of surface roughness. Kaplan et al.15 have suggested that a Ra mean value of less than 10mm is clinically undetectable and therefore clinically acceptable. On the other hand, Borchers et al.8 supports a target Ra threshold of 0.2mm Surface roughness in pm ■ Ra m«sdian ' Ra mean ( ̂ > - H *— I h>— I __ _ < Z100 Hytac Ap Cmpoglss F Prodigy Vltrmr Phtac FI Prodigy Vltrmr Phtac FI Hytac Ap Z100 Cmpoglss F Polished Unpolished Figure 1: Maximum and m inim um range (line and whisker), means and medi­ ans, are shown fo r Ra. The six materials are arranged from least to greatest maximum value in po lished and unpolished groups. Maximum surface roughness in pm ^ P o l i s h e d I U n p o lis h e d 10 Rp (Peak) o Rv (Valley) -10 - 20f / -30 V itrem e r H y ta c A p lit ip Z100 P h o ta c -F il C o m p o g la s s F P ro d ig y Restorative materials Figure 2: H istogram showing the effect o f po lish ing on Rt maximum values for the s ix restorative materials. A ll values above zero indicate Rp maximum val­ ues, those below zero show Rv maximum values. Cumulative percent Polished Unpolished Median ■ Rt mean x Ra mean ♦ Rv mean A Rp mean Figure 3: Histogram o f cumulative percentages o f specimen spread in polished and unpolished conditions. The mean and median is given for Ra, Rv, Rp and Rt www. sadanet.co.za August 2004 Vol. 59 No. 7 SADJ 278 sc i en t i f i c Figure 4: Shallow scratches (arrowed) are visible on the unpolished, m atrix-rich Prodigy surface. Figure 5: The matrix rich surface layer o f this unpolished Vitremer specimen is broken by numer­ ous voids which have been incorporated w ithin the material. Some cracks can be seen extending from the voids. Figure 6: Deep grooves, shallow scratches and areas o f p itting are visible on this po lished Prodigy surface. grooves and pits. for effective plaque prevention on restorative surfaces. The Ra = 10mm threshold for acceptability means that all six resin based restorative materials whether polished or unpolished, have clinically satisfactory sur­ faces as regards plaque prevention. However Figure 8: Irregu lar surface o f a po lished Hytac A p litip specim en show ing num erous p its and scratches. Figure 9: Polishing o f Vitremer specimens removed the matrix rich surface layer exposing fille r pa rti­ cles and voids, both o f which in this fie ld can be up to 50mm at the w idest point. Grooves and scratch­ es can be seen with large crack on the right. An area o f roughened matrix is w ithin the angle o f the crack. Compare this surface to that in Fig. 5. the SEM assessment of surface roughness clearly indicates differences between the materials within each treatment. The Ra threshold of 0.2mm was more in keeping with SEM discrimination of surfaces seen in this study. All polished surfaces had a complex topography at the ultramicroscope level as did the unpolished Photac-Fil and Vitremer: all had Ra’s well over 0.2mm and would be regarded as having unacceptable plaque pre­ vention surfaces8. The inconsistency between SEM and Ra sur­ face roughness measures in this study sug­ gests that other surface roughness values, generated during profilometry, may be more appropriate when examining the effect of pol­ ishing. It seems strange that Rv, which ascer­ tains the valley depth of measured surfaces and as such records a pivotal aspect of restorative material topography, appears to be overlooked in surface roughness studies. We therefore suggest that Rv maximum values are a useful measure as pointers for possible areas of threat to restoration longevity. After all, the site of restoration breakdown will be ini­ tiated through the weakest region of the mate­ rial, in most cases the site o f the largest defect. Rp values of the restorative materials studied were not as large as Rv values. It is not known to what extent peaks could affect the physical and mechanical properties of the restoration, after all mastication would tend to grind down surface pro­ trusions with time. However, projections above the restoration would offer less shelter from the swirling oral tide than the valleys and may be insignificant in the accumulation of oral debris. The above does not imply that Rp values are inconsequential as regards restoration longevity, this can only be assessed in further work. The Rt value could represent an exaggeration of the topography in as much as the Ra values represents a flattening of the surface. This has been recognised previously" 12, but both stud­ ies seem to regard the sensitivity of Rt to a sin­ gle large surface defect along the traverse path as a shortcoming of the method rather than adding to information on potential vulnerabili­ ties of the material. The present data shows that polishing may increase or decrease the valleys and peaks depending on the resin based restorative material. While the overall effect may be to increase Rt, the consequence this has on the material may differ whether Rp or Rv is more affected by the process: and thereby the interpretation of the data. Surface roughness of restorative materials is a term which is loosely applied to the surface roughening effects of finishing and polishing on the material itself. It seems that over the years other equally important consequences of this procedure have been overlooked such as the exposing of voids and cracks within the bulk of the material. While these features were seen in the SEM the Ra values appear not to be suffi­ ciently sensitive, whether analysed by statistical methods or differentiated by threshold values, to reflect these irregularities. CONCLUSION This study highlights once again the importance of examining polished and finished restorative material surfaces using more than one technique as false conclusions can be drawn from just one set of data. While the use of Ra mean values appears to have adequately quantified dental material surfaces in the past it may be that the nature of present day resin based restorative materials require the use of all R values to fully portray the topography of such surfaces when assessing materials for clinical outcomes. The added use of Rv and Rp maximum values would surely assist in a greater understanding of place­ ment and wear phenomena unique to modem restorative materials. References 1. Skjorland, KK. Henston-Petersen, A. Rstevik, D. Soderholm, K-J. Tooth Coloured Dental Restorative Materials: Porosities and Surface Topography in Relation to Bacterial Adhesion. Acta Odontol Scand 1982; 40: 113-120. / V_______________________________ The rest o f this aricle 's references (2 - 22) w ill be published in the online A ugust SADJ, www.sada.co.za SADJ August 2004 Vol. 59 No. 7 www. sadanet.co.za http://www.sada.co.za