Journal of Dental Research http://jdr.sagepub.

com/

Accuracy of Kubelka-Munk Reflectance Theory Applied to Human Dentin and Enamel

J.C. Ragain, JR and W.M. Johnston J DENT RES 2001 80: 449 DOI: 10.1177/00220345010800020901 The online version of this article can be found at: http://jdr.sagepub.com/content/80/2/449

Published by:
http://www.sagepublications.com

On behalf of:
International and American Associations for Dental Research

Additional services and information for Journal of Dental Research can be found at: Email Alerts: http://jdr.sagepub.com/cgi/alerts Subscriptions: http://jdr.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav Citations: http://jdr.sagepub.com/content/80/2/449.refs.html

Downloaded from jdr.sagepub.com by guest on April 20, 2011 For personal use only. No other uses without permission.

-Em.
J.C. Ragain, Jr.l and W.M. Johnston2
Naval Dental Research Institute, 3 1OA, B Street, Bldg 1-H, Great Lakes, IL 60088-5259; 'Captain, Dental Corps, US Navy, Executive Officer and Research Scientist; 2Professor, Section of Restorative Dentistry, Prosthodontics and Endodontics, The Ohio State University College of Dentistry, Columbus, OH; *corresponding author,

james.ragaingndri.med.navy.mil

Accuracy of Kubelka-Munk Reflectance Theory Applied to Human Dentin and Enamel

INTRODUCTION
The appearance of a tooth to a human observer is a complicated psychophysiological sensory process which is influenced by the illuminant and
the specific characteristics of the tooth. These characteristics include shape, size, position, and color of the tooth. The color of a tooth is determined by a combination of its optical properties. When light encounters a tooth, four phenomena associated with the interactions of the tooth with the light flux can be described: (1) specular transmission of the light flux through the tooth, (2) specular reflection at the surface, (3) diffuse light reflection at the surface, and (4) absorption and scattering of the flux within the dental tissues. The coloring material, or pigments, within the object will absorb various wavelengths of light, allowing other wavelengths to scatter out of the object. The process of selective wavelength absorption is the primary source of color (Hunter, 1975). The net effect of pigmented particles is to produce perfect diffusion of light in every direction. The methods of calculus can be used to derive theoretical relations of light fluxes into and out of a turbid or light-scattering material. Kubelka and Munk (1931) applied calculus to solve the differential equations for the change of light fluxes (two flux theory) as a function of scattering, absorption, and distance. The Kubelka-Munk (K-M) theory is one of the most frequently used theories to predict color matches, because it provides a reflectance model for translucent materials placed on backings of different colors. The equations used in this study are described in the

J Dent Res 80(2):449-452, 2001

ABSTRACT
The Kubelka-Munk (K-M) theory provides a reflectance model for translucent materials placed on backings of different colors. We hypothesize that Kubelka-Munk (K-M) theoretical diffuse reflectance spectra of dentin and enamel are in good agreement with observed diffuse reflectance. The aim of this study was to measure the reflectance of enamel and dentin specimens and to compare the measured values of reflectance with K-M theoretical values. Disc-shaped specimens of enamel, dentin, and enamel/dentin were prepared from extracted teeth. Diffuse reflectance spectra were measured on three backings by means of a reflectance spectrophotometer over every wavelength (X) from 400 to 700 nm at three thicknesses. The measured reflectance values were fit by non-linear regression to corrected K-M theory. The low value of the reported errors associated with the application of K-M theory illustrated that the theoretical diffuse reflectance spectra of dentin and enamel are in good agreement with the observed diffuse reflectance.

KEY WORDS: Kubelka-Munk theory, reflectance,

color, dentin, enamel.

Received April 10, 2000; Last revision September 26, 2000; Accepted November 30, 2000
A supplemental appendix to this article is published electronically only at http://www.dentalresearch.org.

Appendix. Spitzer and ten Bosch (1975) measured the reflectance and transmission of thin slabs of dental enamel at wavelengths between 220 and 700 nm. Measurements were made on five specimens by means of a spectrophotometer equipped with an integrating sphere. The theoretical model used was the two-flux model, and the scattering and absorption coefficients were computed. An absorption peak of 270 nm, which is outside the normal human visible light range (from 380 to 780 nm), was found common to all the spectra. The refractive indices of the specimens were very similar. It was concluded that the organic component (aromatic amino acids: tryptophan, tyrosine, and phenylalanine) was responsible for most or all of the observed optical absorption. Cook and McAree (1985) found that the application of the KubelkaMunk theory to the analysis of the diffuse reflectance spectra of selected porcelains and composite resin restorative materials accurately predicted the reflectance spectra for various thicknesses of specimens on different backings. They also found that the range of hue, value, and chroma of these restorative materials was consistent with that of natural tooth sections. However, because of the considerable difference in their reflectance spectra, these restorative materials were highly metameric with the natural dentition. In this study, we tested the hypothesis that K-M theoretical diffuse reflectance spectra of dentin and enamel are in good agreement with

Downloaded from jdr.sagepub.com by guest on April 20, 2011 For personal use only. No other uses without permission.

449

450
1.0 -

Ragain & Johnston

Backing: x Whlte

J Dent Res 80(2) 2001

0.8 0

+ Gray

80.x..
--

3 Black

. 0.4 0

..

--- +-

0.2 -+
0.0 0 0.02 0.04 0.06 0.08 0.1 Thickness (cm)

0.12

0.1

Figure 1. Measured values of diffuse reflectance plotted on curves representing Kubelka-Munk theoretical reflectance values for an enamel specimen at A = 446 nm.
observed diffuse reflectance. The aim of this study was to measure the reflectance of enamel and dentin specimens and to compare the measured values of reflectance with K-M theoretical values. By evaluation of the residuals, it could be determined if K-M theory is valid for enamel and dentin.

MATERIALS & METHODS Materials

Recently extracted teeth were collected and stored in doubledistilled water. The teeth were carefully examined under magnification to ensure that only intact, non-restored, non-carious teeth were selected for this study. Prior to the commencement of the study, the Ohio State University Institutional Review Board found the use of extracted teeth to be exempt from informed patient consent.

area using a diamond core drill bit (6.4 mm diamond drill bit, NO. 322020020, Abrasive Technology, Westerville, OH, USA) and continual water irrigation. Surface debris was removed from the cores by being gently brushed with a soft toothbrush with distilled water. The cores were stored in double-distilled water. The cores were further reduced to produce disc-shaped specimens of either enamel, dentin, or enamel and dentin together. The thickness of these specimens was achieved by grinding of the cored sections in a specimen-holder jig and a revolving polishing wheel (Vari/Pol, VP-50, Leco Corporation, St. Joseph, MI, USA). Depressions cut into the jig's polishing surface were calibrated to the desired thicknesses of the samples. It was determined, through magnified visual inspection, which cores would be used to produce discs of enamel, dentin, or enamel/dentin. The specimens were ground to 1.5 mm ( 0.03 mm) by means of the grinding wheel, increasing grits of sandpaper (from 320 to 600), and continuously flowing water. The specimens were then polished by means of 1.0-,um alumina and a chamois cloth on the polishing wheel. Both top and bottom surfaces were polished. A small notch was placed on the outer border of one surface of the discs by means of a 1/4 round dental carbide bur in a high-speed dental handpiece with water spray. The notch was placed in the disc so that the orientation of the disc in the spectrophotometer (SP) holder would be the same for all measurements. The reflectance spectra were then measured for each specimen in a randomized fashion. The grinding, polishing, notching, and measurement procedures were repeated for each specimen at 1.0 mm ( 0.03 mm) and 0.5 mm ( 0.03 mm). The surface not having the orientation notch was reduced. All polishing debris was thoroughly removed by being gently scrubbed with a soft brush and distilled water and by ultra-sonification. The specimens were stored in double-distilled water at all times when not being prepared or measured.

Measuring the Reflectance of the Enamel and Dentin Specimens

The specimens were placed in a spectrophotometer specimen holder similar to the one described by Johnston et al. (1996), which reduced the effects of scattered light edge-loss from the specimen. Optical contact between the specimen and the backing was improved by an interfacing layer of aqueous sucrose solution having a refractive index (n) of approximately 1.5. Sucrose solution was chosen because its refractive index could be made to approach that of the specimens (n = 1.655) and the tile backings without altering the structure of the specimens. The surface not having the notch was placed against the backing, and the notch was aligned with a mark on the SP holder. Diffuse reflectance spectra were measured on black, white, and gray backings by means of a reflectance spectrophotometer and an integrating sphere (Lambda 6, U/VIS spectrophotometer, The Perkin-Elmer Corporation, Norwalk, CT, USA) over every wavelength (A) from 400 to 700 nm. Absolute reflectance was then calculated for each measurement.

Preparation of the Enamel and Dentin Specimens

The roots of the teeth were sectioned at the CEJ by means of a diamond saw (Vari/Cut, VC-50, Leco Corporation, St. Joseph, MI, USA) with water irrigation. Plastic rings (2.5 cm in diameter) were fixed to a glass slab with sticky wax. The teeth were positioned in the plastic rings such that the enamel rods composing the buccal or lingual surfaces were parallel to the normal plane. We maintained this orientation by securing the teeth to the glass slab using melted base plate wax. Chemically cured provisional restoration acrylic (SNAP, shade clear, Parkell Biomaterials Division, Farmingdale, NY, USA) was mixed and poured between the crown and the ring, thus securing the tooth specimen to the ring. We maintained the temperature of the setting reaction of the acrylic as low as possible by placing the glass slab in cold water. After the acrylic had set, the rings were removed from the glass slab, and all remaining sticky wax was removed. The teeth were then sectioned in half in the mesial-distal direction by means of a diamond saw and copious amounts of water. To produce disc-shaped specimens, we secured the tooth sections on a drill press (Craftsman model No. 113.213100, Sears, Roebuck and Co., Chicago, IL, USA), and cored the mid-tooth

Determinafion of the Absorption and Scattering Coefficients of the Ename and Dentin Specimens
The measured reflectance values (enamel, 3 specimens; dentin, 4 specimens; and enamel/dentin, 2 specimens; each specimen at 3 thicknesses on 3 backings) at each wavelength were fit by the technique of non-linear regression to corrected K-M theory

Downloaded from jdr.sagepub.com by guest on April 20, 2011 For personal use only. No other uses without permission.

J Dent Res 80(2) 2001

Kubelka-Munk Theory Applied to Dentin and Enamel

H .1n v

451
Backing: x White

(Appendix, Equations 3-11, www.dentalresearch.org). The nonlinear regression technique (SAS Proc NLIN, SAS Institute Inc., Cary, NC, USA) uses a least-squares-error solution to find the values of absorption (K) and scattering (S) coefficients that best fit the observed reflectance (R) values, and produced a measure of error between the observed reflectance values and corrected K-M theory.

0.8
0

+ Gray

Statistical Analysis
For each enamel and dentin sample at each integral wavelength, the measured reflectance values for all thicknesses and backings were fit by a least-squares routine to Kubelka-Munk reflectance theory corrected for interfacial reflections. Since the reflectance was measured at three determined thicknesses for each of three measured backings, and since the remaining unknowns were the two K-M (absorption and scattering) coefficients, seven degrees of freedom would be left for the determination of the error in the theory. For analysis of these errors, the wavelengths at which each of the tristimulus values for the Commission International l'clairage (CIE) standard observer reaches a maximum (446, 555, and 599 nm) were chosen to represent the tristimulus characteristics of the human observer. These errors were then compared among the types of tissue, with wavelength as a repeated measure (within-subjects effect), in a repeated-measures analysis of variance (ANOVA). Post hoc pairwise comparisons were made by Tukey's Studentized Range (HSD) Test at a 0.05 significance level.

(a 0.6 ""a
0.4
0

a c

* Black

Xx

0.2 0.0

-7

0.02 0.04 0.06 0.08 0.1 Thickness (cm)

0.12 0.14 0.16

Figure 2. Measured values of diffuse reflectance plotted on curves representing Kubelka-Munk theoretical reflectance values for a dentin specimen at A = 555 nm.
specimens were significantly higher than the more homogeneous enamel and dentin specimens.

RESULTS
To illustrate the close agreement of measured reflectance values with corrected K-M theoretical values of reflectance, we plotted measured reflectance values of specimens on white, gray, and black backings against theoretical K-M values in Figs. 1-3 for representative specimens at the peak human observer tristimulus wavelengths. In these Figs., reflectance is plotted as a function of thickness. It is apparent that the theoretical and measured reflectance values are in close agreement for enamel and dentin at the three thicknesses measured (Figs. 1 and 2, respectively). However, the errors between theoretical and measured reflectance are more pronounced for the enamel/dentin specimens (Fig. 3). The reflectance values predicted by K-M theory were obtained by non-linear regression analyses at all wavelengths from 400 to 700 nm (Figs. 1, 2). All regressions were highly significant (P

DISCUSSION
When color is measured with an instrument that has a relatively small window of a few millimeters in diameter, a considerable fraction of the light entering the tooth is lost, because it emerges at the surface outside of the window of measurement (Van der Burgt et al., 1990). These edge-losses experienced in the measurement of teeth have also been reported by Bolt et al. (1994). The differences found among the various instruments used in colorimetry (Seghi et al., 1989) may be due to these edge-losses. The edge-loss effect was minimized in this study by the use of a stainless steel spectrophotometer specimenholder on which the surfaces that came into contact with the specimen were highly polished. This mirror-like surface reflected the light flux exiting the edges of the specimen back into the specimen (Johnston et al., 1996). While every effort was made to replicate the fabrication and measurement techniques for the specimens, error may be introduced in those steps. It was found that the thin specimen slabs desiccated rapidly, particularly enamel. Being natural tissue, the specimens were not perfectly homogeneous, and the

<0.001). The Table is a summary of errors at the peak human observer tristimulus wavelengths (A = 446, 555, and 599 nm) for enamel and dentin, including post hoc pairwise comparisons that were made Table. Summary of Non-linear Reggression Error Mean Squares for Diffuse Reflectance of Enamel, Dentin, and by Tukey's Studentized Enamel/Dentin Specimens Range (HSD) Test at alpha X = 599 nm X = 555 nm X = 446 nm = 0.05. All of the homogeSD X X SD SD X neous tissue specimens had Tissue (n) mean square errors of less 0.0001 7a 0.00010 0.0001 8a 0.00015 0.00005 0.000133a Enamel (3) than 0.0007 (absolute re0.00018 0.00034a 0.00021 0.00034a 0.00021 0.000265a flectance) in the 400- to Dentin (4) 0.00001 0.00018 0.00163b 0.00005 0.001,49b 1b 0.00121 Enamel/Dentin (2) 700-nm wavelength range. At each wavelength investidifference (p > 0.05) Tukey pairwise comparisons of tissue at each gated, the errors for the a,b The same letter indicates no si ignificant ,rror. wavelength for mean square e heterogenous enamel/dentin
Downloaded from jdr.sagepub.com by guest on April 20, 2011 For personal use only. No other uses without permission.

452
1.10-

Ragain & Johnston

J Dent Res 80(2) 2001

ACKNOWLEDGMENTS
Backing:
x

White

0.1

8
0.'6 % a
0

+ Gray

*K Black

4
0.
2-

This investigation was funded in part by the US Navy and by The Ohio State University. The views expressed in this article are those of the authors and do not reflect the official policy or position of the Department of the Navy, the Department of Defense, or the US Government. This paper is based on a thesis submitted to the graduate faculty, The Ohio State University, in partial fulfillment of the requirements for the PhD degree.

+--

.X
I...I.,

xK

REFERENCES
I

0.1u

0.02 0.04 0.06 0.08 0.1 Thickness (cm)

0.12 0.14 0.16

Figure 3. Measured values of diffuse reflectance plotted on curves representing Kubelka-Munk theoretical reflectance values for an enamel/dentin specimen at X = 555 nm.
degree of homogeneity varied among the specimens. Enamel may be relatively more homogeneous than dentin. The dentin slabs may have been composed of primary, reparative, or tertiary dentin. As the dentin slabs were reduced, the proportions of the various forms of dentin might have changed. This may lead to the larger errors observed in the dentin specimens. Also, if the organic component is responsible for the observed absorption, as concluded by Spitzer and ten Bosch (1975), then the relative quantities of these components could have been varied between and within specimens either naturally or through specimen preparation. The size and orientation of the dentinal tubules could also be altered as the specimens were reduced. This change in size and orientation could also effect the scattering and absorption within the specimens. All of the homogeneous tissue specimens had mean square errors of less than 0.0007 (absolute reflectance) in the 400- to 700-nm wavelength range. However, the low values of the reported errors indicate errors associated with the application of K-M theory to be low, illustrating that the theoretical diffuse reflectance spectra of dentin and enamel are in good agreement with the observed diffuse reflectance. It can be concluded that Kubelka-Munk theory accurately predicts the reflectance of enamel and dentin for various thickness and backing conditions.

Bolt RA, ten Bosch JJ, Coops JC (1994). Influence of window size in small-window colour measurement, particularly of teeth. Physics MedBiol 39:1133-1142. Cook WD, McAree DC (1985). Optical properties of esthetic restorative materials and natural dentition. J Biomed Mater Res 19:469-488. Duntley SQ (1942). The optical properties of diffusing materials. J Optic Soc Am 32:61-70. Hunter RS (1975). The measurement of appearance. New York: Wiley and Sons. Johnston WM, Hesse NS, Davis BK, Seghi RR (1996). Analysis of edge-losses in reflectance measurements of pigmented maxillofacial elastomer. JDent Res 75:752-760. Judd DB (1942). Fresnel reflection of diffuse incident light. Natl Bur Stds JRes 29:329-332. Judd DB, Wyszecki G (1975). Color in business, science, and industry. 3rd ed. New York: Wiley and Sons. Kortum G, Lohr TE (1969). Reflectance spectroscopy. New York: Springer. Kubelka P (1948). New contributions to the optics of intensely lightscattering materials, Part 1. J Opt Soc Am 38:448-457. Kubelka P, Munk F (1931). Ein beitrag zur optik der farbanstriche. Z Techn Physik 12:593-601. Saunderson JL (1942). Calculation of the color of pigmented plastics. J Opt Soc Am 32:727-736. Seghi RR, Johnston WM, O'Brien WJ (1989). Performance assessment of colorimetric devices on dental porcelains. J Dent Res 68:17551759. Spitzer D, ten Bosch JJ (1975). The absorption and scattering of light in bovine and human dental enamel. Calcif Tissue Res 17:129-137. Van der Burgt TP, Ten Bosch JJ, Borsboom PC, Kortsmit WJ (1990). A comparison of new and conventional methods for quantification of tooth color. JProsthet Dent 63:155-162. Wyszecki G, Stiles WS (1967). Color science concepts and methods, quantitative data and formulas. New York: Wiley and Sons.

Downloaded from jdr.sagepub.com by guest on April 20, 2011 For personal use only. No other uses without permission.