- Email: [email protected]
Comparison of two stylus methods for measuring surface texture
dental materials Dental Materials 15 (1999) 79–86 www.elsevier.com/locate/dental
Comparison of two stylus methods for measuring surface texture S.A. Whitehead*, A.C. Shearer, D.C. Watts, N.H.F. Wilson University of Manchester Dental School, Higher Cambridge Street, Manchester M15 6FH, UK Received 2 November 1998; received in revised form 18 December 1998; accepted 21 December 1998
Abstract Objectives: The purpose of this study was to evaluate and compare two ‘‘stylus’’ methods for measuring surface texture of dental tissues and materials Methods: The two styli chosen were a contact diamond stylus and a non-contact focussed laser stylus attached to the same measuring apparatus. Results: These indicate that there are significant differences between those obtained from surface texture measurements using a noncontact laser stylus and a diamond contact stylus method despite being mounted in the same profilometer. This is valid for both the test specimens of known surface texture, provided by the manufacturers, and for a ‘‘real world’’ simulation using contoured and finished Dicor ceramic blocks. The only significant agreement between the two styli was found for the Ra parameter. This should not be used alone to describe the roughness of a surface because the parameter is not sensitive to profile shape. Owing to the properties of the stylus used it is essential that the limitations of surface profilometry be recognised. Significance: Caution should be exercised when comparing the results of surface texture studies of dental hard tissues and restorative materials using varying types of stylus attached to a surface profilometer. q 1999 Academy of Dental Materials. Published by Elsevier Science Ltd. All rights reserved. Keywords: Contact diamond stylus; Non-contact laser stylus; Surface profilometry
1. Introduction The ideal dental restorative material and associated techniques continue to be the goal of material scientists and clinicians alike. The ‘‘gold standard’’ for restoration quality could be considered to be its long term performance in clinical service [1,2]; however, laboratory tests are useful as indicators of clinical performance and have practical benefits. The widely used laboratory technique is to measure the surface texture of a surface and relate the surface texture to a clinical parameter such as wear resistance [3,4] or to use surface texture measurements to compare materials and techniques (polishing for example). There are several methods currently available to measure the texture of a surface; these include contact stylus tracing, laser reflectivity, non-contact laser stylus metrology, scanning electron microscopy, and compressed air measuring [5]. More recently, atomic force microscopy has been evaluated for use in assessing the surface texture [6,7]. Of these methods the contact stylus tracing method is probably the
most commonly used. There are several shortcomings with the use of this method which relate to the mass of the measuring stylus and the physical configuration of the contacting tip. The authors have already described the failings of using a laser reflection method for describing surface texture [8,9] However a laser type ‘‘stylus’’ was developed which replaces a conventional contact stylus that does not rely on the measurement of scattered laser light. The aim of this study was to compare the performance of a contact stylus and non-contact (laser) stylus. Initially, comparisons were made on samples supplied by the manufacturer. Following this, a ‘‘real world’’ situation was selected to validate the performance of the two systems in a dental materials context. The measurement of surface texture of finished ceramic was chosen as such restorations may be difficult to contour and effectively polish in vivo [10–12] and also ceramic materials may be semi-translucent which may influence the performance of the laser stylus.
2. Materials and methods * Corresponding author. Tel.: 1 44-161-275-6660; fax: 1 44-161-2756710. E-mail address: [email protected] (S.A. Whitehead)
Two surface measuring styli were evaluated: a diamond contact stylus (Fig. 1) and a laser stylus (Fig. 2). To ensure
0109-5641/99/$20.00 1 0.00 q 1999 Academy of Dental Materials. Published by Elsevier Science Ltd. All rights reserved. PII: S0109-564 1(99)00017-2
80
S.A. Whitehead et al. / Dental Materials 15 (1999) 79–86
ware translated the data into the selected parameters. Sixty four measurement runs were taken from the centre of the calibration specimen over an area of 5.6 × 5.6 mm 2 (Fig. 3). The calibration specimen was replaced with the optically flat glass plate onto which a groove of 8.8 mm had been prepared. The groove was then profiled using the diamond stylus to assess the effectiveness of the profiling system at reproducing the groove. 2.2. Measurement by non-contact laser stylus tracing
Fig. 1. Photograph of the diamond stylus tracing the surface of a finished block of ceramic. The non-contact laser stylus was attached to the same apparatus to duplicate the measuring setup.
that the results could be compared, the styli were mounted in the same measuring system, a Perthometer S8P (FeinprufPerthen GmbH, Gottingen, Germany). Prior to the evaluation of the ceramic surfaces, the performance of the two styli were evaluated on two test surfaces which were provided by the manufacturer. These surfaces were a groove of 8.8 mm prepared in an optically flat glass plate and a machined metal surface of known surface texture (Ra 1.98 mm, Rpm 9.2 mm, Rz 8.65 mm) which will be referred to as the calibration specimen. 2.1. Measurement of the manufacturers samples by contact stylus tracing The Perthometer S8P is a stylus based surface measuring instrument for the acquisition, graphical presentation, evaluation and documentation of surface profiles. A diamond stylus of 5 mm radius and stylus angle 908 was traversed at a constant speed across the calibration specimen with a force of 0.8 N (Fig. 1). The tracing run was parallel to the long axis of the longitudinal features of the machined calibration specimen. For the purposes of this study a tracing length of 5.6 mm was chosen. Data from an inductive transducer coupled to the stylus were relayed to the central processing unit of the apparatus and on-board firm-
Fig. 2 shows a cut-away diagram of the Focodyn laser stylus. A laser diode built into the casing serves as a coherent light source (the manufacturers only say that the laser source emits at the red end of the visible spectrum). The laser light initially passes through a collimator and is then guided by a prism to the micro objective lens. Another prism in the measuring head guides the beam onto the surface with a spot size of 1 mm. The micro objective lens focuses the beam exactly 0.9 mm below the aperture of the measuring arm. The light reflected back from the surface returns the same way through the optical system where it is directed by a beam splitter to the detector. This detector is part of a sensitive control loop which by means of a powerful linear motor readjusts the measuring arm continuously so that a distance of 0.9 mm to the surface remains constant (dynamic focussing). Thus when the Focodyn stylus is passed across a surface, the measuring arm follows the surface profile and provides a trace of the surface. For the purposes of this study the diamond contact stylus was replaced by the Focodyn laser stylus with all other measuring parameters remaining unchanged and surface texture measurements of the calibration specimen and 8.8 mm groove repeated. 2.3. Ceramic sample preparation To compare and contrast the performance of the two styli systems in a ‘‘real world’’ situation, a series of different surface textures were produced on blocks of machinable ceramic. Eighteen dark shade Dicor MGC (Corning, NY, USA) blocks were used. To simulate contouring of the occlusal and approximal surfaces, a diamond bur
Fig. 2. Cut-away diagram of the Focodyn laser stylus to show the mode of operation and the path of the focussed laser beam.
S.A. Whitehead et al. / Dental Materials 15 (1999) 79–86
81
Fig. 3. Pseudo 3D image from 64 consecutive profiles of the calibration specimen. The profiles occupy an area of 5.6 × 5.6 mm 2 in the centre of the specimen. This was the area profiled by the contact and non-contact (laser) styli. The manufacturer’s data regarding this specimen is: distance between peaks (l ) 0.8 mm, Ra 1.98 mm, Rpm 9.2 mm, Rz 8.65 mm.
(100 mm) was applied to all the blocks (Premier Dental Products, Norristown, PA, USA). The burs were held in a friction grip, high speed handpiece (Mid-west Quiet-air, Sybron Corp., Des Plaines, IL, USA) and used at moderate speed and pressure (0.1 MPa) with coolant. Each bur was Table 1 Outline of the finishing sequences and protocols for the preparation of the ceramic samples used to compare the two measuring systems Instruments used/Method Method A 30 blade tungsten carbide a Shofu porcelain points b Truluster diamond paste c Method B Two striper MFS microfine diamonds 80, 40, 15 mm d Shofu porcelain points Truluster diamond paste Method C 30 blade tungsten carbide Shofu porcelain points Two striper diamond polishing system Method D Baker Curson bur e Truluster diamond paste Method E Two striper MFS microfine diamonds 80, 40, 15 mm 3M Soflex disc f (coarse, medium, fine) Method F 30 blade tungsten carbide Shofu porcelain points Truluster diamond paste a
Protocol
High speed dry Slow speed wet
2.4. Surface roughness parameters selected High speed wet
Slow speed wet High speed dry Slow speed wet
High speed wet
High speed wet
Moderate speed dry
High speed dry Slow speed wet
Kerr/Sybron, Romulus, MI, USA. Shofu, Menlo Park, CA, USA. c Braesseler, Savannah, GA, USA. d Premier Dental Co., Norristown, PA, USA. e Ash, London, UK. f 3M Co., St Paul, MN, USA. b
applied to the surface for 30 s in one direction using light intermittent pressure. Each of the six finishing sequences detailed in Table 1 was applied to three blocks. All burs were applied to the surface in one direction only using light intermittent pressure. Diamond and tungsten carbide burs were used until the ceramic surface appeared uniform and no further improvement apparent. Diamond pastes were applied in a bristle brush for 20 s at moderate speed and pressure. The surface of the Dicor blocks were profiled with the contact and laser styli using an identical protocol to that used for the calibration specimen.
To describe the surface texture of the finished ceramic samples, several parameters were selected. This is because some commonly used surface parameters are true amplitude parameters (Ra for example) and give no information as to the shape of the recorded profile. The selected surface parameters were Ra, Rz, Rpm, Rpm:Rz ratio and the material ratio [Mr] [5,13] and are described in the following: 2.4.1. Ra parameter The Ra parameter describes the overall roughness of a surface and can be defined as the arithmetical average value of all absolute distances of the roughness profile from the centre line within the measuring length and is R derived from the equation:Ra 1=l 10 uy xudx: 2.4.2. Rz parameter The Rz surface parameter can be defined as the average maximum peak to valley height of five consecutive sampling lengths within the measuring length and is used to describe the degree of roughness of the surface of the sample as given by: Rz (1/n)(Z1 1 Z2 1 … 1 Zn). 2.4.3. Rpm parameter The Rpm surface parameter is the mean value of the
82
S.A. Whitehead et al. / Dental Materials 15 (1999) 79–86
levelling depths of five consecutive sampling lengths as given by Rpm Rp=1 1 Rp=2 1 Rp=3 1 Rp=4 1 Rp=5 =5: Exceptional profile peaks are thus only partly considered. By contrast to the surface parameters already described (Ra, Rz), the Rpm gives reliable information on the profile shape. Small Rpm values characterise a surface featuring wide peaks and narrow valleys whilst greater Rpm values indicate a spiky, sharp ridged profile (Fig. 8). 2.4.4. Rpm:Rz ratio The Rpm:Rz ratio is of special interest because the value also gives valuable information on the profile shape. A ratio higher than 0.5 indicates a sharp ridge profile, a ratio smaller than 0.5 indicates that the profile is rounded. 2.4.5. Material ratio (formerly micro bearing ratio (tp)) The material ratio (Mr) parameter is assessed from the roughness profile (R-profile). Long wave components such as waviness are suppressed by a special filter and the Mr calculated as: Mr 1=lm L1 1 L2 Ln 100%: Graphically this can be represented by the Abbot–Firestone curve derived from the serial section levels (Fig. 4). It is a common practice to select a reference level which just fits into the profile (usually Mr 2%). The application of this reference level provides a mathematical line across the sample from which the remainder of the measurements are taken. For the purposes of this investigation the level at Mr 2% was used to determine that percentage of material present 5 mm down into the surface of the material. 2.5. Data analysis The data from the contact and non-contact laser stylus were collected for each of the samples studied and for each surface parameter selected. The data was collated for analysis using the arcus pro-stat statistical package
(Medical Computing, Aughton, West Lancashire, UK) to assess the comparisons between the two measuring systems using the correlation coefficient r 2. P and t values were calculated to determine if any of the surface roughness parameters derived from the contact stylus method had a significantly better correlation to the data derived from the noncontact method.
3. Results 3.1. Test surfaces The results of the two styli on the test surfaces are summarised in Table 2. There are significant differences between all the measurements for the contact and laser styli (P . 0.05). Fig. 5 shows sample traces from the two measuring systems on the calibration specimen and the 8.8 mm groove which indicate how the differing styli interpret the surfaces. 3.2. Finished ceramic (Dicor) The results of the measurements of the surface roughness of the finished ceramic samples are given in Table 3. The correlation coefficients, t and p values are given in Tables 4 and 5 which show the relationship between the results obtained using the contact and non-contact tracing methods. It is clear that there is only one significant agreement (P , 0.05) between the laser non-contact and diamond contact stylus methods; this is when comparing the surface texture parameter Ra. For all other surface texture parameters measured there was no significant correlation between the two measuring systems. Fig. 6 shows the differences in the tracings obtained using the non-contact laser and diamond contact styi. The fidelity of the surface texture as ‘‘seen’’ by the laser stylus can be seen to be different from that quantified by the diamond stylus.
Fig. 4. Derivation of the Abbot–firestone curve from serial section levels. The material ratio chosen in this study was taken as 5 mm from the surface (the ‘‘surface’’ being that line at which there is 2% of material).
S.A. Whitehead et al. / Dental Materials 15 (1999) 79–86
83
Table 2 Results of the laser stylus and the contact stylus profiling the test surfaces compared to the data supplied by the manufacturer (standard deviation in parentheses). There were significant differences (P . 0.05) between the contact and laser styli for all the surface texture parameters measured. The bottom row of the table shows the data for the styli measuring the 8.8-mm groove Parameter
Manufacturer’s data for calibration specimen
Contact stylus measurement of calibration specimen
Laser stylus measurement of calibration specimen
Ra (mm) Rpm (mm) Rz (mm) Rpm:Rz ratio Mr at 2 5 mm Depth of groove in flat glass plate(mm)
1.98 9.2 8.65 1.06 n/a 8.8
1.7(0.03) 4.0(0.2) 6.6(0.4) 0.6 98(2)% 7.41
1.84(0.05) 5.36(0.18) 11.05(0.49) 0.49 45.8(5.5)% 11.9
4. Discussion The aim of this investigation was to compare contact (diamond stylus) and non-contact (laser stylus) methods of surface texture measurement. The measurements of the standard surfaces provided by the manufacturer suggest that the contact stylus system as delivered and calibrated by the manufacturer, underestimates the roughness and spikiness of the surface. This represents a physical constraint on behalf of a diamond stylus to fully represent deep surface features which are narrower than the stylus itself (Fig. 7). This effect may be overcome by using a narrower and longer stylus but in practice this results in a weak stylus which is liable to be easily damaged. The laser stylus gave more comparable results to the manufacturer’s data than the contact stylus apart when measuring the calibration specimen. However, measurement of the 8.8 mm groove produced an ‘‘overshot’’ at the
bottom of the groove which resulted in an artefact (Fig. 5). The results indicate that there are significant differences between the performance of the two measuring systems when measuring the test surfaces. This was surprising as these specimens are used by the manufacturers and support a personnel to calibrate the instrument on delivery and at subsequent maintenance visits. The difference in the results is therefore considered to be a fundamental attribute of the two measuring systems and not an artefact of the study design. The results from the profilometry of the ceramic samples indicate that there is little correlation between the two systems apart from when measuring the Ra surface texture parameter. This correlation is unsurprising as the Ra parameter is a simple amplitude descriptor and gives no indication of profile shape. Thus, if the surface is round edged or spiky then this parameter is unable to differentiate between the two and is a poor predictor of surface
Fig. 5. Sample tracings from the evaluation of the test specimens. Profiles A and B are from the contact and laser stylus respectively, and show the surface of the calibration specimen. Profiles C and D are of the 8.8 mm groove. Profile D is from the laser stylus and illustrates the overshoot of the trace as the stylus fails to find the floor of the groove and overcompensates giving an artefact.
84
S.A. Whitehead et al. / Dental Materials 15 (1999) 79–86
Table 3 Results of the averaged surface measurements using the non-contact laser and contact diamond styli for each finishing method Finishing regimen
Parameter
Laser stylus
Contact stylus
Method A
Ra (mm) Rz (mm) Rpm (mm) Rpm:Rz Tp ( 2 mm) Ra (mm) Rz (mm) Rpm (mm) Rpm:Rz Tp ( 2 mm) Ra (mm) Rz (mm) Rpm (mm) Rpm:Rz Tp ( 2 mm) Ra (mm) Rz (mm) Rpm (mm) Rpm:Rz Tp ( 2 mm) Ra (mm) Rz (mm) Rpm (mm) Rpm:Rz Tp ( 2 mm) Ra (mm) Rz (mm) Rpm (mm) Rpm:Rz Tp ( 2 mm)
0.4 0.36 1.3 0.38 1.37 0.47 4.17 1.4 0.34 2.8 0.43 4.1 1.3 0.32 2.64 1 8.8 2.3 0.27 4.14 0.59 5.3 1.5 0.27 3.69 0.36 3.0 1.1 0.36 2.47
0.33 1.85 0.97 0.55 2.3 0.4 1.93 1.03 0.55 2.57 0.27 1.54 0.76 0.48 1.54 5.37 2.89 1.3 0.47 3.41 0.26 1.11 0.64 0.57 2.63 0.31 1.4 0.79 0.57 2.35
Method B
Method C
Method D
Method E
Method F
Method A Method B Method C Method D Method E Method F
morphology (Fig. 8). The Rpm:Rz ratios determined by the laser stylus were somewhat higher than those found by the contact stylus methods for similar surfaces. Therefore the laser stylus profiled surfaces are more sharp ridged and spiky. This may be a true finding or an artefact caused by the inability of a diamond stylus tip of finite radius to accurately determine a sharp ridge or that the stylus ‘‘knocked off’’ the ridge during measurement. Conversely the laser non-contact stylus may be producing erroneous measurements because of an overshoot phenomenon as the laser stylus fails to find the tip of the ceramic ridge and produces a false ‘‘spike’’ or that the laser light is reflected from the Table 4 Correlation coefficients between the non-contact laser and contact diamond styli for each of the parameters included in the study. The only significant correlation is the Ra surface parameter (P , 0.05)
Ra (mm) Rz (mm) Rpm (mm) Rpm:Rz Mr (%)
Table 5 Correlation coefficients between the non-contact laser and contact diamond styli for each of the finishing regimens. The only significant correlation is for the finishing regimen ‘‘Method C’’ (P , 0.05)
r
r2
t
p
0.941 0.533 0.753 0.404 0.623
0.886 0.284 0.540 0.163 0.387
5.570 1.259 2.160 0.884 1.591
0.050 0.276 0.096 0.426 0.186
r
r2
t
p
0.477 0.850 0.936 0.014 0.621 0.828
0.288 0.723 0.875 0.019 0.386 0.686
0.942 0.279 4.603 0.245 1.372 2.560
0.416 0.067 0.019 0.822 0.264 0.080
subsurface rather than the actual surface. However, as the laser stylus can effectively track across a glass plate then the reflected laser energy from the semi-transparent Dicor would be sufficient to enable effective operation of the laser stylus. Similar differences are noted in the Mr in that the diamond stylus appears to ‘‘look’’ further into the valleys of the ceramic after finishing. This again may be an artefact caused by a ‘‘drop-off’’ phenomenon as the laser stylus overshoots the bottom of a valley to record a valley depth which is greater than that actually present. The material ratio is a useful parameter as it gives an indication of the porosity of a surface and can differentiate between surfaces of similar Ra and Rz (Fig. 8). Again the material ratios derived from the laser stylus were significantly different from the contact stylus both for the calibration specimen and for the ceramic blocks. This further illustrates the problems of profiling a surface which has narrow pits and pores using a diamond stylus which tends to smooth out such surfaces. It would seem therefore that the laser stylus has an advantage in more accurately assessing the calibration specimen but is less accurate on surfaces that have abrupt changes with steep surfaces. Studies which use the profilometer to measure changes in height of a surface, such as measuring the marginal discrepancy of restorations [14], would be best performed using a contact stylus. Similarly, studies of surface profiles for wear comparisons [15,16] would benefit from the smoothing effect of a contact stylus to reduce extraneous ‘‘noise’’ which would interfere with subsequent wear analysis. By contrast, for studies of surface texture where surfaces have to be quantified such as following polishing [17], the laser stylus has the advantage of a more accurate assessment over a conventional contact stylus. This is provided that an appropriate surface parameter is chosen to quantify the surface, it is often not sufficient to measure the Ra and additional parameters will need to be used [8] However, the laser stylus requires a flat surface for optimum results and therefore on curved surfaces such as cast restorations [18] a contact stylus would then be more appropriate. This study indicates that the laser stylus can operate on semi-translucent materials such as Dicor and sputter coating of the samples with a layer of metal to aid reflectivity is not necessary.
S.A. Whitehead et al. / Dental Materials 15 (1999) 79–86
85
Fig. 6. Tracings of the surface of the ceramic finished using method D. The upper trace (A) shows the surface traced by the diamond stylus whilst the lower trace (B) represents the surface traced by the laser stylus. The differences between the traces are quite marked and show the increased sensitivity of the laser stylus.
5. Conclusion The results of this study indicate that there are significant differences between the results obtained from surface
texture measurements using a non-contact laser stylus and a diamond contact stylus method despite being mounted in the same profilometer. This is valid for both the test specimens of known surface texture provided by the
Fig. 7. Diagram to illustrate the problems that a contact stylus has in tracing a narrow pit or groove. The stylus fails to record the full depth of the groove in sample B. Despite the differences in samples A and B the traces are similar.
86
S.A. Whitehead et al. / Dental Materials 15 (1999) 79–86
Fig. 8. Diagram to illustrate the ability of parameters derived from the Abbot–Firestone curve to give meaningful information on profile shape. All the surface profiles are different but have similar Ra or Rz values. The Rpm;Rz ratio can indicate if the surface is spikey and the material ratio gives information as to the porosity of the subsurface. The actual shape of the Abbot–Firestone curve also describes the topography of the subsurface.
manufacturers and for a ‘‘real world’’ simulation using contoured and finished Dicor ceramic blocks. The only significant agreement between the two styli was found for the Ra parameter which should in itself not be used to describe the roughness of a surface because the parameter is not sensitive to profile shape. Owing to the properties of the stylus used it is essential that the limitations of the surface profilometry be recognised. Therefore, caution should be exercised when comparing the results of surface texture studies of dental hard tissues and restorative materials using varying types of stylus attached to a surface profilometer. References [1] A. Jokstad, I.A. Mjor, K. Nilner, et al., Clinical performance of three anterior restorative materials over 10 years, Quint. Int. 25 (1994) 101. [2] M. Ferrari, E. Bertelli, W. Finger, A five year report on a enameldentinal bonding agent and microfilled resin system, Quint. Int. 24 (1993) 735. [3] R. De Long, M. Pintado, W.H. Douglas, The wear of enamel opposing shaded ceramic restorative materials: an in vitro study, J. Prosthet. Dent. 68 (1992) 42. [4] Y. Momoi, K. Hirosaki, A. Kohno, F. McCabe, In vitro toothbrushdentifrice abrasion of resin modified glass ionomers, Dent. Mater. 13 (1997) 82. [5] M. Sander, A Practical Guide to the Assessment of Surface Texture, Feinpruf-Perthen GmbH, Gottingen, Germany, 1991. [6] G.W. Marshall, M. Balooch, X. Tench, et al., Atomic force microscopy of acid effects on dentin, Dent. Mater. 9 (1993) 265. [7] A. Wennerberg, R. Ohlsson, B.G. Rose´n, B. Andersson, Characterizing
[8]
[9]
[10]
[11]
[12]
[13] [14] [15]
[16] [17]
[18]
threedimensional topography of engineering and biomaterial surfaces by confocal laser scanning and stylus techniques, Med. Eng. Phys. 18 (7) (1996) 548. S.A. Whitehead, A.C.S. Shearer, N.H.F. Wilson, et al., Comparison of methods for measuring surface roughness of ceramic, J. Oral Rehabil. 22 (1995) 421. S.A. Whitehead, A.C. Shearer, D.C. Watts, N.H.F. Wilson, Surface texture changes of a composite brushed with tooth whitening dentifrices , Dent. Mater. 12 (1996) 315. V.B. Haywood, Ho. Heymann, R.P. Kusy, et al., Polishing porcelain veneers: a SEM and specular reflectance analysis, Dent. Mater. 4 (1988) 116. V.B. Haywood, H.O. Heymann, M.S. Scurria, Effects of water, speed and experimental instrumentation on finishing and polishing porcelain intra-orally, Dent. Mater. 5 (1989) 118. I. Krejci, in: W.H. Mo¨rmann (Ed.), International Symposium on Computer Restorations: State of the Art of the CEREC Method, Quintessence Publications, Chicago, USA, 1991, p. 245. S.A. Whitehead, L.Y. Lo, D.A. Watts, N.H.F. Wilson, Changes of surface texture of enamel in vivo, J. Oral Rehabil. 24 (1997) 449. B. Tarim, S. Suzuki, S. Suzuki, C.F. Cox, Marginal integrity of bonded amalgam restorations, Am. J. Dent. 9 (2) (1996) 72. M.R. Pintado, G.C. Anderson, R. DeLong, W.H. Douglas, Variation in tooth wear in young adults over a two-year period, J. Prosthet. Dent. 77 (3) (1997) 313. K. Kawai, K.F. Leinfelder, In vitro evaluation of OCA wear resistance of posterior composites, Dent. Mater. 11 (4) (1995) 246. D. Bouvier, J.P. Duprez, M. Lissac, Comparative evaluation of polishing systems on the surface of three aesthetic materials, J. Oral Rehabil. 24 (12) (1997) 888. D. Chan, V. Guillory, R. Blackman, K.H. Chung, The effects of sprue design on the roughness and porosity of titanium castings, J. Prosthet. Dent. 78 (1997) 400.
Comparison of two stylus methods for measuring surface texture S.A. Whitehead*, A.C. Shearer, D.C. Watts, N.H.F. Wilson University of Manchester Dental School, Higher Cambridge Street, Manchester M15 6FH, UK Received 2 November 1998; received in revised form 18 December 1998; accepted 21 December 1998
Abstract Objectives: The purpose of this study was to evaluate and compare two ‘‘stylus’’ methods for measuring surface texture of dental tissues and materials Methods: The two styli chosen were a contact diamond stylus and a non-contact focussed laser stylus attached to the same measuring apparatus. Results: These indicate that there are significant differences between those obtained from surface texture measurements using a noncontact laser stylus and a diamond contact stylus method despite being mounted in the same profilometer. This is valid for both the test specimens of known surface texture, provided by the manufacturers, and for a ‘‘real world’’ simulation using contoured and finished Dicor ceramic blocks. The only significant agreement between the two styli was found for the Ra parameter. This should not be used alone to describe the roughness of a surface because the parameter is not sensitive to profile shape. Owing to the properties of the stylus used it is essential that the limitations of surface profilometry be recognised. Significance: Caution should be exercised when comparing the results of surface texture studies of dental hard tissues and restorative materials using varying types of stylus attached to a surface profilometer. q 1999 Academy of Dental Materials. Published by Elsevier Science Ltd. All rights reserved. Keywords: Contact diamond stylus; Non-contact laser stylus; Surface profilometry
1. Introduction The ideal dental restorative material and associated techniques continue to be the goal of material scientists and clinicians alike. The ‘‘gold standard’’ for restoration quality could be considered to be its long term performance in clinical service [1,2]; however, laboratory tests are useful as indicators of clinical performance and have practical benefits. The widely used laboratory technique is to measure the surface texture of a surface and relate the surface texture to a clinical parameter such as wear resistance [3,4] or to use surface texture measurements to compare materials and techniques (polishing for example). There are several methods currently available to measure the texture of a surface; these include contact stylus tracing, laser reflectivity, non-contact laser stylus metrology, scanning electron microscopy, and compressed air measuring [5]. More recently, atomic force microscopy has been evaluated for use in assessing the surface texture [6,7]. Of these methods the contact stylus tracing method is probably the
most commonly used. There are several shortcomings with the use of this method which relate to the mass of the measuring stylus and the physical configuration of the contacting tip. The authors have already described the failings of using a laser reflection method for describing surface texture [8,9] However a laser type ‘‘stylus’’ was developed which replaces a conventional contact stylus that does not rely on the measurement of scattered laser light. The aim of this study was to compare the performance of a contact stylus and non-contact (laser) stylus. Initially, comparisons were made on samples supplied by the manufacturer. Following this, a ‘‘real world’’ situation was selected to validate the performance of the two systems in a dental materials context. The measurement of surface texture of finished ceramic was chosen as such restorations may be difficult to contour and effectively polish in vivo [10–12] and also ceramic materials may be semi-translucent which may influence the performance of the laser stylus.
2. Materials and methods * Corresponding author. Tel.: 1 44-161-275-6660; fax: 1 44-161-2756710. E-mail address: [email protected] (S.A. Whitehead)
Two surface measuring styli were evaluated: a diamond contact stylus (Fig. 1) and a laser stylus (Fig. 2). To ensure
0109-5641/99/$20.00 1 0.00 q 1999 Academy of Dental Materials. Published by Elsevier Science Ltd. All rights reserved. PII: S0109-564 1(99)00017-2
80
S.A. Whitehead et al. / Dental Materials 15 (1999) 79–86
ware translated the data into the selected parameters. Sixty four measurement runs were taken from the centre of the calibration specimen over an area of 5.6 × 5.6 mm 2 (Fig. 3). The calibration specimen was replaced with the optically flat glass plate onto which a groove of 8.8 mm had been prepared. The groove was then profiled using the diamond stylus to assess the effectiveness of the profiling system at reproducing the groove. 2.2. Measurement by non-contact laser stylus tracing
Fig. 1. Photograph of the diamond stylus tracing the surface of a finished block of ceramic. The non-contact laser stylus was attached to the same apparatus to duplicate the measuring setup.
that the results could be compared, the styli were mounted in the same measuring system, a Perthometer S8P (FeinprufPerthen GmbH, Gottingen, Germany). Prior to the evaluation of the ceramic surfaces, the performance of the two styli were evaluated on two test surfaces which were provided by the manufacturer. These surfaces were a groove of 8.8 mm prepared in an optically flat glass plate and a machined metal surface of known surface texture (Ra 1.98 mm, Rpm 9.2 mm, Rz 8.65 mm) which will be referred to as the calibration specimen. 2.1. Measurement of the manufacturers samples by contact stylus tracing The Perthometer S8P is a stylus based surface measuring instrument for the acquisition, graphical presentation, evaluation and documentation of surface profiles. A diamond stylus of 5 mm radius and stylus angle 908 was traversed at a constant speed across the calibration specimen with a force of 0.8 N (Fig. 1). The tracing run was parallel to the long axis of the longitudinal features of the machined calibration specimen. For the purposes of this study a tracing length of 5.6 mm was chosen. Data from an inductive transducer coupled to the stylus were relayed to the central processing unit of the apparatus and on-board firm-
Fig. 2 shows a cut-away diagram of the Focodyn laser stylus. A laser diode built into the casing serves as a coherent light source (the manufacturers only say that the laser source emits at the red end of the visible spectrum). The laser light initially passes through a collimator and is then guided by a prism to the micro objective lens. Another prism in the measuring head guides the beam onto the surface with a spot size of 1 mm. The micro objective lens focuses the beam exactly 0.9 mm below the aperture of the measuring arm. The light reflected back from the surface returns the same way through the optical system where it is directed by a beam splitter to the detector. This detector is part of a sensitive control loop which by means of a powerful linear motor readjusts the measuring arm continuously so that a distance of 0.9 mm to the surface remains constant (dynamic focussing). Thus when the Focodyn stylus is passed across a surface, the measuring arm follows the surface profile and provides a trace of the surface. For the purposes of this study the diamond contact stylus was replaced by the Focodyn laser stylus with all other measuring parameters remaining unchanged and surface texture measurements of the calibration specimen and 8.8 mm groove repeated. 2.3. Ceramic sample preparation To compare and contrast the performance of the two styli systems in a ‘‘real world’’ situation, a series of different surface textures were produced on blocks of machinable ceramic. Eighteen dark shade Dicor MGC (Corning, NY, USA) blocks were used. To simulate contouring of the occlusal and approximal surfaces, a diamond bur
Fig. 2. Cut-away diagram of the Focodyn laser stylus to show the mode of operation and the path of the focussed laser beam.
S.A. Whitehead et al. / Dental Materials 15 (1999) 79–86
81
Fig. 3. Pseudo 3D image from 64 consecutive profiles of the calibration specimen. The profiles occupy an area of 5.6 × 5.6 mm 2 in the centre of the specimen. This was the area profiled by the contact and non-contact (laser) styli. The manufacturer’s data regarding this specimen is: distance between peaks (l ) 0.8 mm, Ra 1.98 mm, Rpm 9.2 mm, Rz 8.65 mm.
(100 mm) was applied to all the blocks (Premier Dental Products, Norristown, PA, USA). The burs were held in a friction grip, high speed handpiece (Mid-west Quiet-air, Sybron Corp., Des Plaines, IL, USA) and used at moderate speed and pressure (0.1 MPa) with coolant. Each bur was Table 1 Outline of the finishing sequences and protocols for the preparation of the ceramic samples used to compare the two measuring systems Instruments used/Method Method A 30 blade tungsten carbide a Shofu porcelain points b Truluster diamond paste c Method B Two striper MFS microfine diamonds 80, 40, 15 mm d Shofu porcelain points Truluster diamond paste Method C 30 blade tungsten carbide Shofu porcelain points Two striper diamond polishing system Method D Baker Curson bur e Truluster diamond paste Method E Two striper MFS microfine diamonds 80, 40, 15 mm 3M Soflex disc f (coarse, medium, fine) Method F 30 blade tungsten carbide Shofu porcelain points Truluster diamond paste a
Protocol
High speed dry Slow speed wet
2.4. Surface roughness parameters selected High speed wet
Slow speed wet High speed dry Slow speed wet
High speed wet
High speed wet
Moderate speed dry
High speed dry Slow speed wet
Kerr/Sybron, Romulus, MI, USA. Shofu, Menlo Park, CA, USA. c Braesseler, Savannah, GA, USA. d Premier Dental Co., Norristown, PA, USA. e Ash, London, UK. f 3M Co., St Paul, MN, USA. b
applied to the surface for 30 s in one direction using light intermittent pressure. Each of the six finishing sequences detailed in Table 1 was applied to three blocks. All burs were applied to the surface in one direction only using light intermittent pressure. Diamond and tungsten carbide burs were used until the ceramic surface appeared uniform and no further improvement apparent. Diamond pastes were applied in a bristle brush for 20 s at moderate speed and pressure. The surface of the Dicor blocks were profiled with the contact and laser styli using an identical protocol to that used for the calibration specimen.
To describe the surface texture of the finished ceramic samples, several parameters were selected. This is because some commonly used surface parameters are true amplitude parameters (Ra for example) and give no information as to the shape of the recorded profile. The selected surface parameters were Ra, Rz, Rpm, Rpm:Rz ratio and the material ratio [Mr] [5,13] and are described in the following: 2.4.1. Ra parameter The Ra parameter describes the overall roughness of a surface and can be defined as the arithmetical average value of all absolute distances of the roughness profile from the centre line within the measuring length and is R derived from the equation:Ra 1=l 10 uy xudx: 2.4.2. Rz parameter The Rz surface parameter can be defined as the average maximum peak to valley height of five consecutive sampling lengths within the measuring length and is used to describe the degree of roughness of the surface of the sample as given by: Rz (1/n)(Z1 1 Z2 1 … 1 Zn). 2.4.3. Rpm parameter The Rpm surface parameter is the mean value of the
82
S.A. Whitehead et al. / Dental Materials 15 (1999) 79–86
levelling depths of five consecutive sampling lengths as given by Rpm Rp=1 1 Rp=2 1 Rp=3 1 Rp=4 1 Rp=5 =5: Exceptional profile peaks are thus only partly considered. By contrast to the surface parameters already described (Ra, Rz), the Rpm gives reliable information on the profile shape. Small Rpm values characterise a surface featuring wide peaks and narrow valleys whilst greater Rpm values indicate a spiky, sharp ridged profile (Fig. 8). 2.4.4. Rpm:Rz ratio The Rpm:Rz ratio is of special interest because the value also gives valuable information on the profile shape. A ratio higher than 0.5 indicates a sharp ridge profile, a ratio smaller than 0.5 indicates that the profile is rounded. 2.4.5. Material ratio (formerly micro bearing ratio (tp)) The material ratio (Mr) parameter is assessed from the roughness profile (R-profile). Long wave components such as waviness are suppressed by a special filter and the Mr calculated as: Mr 1=lm L1 1 L2 Ln 100%: Graphically this can be represented by the Abbot–Firestone curve derived from the serial section levels (Fig. 4). It is a common practice to select a reference level which just fits into the profile (usually Mr 2%). The application of this reference level provides a mathematical line across the sample from which the remainder of the measurements are taken. For the purposes of this investigation the level at Mr 2% was used to determine that percentage of material present 5 mm down into the surface of the material. 2.5. Data analysis The data from the contact and non-contact laser stylus were collected for each of the samples studied and for each surface parameter selected. The data was collated for analysis using the arcus pro-stat statistical package
(Medical Computing, Aughton, West Lancashire, UK) to assess the comparisons between the two measuring systems using the correlation coefficient r 2. P and t values were calculated to determine if any of the surface roughness parameters derived from the contact stylus method had a significantly better correlation to the data derived from the noncontact method.
3. Results 3.1. Test surfaces The results of the two styli on the test surfaces are summarised in Table 2. There are significant differences between all the measurements for the contact and laser styli (P . 0.05). Fig. 5 shows sample traces from the two measuring systems on the calibration specimen and the 8.8 mm groove which indicate how the differing styli interpret the surfaces. 3.2. Finished ceramic (Dicor) The results of the measurements of the surface roughness of the finished ceramic samples are given in Table 3. The correlation coefficients, t and p values are given in Tables 4 and 5 which show the relationship between the results obtained using the contact and non-contact tracing methods. It is clear that there is only one significant agreement (P , 0.05) between the laser non-contact and diamond contact stylus methods; this is when comparing the surface texture parameter Ra. For all other surface texture parameters measured there was no significant correlation between the two measuring systems. Fig. 6 shows the differences in the tracings obtained using the non-contact laser and diamond contact styi. The fidelity of the surface texture as ‘‘seen’’ by the laser stylus can be seen to be different from that quantified by the diamond stylus.
Fig. 4. Derivation of the Abbot–firestone curve from serial section levels. The material ratio chosen in this study was taken as 5 mm from the surface (the ‘‘surface’’ being that line at which there is 2% of material).
S.A. Whitehead et al. / Dental Materials 15 (1999) 79–86
83
Table 2 Results of the laser stylus and the contact stylus profiling the test surfaces compared to the data supplied by the manufacturer (standard deviation in parentheses). There were significant differences (P . 0.05) between the contact and laser styli for all the surface texture parameters measured. The bottom row of the table shows the data for the styli measuring the 8.8-mm groove Parameter
Manufacturer’s data for calibration specimen
Contact stylus measurement of calibration specimen
Laser stylus measurement of calibration specimen
Ra (mm) Rpm (mm) Rz (mm) Rpm:Rz ratio Mr at 2 5 mm Depth of groove in flat glass plate(mm)
1.98 9.2 8.65 1.06 n/a 8.8
1.7(0.03) 4.0(0.2) 6.6(0.4) 0.6 98(2)% 7.41
1.84(0.05) 5.36(0.18) 11.05(0.49) 0.49 45.8(5.5)% 11.9
4. Discussion The aim of this investigation was to compare contact (diamond stylus) and non-contact (laser stylus) methods of surface texture measurement. The measurements of the standard surfaces provided by the manufacturer suggest that the contact stylus system as delivered and calibrated by the manufacturer, underestimates the roughness and spikiness of the surface. This represents a physical constraint on behalf of a diamond stylus to fully represent deep surface features which are narrower than the stylus itself (Fig. 7). This effect may be overcome by using a narrower and longer stylus but in practice this results in a weak stylus which is liable to be easily damaged. The laser stylus gave more comparable results to the manufacturer’s data than the contact stylus apart when measuring the calibration specimen. However, measurement of the 8.8 mm groove produced an ‘‘overshot’’ at the
bottom of the groove which resulted in an artefact (Fig. 5). The results indicate that there are significant differences between the performance of the two measuring systems when measuring the test surfaces. This was surprising as these specimens are used by the manufacturers and support a personnel to calibrate the instrument on delivery and at subsequent maintenance visits. The difference in the results is therefore considered to be a fundamental attribute of the two measuring systems and not an artefact of the study design. The results from the profilometry of the ceramic samples indicate that there is little correlation between the two systems apart from when measuring the Ra surface texture parameter. This correlation is unsurprising as the Ra parameter is a simple amplitude descriptor and gives no indication of profile shape. Thus, if the surface is round edged or spiky then this parameter is unable to differentiate between the two and is a poor predictor of surface
Fig. 5. Sample tracings from the evaluation of the test specimens. Profiles A and B are from the contact and laser stylus respectively, and show the surface of the calibration specimen. Profiles C and D are of the 8.8 mm groove. Profile D is from the laser stylus and illustrates the overshoot of the trace as the stylus fails to find the floor of the groove and overcompensates giving an artefact.
84
S.A. Whitehead et al. / Dental Materials 15 (1999) 79–86
Table 3 Results of the averaged surface measurements using the non-contact laser and contact diamond styli for each finishing method Finishing regimen
Parameter
Laser stylus
Contact stylus
Method A
Ra (mm) Rz (mm) Rpm (mm) Rpm:Rz Tp ( 2 mm) Ra (mm) Rz (mm) Rpm (mm) Rpm:Rz Tp ( 2 mm) Ra (mm) Rz (mm) Rpm (mm) Rpm:Rz Tp ( 2 mm) Ra (mm) Rz (mm) Rpm (mm) Rpm:Rz Tp ( 2 mm) Ra (mm) Rz (mm) Rpm (mm) Rpm:Rz Tp ( 2 mm) Ra (mm) Rz (mm) Rpm (mm) Rpm:Rz Tp ( 2 mm)
0.4 0.36 1.3 0.38 1.37 0.47 4.17 1.4 0.34 2.8 0.43 4.1 1.3 0.32 2.64 1 8.8 2.3 0.27 4.14 0.59 5.3 1.5 0.27 3.69 0.36 3.0 1.1 0.36 2.47
0.33 1.85 0.97 0.55 2.3 0.4 1.93 1.03 0.55 2.57 0.27 1.54 0.76 0.48 1.54 5.37 2.89 1.3 0.47 3.41 0.26 1.11 0.64 0.57 2.63 0.31 1.4 0.79 0.57 2.35
Method B
Method C
Method D
Method E
Method F
Method A Method B Method C Method D Method E Method F
morphology (Fig. 8). The Rpm:Rz ratios determined by the laser stylus were somewhat higher than those found by the contact stylus methods for similar surfaces. Therefore the laser stylus profiled surfaces are more sharp ridged and spiky. This may be a true finding or an artefact caused by the inability of a diamond stylus tip of finite radius to accurately determine a sharp ridge or that the stylus ‘‘knocked off’’ the ridge during measurement. Conversely the laser non-contact stylus may be producing erroneous measurements because of an overshoot phenomenon as the laser stylus fails to find the tip of the ceramic ridge and produces a false ‘‘spike’’ or that the laser light is reflected from the Table 4 Correlation coefficients between the non-contact laser and contact diamond styli for each of the parameters included in the study. The only significant correlation is the Ra surface parameter (P , 0.05)
Ra (mm) Rz (mm) Rpm (mm) Rpm:Rz Mr (%)
Table 5 Correlation coefficients between the non-contact laser and contact diamond styli for each of the finishing regimens. The only significant correlation is for the finishing regimen ‘‘Method C’’ (P , 0.05)
r
r2
t
p
0.941 0.533 0.753 0.404 0.623
0.886 0.284 0.540 0.163 0.387
5.570 1.259 2.160 0.884 1.591
0.050 0.276 0.096 0.426 0.186
r
r2
t
p
0.477 0.850 0.936 0.014 0.621 0.828
0.288 0.723 0.875 0.019 0.386 0.686
0.942 0.279 4.603 0.245 1.372 2.560
0.416 0.067 0.019 0.822 0.264 0.080
subsurface rather than the actual surface. However, as the laser stylus can effectively track across a glass plate then the reflected laser energy from the semi-transparent Dicor would be sufficient to enable effective operation of the laser stylus. Similar differences are noted in the Mr in that the diamond stylus appears to ‘‘look’’ further into the valleys of the ceramic after finishing. This again may be an artefact caused by a ‘‘drop-off’’ phenomenon as the laser stylus overshoots the bottom of a valley to record a valley depth which is greater than that actually present. The material ratio is a useful parameter as it gives an indication of the porosity of a surface and can differentiate between surfaces of similar Ra and Rz (Fig. 8). Again the material ratios derived from the laser stylus were significantly different from the contact stylus both for the calibration specimen and for the ceramic blocks. This further illustrates the problems of profiling a surface which has narrow pits and pores using a diamond stylus which tends to smooth out such surfaces. It would seem therefore that the laser stylus has an advantage in more accurately assessing the calibration specimen but is less accurate on surfaces that have abrupt changes with steep surfaces. Studies which use the profilometer to measure changes in height of a surface, such as measuring the marginal discrepancy of restorations [14], would be best performed using a contact stylus. Similarly, studies of surface profiles for wear comparisons [15,16] would benefit from the smoothing effect of a contact stylus to reduce extraneous ‘‘noise’’ which would interfere with subsequent wear analysis. By contrast, for studies of surface texture where surfaces have to be quantified such as following polishing [17], the laser stylus has the advantage of a more accurate assessment over a conventional contact stylus. This is provided that an appropriate surface parameter is chosen to quantify the surface, it is often not sufficient to measure the Ra and additional parameters will need to be used [8] However, the laser stylus requires a flat surface for optimum results and therefore on curved surfaces such as cast restorations [18] a contact stylus would then be more appropriate. This study indicates that the laser stylus can operate on semi-translucent materials such as Dicor and sputter coating of the samples with a layer of metal to aid reflectivity is not necessary.
S.A. Whitehead et al. / Dental Materials 15 (1999) 79–86
85
Fig. 6. Tracings of the surface of the ceramic finished using method D. The upper trace (A) shows the surface traced by the diamond stylus whilst the lower trace (B) represents the surface traced by the laser stylus. The differences between the traces are quite marked and show the increased sensitivity of the laser stylus.
5. Conclusion The results of this study indicate that there are significant differences between the results obtained from surface
texture measurements using a non-contact laser stylus and a diamond contact stylus method despite being mounted in the same profilometer. This is valid for both the test specimens of known surface texture provided by the
Fig. 7. Diagram to illustrate the problems that a contact stylus has in tracing a narrow pit or groove. The stylus fails to record the full depth of the groove in sample B. Despite the differences in samples A and B the traces are similar.
86
S.A. Whitehead et al. / Dental Materials 15 (1999) 79–86
Fig. 8. Diagram to illustrate the ability of parameters derived from the Abbot–Firestone curve to give meaningful information on profile shape. All the surface profiles are different but have similar Ra or Rz values. The Rpm;Rz ratio can indicate if the surface is spikey and the material ratio gives information as to the porosity of the subsurface. The actual shape of the Abbot–Firestone curve also describes the topography of the subsurface.
manufacturers and for a ‘‘real world’’ simulation using contoured and finished Dicor ceramic blocks. The only significant agreement between the two styli was found for the Ra parameter which should in itself not be used to describe the roughness of a surface because the parameter is not sensitive to profile shape. Owing to the properties of the stylus used it is essential that the limitations of the surface profilometry be recognised. Therefore, caution should be exercised when comparing the results of surface texture studies of dental hard tissues and restorative materials using varying types of stylus attached to a surface profilometer. References [1] A. Jokstad, I.A. Mjor, K. Nilner, et al., Clinical performance of three anterior restorative materials over 10 years, Quint. Int. 25 (1994) 101. [2] M. Ferrari, E. Bertelli, W. Finger, A five year report on a enameldentinal bonding agent and microfilled resin system, Quint. Int. 24 (1993) 735. [3] R. De Long, M. Pintado, W.H. Douglas, The wear of enamel opposing shaded ceramic restorative materials: an in vitro study, J. Prosthet. Dent. 68 (1992) 42. [4] Y. Momoi, K. Hirosaki, A. Kohno, F. McCabe, In vitro toothbrushdentifrice abrasion of resin modified glass ionomers, Dent. Mater. 13 (1997) 82. [5] M. Sander, A Practical Guide to the Assessment of Surface Texture, Feinpruf-Perthen GmbH, Gottingen, Germany, 1991. [6] G.W. Marshall, M. Balooch, X. Tench, et al., Atomic force microscopy of acid effects on dentin, Dent. Mater. 9 (1993) 265. [7] A. Wennerberg, R. Ohlsson, B.G. Rose´n, B. Andersson, Characterizing
[8]
[9]
[10]
[11]
[12]
[13] [14] [15]
[16] [17]
[18]
threedimensional topography of engineering and biomaterial surfaces by confocal laser scanning and stylus techniques, Med. Eng. Phys. 18 (7) (1996) 548. S.A. Whitehead, A.C.S. Shearer, N.H.F. Wilson, et al., Comparison of methods for measuring surface roughness of ceramic, J. Oral Rehabil. 22 (1995) 421. S.A. Whitehead, A.C. Shearer, D.C. Watts, N.H.F. Wilson, Surface texture changes of a composite brushed with tooth whitening dentifrices , Dent. Mater. 12 (1996) 315. V.B. Haywood, Ho. Heymann, R.P. Kusy, et al., Polishing porcelain veneers: a SEM and specular reflectance analysis, Dent. Mater. 4 (1988) 116. V.B. Haywood, H.O. Heymann, M.S. Scurria, Effects of water, speed and experimental instrumentation on finishing and polishing porcelain intra-orally, Dent. Mater. 5 (1989) 118. I. Krejci, in: W.H. Mo¨rmann (Ed.), International Symposium on Computer Restorations: State of the Art of the CEREC Method, Quintessence Publications, Chicago, USA, 1991, p. 245. S.A. Whitehead, L.Y. Lo, D.A. Watts, N.H.F. Wilson, Changes of surface texture of enamel in vivo, J. Oral Rehabil. 24 (1997) 449. B. Tarim, S. Suzuki, S. Suzuki, C.F. Cox, Marginal integrity of bonded amalgam restorations, Am. J. Dent. 9 (2) (1996) 72. M.R. Pintado, G.C. Anderson, R. DeLong, W.H. Douglas, Variation in tooth wear in young adults over a two-year period, J. Prosthet. Dent. 77 (3) (1997) 313. K. Kawai, K.F. Leinfelder, In vitro evaluation of OCA wear resistance of posterior composites, Dent. Mater. 11 (4) (1995) 246. D. Bouvier, J.P. Duprez, M. Lissac, Comparative evaluation of polishing systems on the surface of three aesthetic materials, J. Oral Rehabil. 24 (12) (1997) 888. D. Chan, V. Guillory, R. Blackman, K.H. Chung, The effects of sprue design on the roughness and porosity of titanium castings, J. Prosthet. Dent. 78 (1997) 400.