- Email: [email protected]
Elimination, via high-rate laser dilatometry, of structural relaxation during thermal expansion measurement of dental porcelains
dental materials Dental Materials 15 (1999) 390–396 www.elsevier.com/locate/dental
Elimination, via high-rate laser dilatometry, of structural relaxation during thermal expansion measurement of dental porcelains S.S. Khajotia a, J.R. Mackert Jr b,*, S.W. Twiggs b, C.M. Russell c, A.L. Williams b a
b
Department of Dental Materials, University of Oklahoma, Oklahoma City, OK, USA Section of Dental Materials, Department of Oral Rehabilitation, Medical College of Georgia, Augusta, GA, USA c Office of Biostatistics, Medical College of Georgia, Augusta, GA, USA Received 2 March 1999; accepted 20 May 1999
Abstract Objectives: Thermal expansion measurement of glassy materials is complicated by thermal history effects. Excess volume—trapped in quenched dental porcelains after firing—collapses via structural relaxation on first slow heating during conventional dilatometry, making the thermal expansion coefficient (a ) obtained on first heating unreliable. The purpose of this study was to determine whether porcelain thermal expansion measurement at high thermal rates could minimize the influence of thermal history. Methods: Eight thermal expansion specimens each of six body porcelains and the Component No.1 (leucite-containing) frit prepared according to the patent by Weinstein et al. (US Patent No. 3,052,982) were subjected to three heat–cool conventional dilatometry runs at 38C/ min, while eight thermal expansion specimens of each porcelain were reserved as untreated controls. Eight hollow, cylindrical specimens of the same brands were subjected to three heat–cool laser dilatometer thermal expansion runs at 6008C/min, while eight cylindrical specimens of each porcelain were reserved as untreated controls. Thermal expansion data (25–5008C) of all specimens were subjected to repeated measures analysis of variance. Results: The a obtained on first slow heating was significantly lower than values for succeeding slow heat and cool runs in all porcelains P , 0:001: High-rate a obtained on first heating was not significantly different from values of succeeding heat and cool runs in all porcelains P . 0:05: Significance: Conventional dilatometer measurements demonstrated occurrence of structural relaxation, as evidenced by the significant difference in the first heating and subsequent runs. High-rate laser dilatometry eliminated structural relaxation, thereby providing a thermal expansion measurement that is free of interference from thermal history effects. q 1999 Published by Elsevier Science Ltd. All rights reserved. Keywords: Laser dilatometry; Thermal expansion measurement; Dental porcelains
1. Introduction Dental porcelain is a complex composite of leucite (K2O·Al2O3·4SiO2) and oxides in a glass matrix. During the fabrication of a typical porcelain-fused-to-metal (PFM) restoration, the restoration is removed from the furnace at high firing temperatures (typically 10008C) directly into room air. This air quench results in the rapid cooling of the porcelain from a molten state (through its glass transition temperature range) down to room temperature with cooling rates in the region of 6008C/min [1]. During such rapid cooling, the molecules do not have
* Corresponding author. Tel.: 1 706-721-3354; fax: 1 706-721-6276. E-mail address: [email protected] (J.R. Mackert Jr)
sufficient time to undergo re-ordering to form the most dense configuration. As a result, excess volume is trapped in the glass matrix, and the glass has a lower density, and hence, a higher specific volume [2,3]. When the porcelain is re-heated at a slow rate, this trapped excess volume undergoes structural relaxation [2,4,5]. If the expansion of the porcelain is measured during such re-heating, the porcelain expands normally until it is close to the glass transition temperature, at which point sufficient thermal energy is available for the molecules to rearrange themselves into a more dense structure having a smaller volume. This densification process competes with the porcelain thermal expansion, resulting in an overall effective expansion that is lower than would be the case if the porcelain was annealed prior to measurement [2,5]. Annealing the porcelain to a temperature close to, but less than, the glass transition temperature
0109-5641/99/$20.00 + 0.00 q 1999 Published by Elsevier Science Ltd. All rights reserved. PII: S01 09-5641(99)0006 2-7
S.S. Khajotia et al. / Dental Materials 15 (1999) 390–396
391
Table 1 Porcelains included in study Code
Brand
Manufacturer
Type, lot #
BIO CER CII CRY VIT WIL CNO
Biobond Ceramco Ceramco II Crystar Vita Will-Ceram Component no. 1
Dentsply International, Inc., York, PA, USA Ceramco, Inc., Burlington, NJ, USA Ceramco, Inc., Burlington, NJ, USA Unitek/3M (Shofu Products), Monrovia, CA, USA Vident, Baldwin Park, CA, USA Ivoclar Williams, Amherst, NY, USA J. F. Jelenko & Co., Armonk, NY, USA (custom fabricated for MCG)
Body porcelain, D313 Body porcelain, 94082222 Body porcelain, 1474 Body porcelain, 227906 Body porcelain, 238 Body porcelain, 0689 Leucite-containing frit from Weinstein et al. [18]
has been shown to cause relief of trapped excess volume, as demonstrated in glass-to-metal seals [6–8]. If the porcelain is slowly cooled down to room temperature after re-heating close to the glass transition temperature, it occupies a smaller volume than it did after the initial firing. Such a situation will create tangential tensile stresses, which could cause the porcelain to crack if the stresses are large enough, resulting in failure of the restoration [5]. The thermal expansion of glass depends largely on prior heat treatment and cooling rate [2]. A glass has considerably different thermal expansion coefficients depending on whether the glass has been cooled slowly or rapidly, or has been annealed. The collapse of trapped excess volume is most clearly demonstrated during the conventional dilatometric measurement of the coefficient of linear thermal expansion of a glass [2]. The importance of thermal history (i.e. cooling rate after firing and number of firings) in the thermal expansion of PFM porcelains has been amply demonstrated by several investigators. Fairhurst et al. [4] established that the a -measurement obtained on first heating was not reproducible and therefore was unreliable, and they recommended that a be measured on cooling or following a second heating run. Dorsch [9] demonstrated the dependency of dental porcelain thermal expansion on cooling rate by showing an increase in a -values following protracted slow cooling. He also proved that the a -values of the body porcelains investigated increased with increasing numbers of firing cycles. These effects may have been the result of precipitation of additional leucite during the slow cooling [10] or multiple firing [11]. Fairhurst et al. [1] and Twiggs et al. [12] were able to show that the glass transition ranges of dental porcelains shifted to higher temperatures with increasing heating and cooling rates, thus possibly having a significant deleterious effect on the thermal expansion coefficient match of porcelain with alloy. The leucite content of feldspathic dental porcelains is probably the most significant factor influencing their thermal expansion [13,14]. Therefore, precipitation of additional leucite during the slow heating or cooling of dental porcelain specimens during the thermal expansion measurement itself may result in an a -measurement that is higher than the true value of a (i.e. the value of a without the effect of any confounding
factors on it). Under the thermal expansion measurement conditions used in the present work, however, the porcelains studied appear not to be subject to leucite precipitation during thermal expansion measurement [15]. Conventional dilatometry is an important tool in PFM research and has been widely used to study the thermal compatibility of porcelains with various alloys and to monitor factors that induce changes in their thermal expansion. Commercially available dilatometers are limited to controlled heating rates of less than 308C/min and to temperatures below the glass transition range of the porcelains undergoing thermal expansion measurement. Therefore, the coefficients of thermal expansion of different commercial dental porcelains have only been characterized at low heating and cooling rates below their glass transition range [1,4,9,14]. It would be advantageous to measure the thermal expansion coefficients of dental porcelains at higher heating rates for two reasons. The first reason is that at high heating rates there is insufficient time for additional amounts of leucite to crystallize [10], although at temperatures below about 7008C, the crystallization rate of leucite in dental porcelains is very low [16]. The second reason for the use of higher heating rates in thermal expansion measurement is that there is not sufficient time for the relaxation of trapped excess volume to occur. As discussed above, if the thermal expansion of a PFM porcelain is measured using a conventional dilatometer, the porcelain molecules are able to achieve their most dense configuration during the slow temperature rise of the first heating run due to collapse of trapped excess volume [2,3]. Thus, the structural relaxation process interferes with the thermal expansion measurement itself. To avoid this artifact during the thermal expansion measurement, investigators in this laboratory developed an experimental laser dilatometer [17] that is capable of measuring the thermal expansion of porcelains beyond their glass transition temperatures at rapid heating and cooling rates of 6008C/min. The purpose of this study, therefore, was to determine whether the measurement of porcelain thermal expansion at high heating rates could minimize the influence of thermal history effects—particularly structural relaxation—on thermal expansion measurement of dental porcelains.
392
S.S. Khajotia et al. / Dental Materials 15 (1999) 390–396
Fig. 1. Linear coefficients of thermal expansion measured during heating and cooling in a conventional double-pushrod dilatometer and a laser dilatometer for porcelain BIO.
2. Materials and methods Six commercial dental body porcelains and the Component No. 1 frit of the patent by Weinstein et al. [18] were selected for the study (Table 1). The porcelains were subjected to low-rate (conventional) dilatometry and highrate (laser) dilatometry to determine the effect of thermal history on porcelain thermal expansion. Fifty-six thermal expansion specimens were prepared for use in a conventional dilatometer, eight specimens for each of the porcelains listed in Table 1. Each specimen was fabricated by vibration-condensing a slurry of porcelain powder and distilled water into a split brass mold having inside dimensions of 35:6 × 3:4 × 1:6 mm3 : The porcelain bars were fired on non-sticking porcelain firing trays (Dentecon, Inc., Los Angeles, CA, USA) in a computercontrolled vertical-muffle dental porcelain furnace (Sunfire 10, The J.M. Ney Co., Bloomfield, CT, USA) as per manufacturer’s instructions for each brand of porcelain. The dimensions of each specimen were then reduced by grinding to 25:3 × 3:4 × 1:5 mm3 : The bars for each porcelain brand were subjected to three
successive heating and cooling runs at 38C/min between room temperature and slightly above the softening temperature of the porcelain in a horizontal double-pushrod LVDT dilatometer (Model Research II, Theta Industries, Inc., Port Washington, NY, USA). The specimen design used in the experimental laser dilatometer was a hollow, thin-walled cylinder with parallel ends having the necessary dimensions to fit inside an evacuated Vycorw tube in the laser dilatometer’s radiant furnace (Model E4, Research, Inc., Minneapolis, MN, USA). Each specimen was fabricated by vibration-condensing a slurry of porcelain powder and distilled water into a cylindrical mold 13.5 mm in height and 10.5 mm in outer diameter. The porcelain cylinders were fired on non-sticking porcelain firing trays (Dentecon, Inc., Los Angeles, CA, USA) in a computer-controlled vertical-muffle dental porcelain furnace (Sunfire 10, The J.M. Ney Co., Bloomfield, CT, USA) as per manufacturer’s instructions for each brand of porcelain. The height of each specimen was then reduced by grinding to 14.0 mm, to obtain cylinders with parallel ends (i.d. 8.0 mm; o.d. 10.0 mm). Each specimen was coated with a high-emissivity, silicone-based spray paint (Denplex 12008, “X” Laboratories, Inc., Wheeling, IL, USA) to permit efficient heating by the infrared furnace [19]. The specimen was then placed in a burnout oven at 3008C for 30 min to drive off all the volatile components from the spray paint so that the insides of the Vycorw tube in the laser dilatometer would not become clouded during thermal expansion measurement. The cylinders were subjected to three successive heating and cooling runs at 6008C/minute between room temperature and slightly above the softening temperature of the porcelain in a high heating-rate laser dilatometer. Thermal expansion data over the temperature range 25– 5008C were analyzed for all the runs using repeated measures ANOVA (SAS Institute, Cary, NC, USA; NCSS Statistical Software, Kaysville, UT, USA). P-values less than or equal to 0.05 were considered significant.
3. Results
Fig. 2. Linear coefficients of thermal expansion measured during heating and cooling in a conventional double-pushrod dilatometer and a laser dilatometer for porcelain CER.
The repeated measures ANOVA showed that for the lowrate dilatometer measurements, the mean value of the thermal expansion coefficient (a —in units of 10 6/8C) obtained on first heating was significantly lower than the values for succeeding heating and cooling runs in all the porcelains studied P , 0:001: Comparison of the pooled heating data (excluding the first heating run, which exhibited a significantly lower mean a than the second and third heating runs) with the pooled cooling data for each porcelain via repeated measures, ANOVA showed that for all porcelains except VIT and WIL, the cooling runs resulted in a higher measured value of a than the heating runs. Figs. 1–7 show the mean low-rate coefficient of thermal expansion value for individual heating and cooling runs of all the materials
S.S. Khajotia et al. / Dental Materials 15 (1999) 390–396
393
Fig. 3. Linear coefficients of thermal expansion measured during heating and cooling in a conventional double-pushrod dilatometer and a laser dilatometer for porcelain CII.
Fig. 5. Linear coefficients of thermal expansion measured during heating and cooling in a conventional double-pushrod dilatometer and a laser dilatometer for porcelain VIT.
investigated in this study. The 95% confidence intervals of the means for each material are represented in Figs. 1–7 by error bars. Based on the results of the repeated measures ANOVA for the high-rate (laser) dilatometer data, the value of the thermal expansion coefficient (a ) obtained on first heating was not statistically different from values for succeeding heating and cooling runs for all the porcelains studied P . 0:05: Comparison of the pooled heating data with the pooled cooling data for each porcelain via repeated measures ANOVA showed that for all porcelains, the cooling runs resulted in a higher measured value of a than the heating runs. The mean high-rate coefficient of thermal expansion value (10 6/8C) for individual heating and cooling runs of all the materials studied are shown in Fig. 2. The 95% confidence intervals of the mean for each material are represented by error bars in Fig. 2.
Glasses that are quenched in air (i.e. rapidly cooled) from
high firing temperatures undergo minimal re-ordering to form a structure that has excess volume trapped in it [2,3]. This is because of a lack of sufficient time for the most dense configuration to be reached. If the quenched porcelain is reheated at a slow rate, as in a conventional dilatometer measurement, it expands normally until it is close to its glass transition temperature, at which point sufficient thermal energy is available for the molecules to rearrange themselves into a more dense structure. Kingery et al. [2] noted that the collapse of trapped excess volume via structural relaxation is most clearly demonstrated during the conventional dilatometric thermal expansion measurement of a glass. The results of the low-rate thermal expansion measurements in the present investigation substantiate that premise. The coefficient of thermal expansion values measured on first slow heating were significantly lower than the values for succeeding low-rate heat and cool runs in all of the materials investigated, as expected, providing evidence for the occurrence of structural relaxation during the first slow rate thermal expansion measurement. The a values mean ^ standard deviation obtained for the six commercial dental porcelains on first slow heating ranged
Fig. 4. Linear coefficients of thermal expansion measured during heating and cooling in a conventional double-pushrod dilatometer and a laser dilatometer for porcelain CRY.
Fig. 6. Linear coefficients of thermal expansion measured during heating and cooling in a conventional double-pushrod dilatometer and a laser dilatometer for porcelain WIL.
4. Discussion
394
S.S. Khajotia et al. / Dental Materials 15 (1999) 390–396
Fig. 7. Linear coefficients of thermal expansion measured during heating and cooling in a conventional double-pushrod dilatometer and a laser dilatometer for porcelain CNO.
between 11:4 ^ 0:4 × 1026 =8C and 12:5 ^ 0:2 × 1026 =8C: These values compare favorably with values reported for the same brands of porcelain (11:9 × 1026 =8C to 13:7 × 1026 =8C by other investigators [4,9] over the temperature range 25–5008C after a single firing. However, the thermal expansion coefficient 12:6 ^ 0:2 × 1026 =8C for porcelain Ceramco II, measured on the third slow heating run appeared to be slightly lower than the a value of 13:4 ^ 0:1 × 1026 =8C reported by Piche´ et al. [14] for the same porcelain. This small difference could be attributable to instrumentation differences (single pushrod in the Piche´ et al. study vs. double pushrod in the present study) or other factors. The mean thermal expansion coefficient of Component No. 1 [18] was higher than the a -values of the commercial PFM porcelains. This result is consistent with the higher leucite content (45 wt%) of Component No. 1 compared with commercial dental PFM porcelains (20–25 wt% leucite). The importance of thermal history (i.e. cooling rate after firing and number of firings) in the thermal expansion of PFM porcelains is well documented. Fairhurst et al. [4] originally demonstrated that thermal expansion measurements obtained on first heating were unreliable. The results of this study provide convincing evidence that the thermal expansion coefficients measured on first heating are confounded by the structural relaxation of the material during the measurement itself, thereby making them inaccurate. Fairhurst et al. [4] have suggested that the measurement of a should be made on a cooling run or second heating run instead of on the first slow heating. For most of the porcelains studied in the present investigation (all except VIT and WIL), however, the conventional thermal expansion measurements made on cooling were statistically different P , 0:05 from those made on second and third heating. (The first heating run was excluded from the repeated measures ANOVA, because it was already lower owing to the structural relaxation of the glass matrix during the first heating.) All of the laser (high-rate) dilatometry runs exhibited significantly higher a -values during cooling
measurement compared with heating measurement. This behavior is consistent with the formation of a hysteresis loop in the heating and cooling curve of a glassy material, as described by Rekhson [20], wherein the thermal expansion coefficient of a glass is higher on cooling than on heating. This lag in expansion values can be attributed to the lag in achieving dynamic equilibrium conditions in the glass [20]. While the thermal expansion measurements made during cooling in the conventional (low-rate) dilatometer are not equivalent to those made during heating, the second and third heating runs are not statistically different from each other, and the first, second and third cooling runs are not different from each other. This finding is an important discovery with regard to the actual practice of dilatometer measurements on dental porcelains. It demonstrates that the thermal treatment the specimen receives during the thermal expansion run itself can be safely ignored. In the event of a malfunction or error in measurement during the first cooling or second heating run, it would be possible to re-run the same specimen without expecting a significant change in the measured thermal expansion value obtained, provided heating measurements are compared to heating measurements, and cooling measurements are compared to cooling measurements. Measurement of the coefficient of thermal expansion at high-rates has two major advantages, a lack of sufficient time for additional amounts of leucite to crystallize, and insufficient time for the collapse of trapped excess volume to occur via structural relaxation. Since conventional dilatometer measurements are plagued by the occurrence of structural relaxation in the specimen during the measurement itself, investigators at this laboratory developed an experimental laser dilatometer [12,17] that is capable of measuring the thermal expansion of porcelains beyond their glass transition temperatures at rapid heating and cooling rates of 6008C/min. The results of the high-rate thermal expansion measurements made using the laser dilatometer show that the thermal expansion coefficient obtained on first heating was not statistically different from values for succeeding heat and cool runs for all the materials studied. These results are consistent with the hypothesis that, at highrates, there is insufficient time for the relaxation of trapped excess volume to occur. Therefore, high-rate laser dilatometry eliminated structural relaxation, thereby providing a thermal expansion measurement that is free of interference from thermal history effects. For all except the first heating runs, the thermal expansion coefficients determined via the laser dilatometer were somewhat lower than those determined via the conventional dilatometer. For a few of the porcelains (BIO, CII, WIL), this difference was slight, whereas for others (CER, CRY) it was more pronounced. There are several possible reasons for this difference. One possibility is that the different methods of measuring temperature (optical pyrometer vs. thermocouple) and change in length (laser interferometry vs.
S.S. Khajotia et al. / Dental Materials 15 (1999) 390–396
double pushrod LVDT) in the laser dilatometer and conventional dilatometer could have produced systematic differences in the thermal expansion measurements. One known systematic difference between the two systems is the thermal gradient that existed in the laser dilatometry specimens during heating. The infrared heating of the hollow cylindrical specimens was modeled by Finite Element Analysis (FEA) to assess the thermal gradients that might exist during heating [21]. In this FEA analysis, the ends of the cylinders—particularly the end resting on the base mirror—were found to be somewhat cooler than the center portion throughout the laser dilatometer heating run. Having a portion of the specimen at a lower temperature than the bulk would cause the measured thermal expansion to be lower than the true value for the material. It is unlikely that this or similar systematic differences would be the only reason for the discrepancy, however, because the differences were not uniform across all porcelains—the laser and conventional measurements were quite similar for some of the porcelains. Another possible reason for the observed differences between the laser and conventional dilatometry measurements is the inherent difference in expansivity of the lower-density structure of a quenched glass and the higher-density structure of an annealed glass [3]. Since the laser dilatometry specimen is heated and cooled too rapidly for structural relaxation to occur, it is the lower-density (quenched) structure on which expansion is actually measured in the laser dilatometer. In the conventional dilatometer, structural relaxation occurs during the first heating run, so expansion is measured on the higherdensity (annealed) structure. The relevant structure for porcelain–metal thermal compatibility considerations is the lower-density (quenched) structure, because during the rapid cooling of a PFM restoration from the glass transition temperature down to room temperature, the glassy porcelain matrix has insufficient time to rearrange to the most dense configuration. Therefore, it is the thermal expansion coefficient of the lower-density structure that must be compatible with the metal. Thus, if the thermal expansion coefficients of the annealed and quenched structures were indeed different, it would be preferable to measure the expansion coefficient of the quenched (lower-density) structure.
5. Conclusions • Conventional dilatometer measurements exhibited evidence of structural relaxation during the first heating run, as evidenced by the significant difference in thermal expansion coefficients between the first heating and subsequent runs. • Laser dilatometer measurements were not affected by this thermal history effect. • For both conventional and laser dilatometry, thermal expansion coefficients obtained during cooling
395
measurements are slightly higher than those obtained during heating measurements. • Elimination of structural relaxation in dental porcelain can be accomplished by measuring thermal expansion at heating and cooling rates as high as 6008C/min, thereby providing a thermal expansion measurement that is free of interference from thermal history effects.
Acknowledgements The support of this work through NIH/NIDCR Grant No. R01 DE07806, the Dental Materials Section of the Department of Oral Rehabilitation, and the Department of Oral Biology of the Medical College of Georgia is gratefully acknowledged.
References [1] Fairhurst CW, Hashinger DT, Twiggs SW. The effect of thermal history on porcelain expansion behavior. J Dent Res 1989;68:1313– 1315. [2] Kingery WD, Bowen HK, Uhlmann DR. Introduction to ceramics, 2. New York: Wiley, 1976. [3] Touloukian YS, Kirby RK, Taylor RE, Lee TYR. Theory of thermal expansion of solids, Thermophysical properties of matter, 13. New York: Plenum Press, 1978. pp. 3a–16a. [4] Fairhurst CW, Anusavice KJ, Hashinger DT, Ringle RD, Twiggs SW. Thermal expansion of dental alloys and porcelains. J Biomed Mater Res 1980;14:435–446. [5] Mackert Jr JR. Effects of thermally induced changes on porcelain– metal compatibility. In: Preston JD, editor. Perspectives in dental ceramics, Proceedings of the Fourth International Symposium on Ceramics, Chicago: Quintessence, 1988. pp. 53–64. [6] Rekhson SM, Mazurin OV. Stress relaxation in glass and glass-tometal seals. Glass Technol 1977;18:7–14. [7] Rekhson SM. Annealing of glass-to-metal and glass-to-ceramic seals. Part 1. Theory. Glass Technol 1979;20:27–34. [8] Rekhson SM. Annealing of glass-to-metal and glass-to-ceramic seals. Part 2. Experimental. Glass Technol 1979;20:132–143. [9] Dorsch P. Thermal compatibility of materials for porcelain-fused-tometal (PFM) restorations. Ceramic Forum Int/Ber Dt Keram Ges 1982;59:1–5. [10] Mackert Jr JR, Evans AL. Effect of cooling rate on leucite volume fraction in dental porcelains. J Dent Res 1991;70:137–139. [11] Mackert Jr JR, Evans AL. Quantitative X-ray diffraction determination of leucite thermal instability in dental porcelain. J Am Ceram Soc 1991;74:450–453. [12] Twiggs SW, Searle JR, Ringle RD, Fairhurst CW. A rapid heating and cooling rate dilatometer for measuring thermal expansion in dental porcelain. J Dent Res 1989;68:1316–1318. [13] Mackert Jr JR, Butts MB, Fairhurst CW. The effect of the leucite transformation on dental porcelain expansion. Dent Mater 1986;2:32–36. [14] Piche´ PW, O’Brien WJ, Groh CL, Boenke KM. Leucite content of selected dental porcelains. J Biomed Mater Res 1994;28:603–609. [15] Mackert Jr JR, Khajotia SS, Russell CM, Williams AL. Potential interference of leucite crystallization during porcelain thermal expansion measurement. Dent Mater 1996;12:8–12. [16] Mackert Jr JR, Williams AL, Russell CM. Crystallization of leucite in dental porcelains in the temperature range 6508–10008C. J Dent Res 1999;78:127 Abstr. No. 174.
396
S.S. Khajotia et al. / Dental Materials 15 (1999) 390–396
[17] Twiggs SW, Mackert Jr JR, Khajotia SS. High-rate dilatometer. J Dent Res 1997;76:21 Abstr. No. 62. [18] Weinstein M, Katz S, Weinstein AB. Fused porcelain-to-metal teeth. US Patent No. 3,052,982, September 11, 1962. [19] Twiggs SW, Hashinger DT, Morena R, Fairhurst CW. Glass transition temperatures of dental porcelains at high heating rates. J Biomed Mater Res 1986;20:293–300.
[20] Rekhson SM. Thermal stresses, relaxation and hysteresis in glass. J Am Ceram Soc 1993;76:1113–1123. [21] Khajotia SS. The effect of leucite crystallization and thermal history on thermal expansion measurement of dental porcelains (dissertation). Augusta, GA: Medical College of Georgia, 1997.
Elimination, via high-rate laser dilatometry, of structural relaxation during thermal expansion measurement of dental porcelains S.S. Khajotia a, J.R. Mackert Jr b,*, S.W. Twiggs b, C.M. Russell c, A.L. Williams b a
b
Department of Dental Materials, University of Oklahoma, Oklahoma City, OK, USA Section of Dental Materials, Department of Oral Rehabilitation, Medical College of Georgia, Augusta, GA, USA c Office of Biostatistics, Medical College of Georgia, Augusta, GA, USA Received 2 March 1999; accepted 20 May 1999
Abstract Objectives: Thermal expansion measurement of glassy materials is complicated by thermal history effects. Excess volume—trapped in quenched dental porcelains after firing—collapses via structural relaxation on first slow heating during conventional dilatometry, making the thermal expansion coefficient (a ) obtained on first heating unreliable. The purpose of this study was to determine whether porcelain thermal expansion measurement at high thermal rates could minimize the influence of thermal history. Methods: Eight thermal expansion specimens each of six body porcelains and the Component No.1 (leucite-containing) frit prepared according to the patent by Weinstein et al. (US Patent No. 3,052,982) were subjected to three heat–cool conventional dilatometry runs at 38C/ min, while eight thermal expansion specimens of each porcelain were reserved as untreated controls. Eight hollow, cylindrical specimens of the same brands were subjected to three heat–cool laser dilatometer thermal expansion runs at 6008C/min, while eight cylindrical specimens of each porcelain were reserved as untreated controls. Thermal expansion data (25–5008C) of all specimens were subjected to repeated measures analysis of variance. Results: The a obtained on first slow heating was significantly lower than values for succeeding slow heat and cool runs in all porcelains P , 0:001: High-rate a obtained on first heating was not significantly different from values of succeeding heat and cool runs in all porcelains P . 0:05: Significance: Conventional dilatometer measurements demonstrated occurrence of structural relaxation, as evidenced by the significant difference in the first heating and subsequent runs. High-rate laser dilatometry eliminated structural relaxation, thereby providing a thermal expansion measurement that is free of interference from thermal history effects. q 1999 Published by Elsevier Science Ltd. All rights reserved. Keywords: Laser dilatometry; Thermal expansion measurement; Dental porcelains
1. Introduction Dental porcelain is a complex composite of leucite (K2O·Al2O3·4SiO2) and oxides in a glass matrix. During the fabrication of a typical porcelain-fused-to-metal (PFM) restoration, the restoration is removed from the furnace at high firing temperatures (typically 10008C) directly into room air. This air quench results in the rapid cooling of the porcelain from a molten state (through its glass transition temperature range) down to room temperature with cooling rates in the region of 6008C/min [1]. During such rapid cooling, the molecules do not have
* Corresponding author. Tel.: 1 706-721-3354; fax: 1 706-721-6276. E-mail address: [email protected] (J.R. Mackert Jr)
sufficient time to undergo re-ordering to form the most dense configuration. As a result, excess volume is trapped in the glass matrix, and the glass has a lower density, and hence, a higher specific volume [2,3]. When the porcelain is re-heated at a slow rate, this trapped excess volume undergoes structural relaxation [2,4,5]. If the expansion of the porcelain is measured during such re-heating, the porcelain expands normally until it is close to the glass transition temperature, at which point sufficient thermal energy is available for the molecules to rearrange themselves into a more dense structure having a smaller volume. This densification process competes with the porcelain thermal expansion, resulting in an overall effective expansion that is lower than would be the case if the porcelain was annealed prior to measurement [2,5]. Annealing the porcelain to a temperature close to, but less than, the glass transition temperature
0109-5641/99/$20.00 + 0.00 q 1999 Published by Elsevier Science Ltd. All rights reserved. PII: S01 09-5641(99)0006 2-7
S.S. Khajotia et al. / Dental Materials 15 (1999) 390–396
391
Table 1 Porcelains included in study Code
Brand
Manufacturer
Type, lot #
BIO CER CII CRY VIT WIL CNO
Biobond Ceramco Ceramco II Crystar Vita Will-Ceram Component no. 1
Dentsply International, Inc., York, PA, USA Ceramco, Inc., Burlington, NJ, USA Ceramco, Inc., Burlington, NJ, USA Unitek/3M (Shofu Products), Monrovia, CA, USA Vident, Baldwin Park, CA, USA Ivoclar Williams, Amherst, NY, USA J. F. Jelenko & Co., Armonk, NY, USA (custom fabricated for MCG)
Body porcelain, D313 Body porcelain, 94082222 Body porcelain, 1474 Body porcelain, 227906 Body porcelain, 238 Body porcelain, 0689 Leucite-containing frit from Weinstein et al. [18]
has been shown to cause relief of trapped excess volume, as demonstrated in glass-to-metal seals [6–8]. If the porcelain is slowly cooled down to room temperature after re-heating close to the glass transition temperature, it occupies a smaller volume than it did after the initial firing. Such a situation will create tangential tensile stresses, which could cause the porcelain to crack if the stresses are large enough, resulting in failure of the restoration [5]. The thermal expansion of glass depends largely on prior heat treatment and cooling rate [2]. A glass has considerably different thermal expansion coefficients depending on whether the glass has been cooled slowly or rapidly, or has been annealed. The collapse of trapped excess volume is most clearly demonstrated during the conventional dilatometric measurement of the coefficient of linear thermal expansion of a glass [2]. The importance of thermal history (i.e. cooling rate after firing and number of firings) in the thermal expansion of PFM porcelains has been amply demonstrated by several investigators. Fairhurst et al. [4] established that the a -measurement obtained on first heating was not reproducible and therefore was unreliable, and they recommended that a be measured on cooling or following a second heating run. Dorsch [9] demonstrated the dependency of dental porcelain thermal expansion on cooling rate by showing an increase in a -values following protracted slow cooling. He also proved that the a -values of the body porcelains investigated increased with increasing numbers of firing cycles. These effects may have been the result of precipitation of additional leucite during the slow cooling [10] or multiple firing [11]. Fairhurst et al. [1] and Twiggs et al. [12] were able to show that the glass transition ranges of dental porcelains shifted to higher temperatures with increasing heating and cooling rates, thus possibly having a significant deleterious effect on the thermal expansion coefficient match of porcelain with alloy. The leucite content of feldspathic dental porcelains is probably the most significant factor influencing their thermal expansion [13,14]. Therefore, precipitation of additional leucite during the slow heating or cooling of dental porcelain specimens during the thermal expansion measurement itself may result in an a -measurement that is higher than the true value of a (i.e. the value of a without the effect of any confounding
factors on it). Under the thermal expansion measurement conditions used in the present work, however, the porcelains studied appear not to be subject to leucite precipitation during thermal expansion measurement [15]. Conventional dilatometry is an important tool in PFM research and has been widely used to study the thermal compatibility of porcelains with various alloys and to monitor factors that induce changes in their thermal expansion. Commercially available dilatometers are limited to controlled heating rates of less than 308C/min and to temperatures below the glass transition range of the porcelains undergoing thermal expansion measurement. Therefore, the coefficients of thermal expansion of different commercial dental porcelains have only been characterized at low heating and cooling rates below their glass transition range [1,4,9,14]. It would be advantageous to measure the thermal expansion coefficients of dental porcelains at higher heating rates for two reasons. The first reason is that at high heating rates there is insufficient time for additional amounts of leucite to crystallize [10], although at temperatures below about 7008C, the crystallization rate of leucite in dental porcelains is very low [16]. The second reason for the use of higher heating rates in thermal expansion measurement is that there is not sufficient time for the relaxation of trapped excess volume to occur. As discussed above, if the thermal expansion of a PFM porcelain is measured using a conventional dilatometer, the porcelain molecules are able to achieve their most dense configuration during the slow temperature rise of the first heating run due to collapse of trapped excess volume [2,3]. Thus, the structural relaxation process interferes with the thermal expansion measurement itself. To avoid this artifact during the thermal expansion measurement, investigators in this laboratory developed an experimental laser dilatometer [17] that is capable of measuring the thermal expansion of porcelains beyond their glass transition temperatures at rapid heating and cooling rates of 6008C/min. The purpose of this study, therefore, was to determine whether the measurement of porcelain thermal expansion at high heating rates could minimize the influence of thermal history effects—particularly structural relaxation—on thermal expansion measurement of dental porcelains.
392
S.S. Khajotia et al. / Dental Materials 15 (1999) 390–396
Fig. 1. Linear coefficients of thermal expansion measured during heating and cooling in a conventional double-pushrod dilatometer and a laser dilatometer for porcelain BIO.
2. Materials and methods Six commercial dental body porcelains and the Component No. 1 frit of the patent by Weinstein et al. [18] were selected for the study (Table 1). The porcelains were subjected to low-rate (conventional) dilatometry and highrate (laser) dilatometry to determine the effect of thermal history on porcelain thermal expansion. Fifty-six thermal expansion specimens were prepared for use in a conventional dilatometer, eight specimens for each of the porcelains listed in Table 1. Each specimen was fabricated by vibration-condensing a slurry of porcelain powder and distilled water into a split brass mold having inside dimensions of 35:6 × 3:4 × 1:6 mm3 : The porcelain bars were fired on non-sticking porcelain firing trays (Dentecon, Inc., Los Angeles, CA, USA) in a computercontrolled vertical-muffle dental porcelain furnace (Sunfire 10, The J.M. Ney Co., Bloomfield, CT, USA) as per manufacturer’s instructions for each brand of porcelain. The dimensions of each specimen were then reduced by grinding to 25:3 × 3:4 × 1:5 mm3 : The bars for each porcelain brand were subjected to three
successive heating and cooling runs at 38C/min between room temperature and slightly above the softening temperature of the porcelain in a horizontal double-pushrod LVDT dilatometer (Model Research II, Theta Industries, Inc., Port Washington, NY, USA). The specimen design used in the experimental laser dilatometer was a hollow, thin-walled cylinder with parallel ends having the necessary dimensions to fit inside an evacuated Vycorw tube in the laser dilatometer’s radiant furnace (Model E4, Research, Inc., Minneapolis, MN, USA). Each specimen was fabricated by vibration-condensing a slurry of porcelain powder and distilled water into a cylindrical mold 13.5 mm in height and 10.5 mm in outer diameter. The porcelain cylinders were fired on non-sticking porcelain firing trays (Dentecon, Inc., Los Angeles, CA, USA) in a computer-controlled vertical-muffle dental porcelain furnace (Sunfire 10, The J.M. Ney Co., Bloomfield, CT, USA) as per manufacturer’s instructions for each brand of porcelain. The height of each specimen was then reduced by grinding to 14.0 mm, to obtain cylinders with parallel ends (i.d. 8.0 mm; o.d. 10.0 mm). Each specimen was coated with a high-emissivity, silicone-based spray paint (Denplex 12008, “X” Laboratories, Inc., Wheeling, IL, USA) to permit efficient heating by the infrared furnace [19]. The specimen was then placed in a burnout oven at 3008C for 30 min to drive off all the volatile components from the spray paint so that the insides of the Vycorw tube in the laser dilatometer would not become clouded during thermal expansion measurement. The cylinders were subjected to three successive heating and cooling runs at 6008C/minute between room temperature and slightly above the softening temperature of the porcelain in a high heating-rate laser dilatometer. Thermal expansion data over the temperature range 25– 5008C were analyzed for all the runs using repeated measures ANOVA (SAS Institute, Cary, NC, USA; NCSS Statistical Software, Kaysville, UT, USA). P-values less than or equal to 0.05 were considered significant.
3. Results
Fig. 2. Linear coefficients of thermal expansion measured during heating and cooling in a conventional double-pushrod dilatometer and a laser dilatometer for porcelain CER.
The repeated measures ANOVA showed that for the lowrate dilatometer measurements, the mean value of the thermal expansion coefficient (a —in units of 10 6/8C) obtained on first heating was significantly lower than the values for succeeding heating and cooling runs in all the porcelains studied P , 0:001: Comparison of the pooled heating data (excluding the first heating run, which exhibited a significantly lower mean a than the second and third heating runs) with the pooled cooling data for each porcelain via repeated measures, ANOVA showed that for all porcelains except VIT and WIL, the cooling runs resulted in a higher measured value of a than the heating runs. Figs. 1–7 show the mean low-rate coefficient of thermal expansion value for individual heating and cooling runs of all the materials
S.S. Khajotia et al. / Dental Materials 15 (1999) 390–396
393
Fig. 3. Linear coefficients of thermal expansion measured during heating and cooling in a conventional double-pushrod dilatometer and a laser dilatometer for porcelain CII.
Fig. 5. Linear coefficients of thermal expansion measured during heating and cooling in a conventional double-pushrod dilatometer and a laser dilatometer for porcelain VIT.
investigated in this study. The 95% confidence intervals of the means for each material are represented in Figs. 1–7 by error bars. Based on the results of the repeated measures ANOVA for the high-rate (laser) dilatometer data, the value of the thermal expansion coefficient (a ) obtained on first heating was not statistically different from values for succeeding heating and cooling runs for all the porcelains studied P . 0:05: Comparison of the pooled heating data with the pooled cooling data for each porcelain via repeated measures ANOVA showed that for all porcelains, the cooling runs resulted in a higher measured value of a than the heating runs. The mean high-rate coefficient of thermal expansion value (10 6/8C) for individual heating and cooling runs of all the materials studied are shown in Fig. 2. The 95% confidence intervals of the mean for each material are represented by error bars in Fig. 2.
Glasses that are quenched in air (i.e. rapidly cooled) from
high firing temperatures undergo minimal re-ordering to form a structure that has excess volume trapped in it [2,3]. This is because of a lack of sufficient time for the most dense configuration to be reached. If the quenched porcelain is reheated at a slow rate, as in a conventional dilatometer measurement, it expands normally until it is close to its glass transition temperature, at which point sufficient thermal energy is available for the molecules to rearrange themselves into a more dense structure. Kingery et al. [2] noted that the collapse of trapped excess volume via structural relaxation is most clearly demonstrated during the conventional dilatometric thermal expansion measurement of a glass. The results of the low-rate thermal expansion measurements in the present investigation substantiate that premise. The coefficient of thermal expansion values measured on first slow heating were significantly lower than the values for succeeding low-rate heat and cool runs in all of the materials investigated, as expected, providing evidence for the occurrence of structural relaxation during the first slow rate thermal expansion measurement. The a values mean ^ standard deviation obtained for the six commercial dental porcelains on first slow heating ranged
Fig. 4. Linear coefficients of thermal expansion measured during heating and cooling in a conventional double-pushrod dilatometer and a laser dilatometer for porcelain CRY.
Fig. 6. Linear coefficients of thermal expansion measured during heating and cooling in a conventional double-pushrod dilatometer and a laser dilatometer for porcelain WIL.
4. Discussion
394
S.S. Khajotia et al. / Dental Materials 15 (1999) 390–396
Fig. 7. Linear coefficients of thermal expansion measured during heating and cooling in a conventional double-pushrod dilatometer and a laser dilatometer for porcelain CNO.
between 11:4 ^ 0:4 × 1026 =8C and 12:5 ^ 0:2 × 1026 =8C: These values compare favorably with values reported for the same brands of porcelain (11:9 × 1026 =8C to 13:7 × 1026 =8C by other investigators [4,9] over the temperature range 25–5008C after a single firing. However, the thermal expansion coefficient 12:6 ^ 0:2 × 1026 =8C for porcelain Ceramco II, measured on the third slow heating run appeared to be slightly lower than the a value of 13:4 ^ 0:1 × 1026 =8C reported by Piche´ et al. [14] for the same porcelain. This small difference could be attributable to instrumentation differences (single pushrod in the Piche´ et al. study vs. double pushrod in the present study) or other factors. The mean thermal expansion coefficient of Component No. 1 [18] was higher than the a -values of the commercial PFM porcelains. This result is consistent with the higher leucite content (45 wt%) of Component No. 1 compared with commercial dental PFM porcelains (20–25 wt% leucite). The importance of thermal history (i.e. cooling rate after firing and number of firings) in the thermal expansion of PFM porcelains is well documented. Fairhurst et al. [4] originally demonstrated that thermal expansion measurements obtained on first heating were unreliable. The results of this study provide convincing evidence that the thermal expansion coefficients measured on first heating are confounded by the structural relaxation of the material during the measurement itself, thereby making them inaccurate. Fairhurst et al. [4] have suggested that the measurement of a should be made on a cooling run or second heating run instead of on the first slow heating. For most of the porcelains studied in the present investigation (all except VIT and WIL), however, the conventional thermal expansion measurements made on cooling were statistically different P , 0:05 from those made on second and third heating. (The first heating run was excluded from the repeated measures ANOVA, because it was already lower owing to the structural relaxation of the glass matrix during the first heating.) All of the laser (high-rate) dilatometry runs exhibited significantly higher a -values during cooling
measurement compared with heating measurement. This behavior is consistent with the formation of a hysteresis loop in the heating and cooling curve of a glassy material, as described by Rekhson [20], wherein the thermal expansion coefficient of a glass is higher on cooling than on heating. This lag in expansion values can be attributed to the lag in achieving dynamic equilibrium conditions in the glass [20]. While the thermal expansion measurements made during cooling in the conventional (low-rate) dilatometer are not equivalent to those made during heating, the second and third heating runs are not statistically different from each other, and the first, second and third cooling runs are not different from each other. This finding is an important discovery with regard to the actual practice of dilatometer measurements on dental porcelains. It demonstrates that the thermal treatment the specimen receives during the thermal expansion run itself can be safely ignored. In the event of a malfunction or error in measurement during the first cooling or second heating run, it would be possible to re-run the same specimen without expecting a significant change in the measured thermal expansion value obtained, provided heating measurements are compared to heating measurements, and cooling measurements are compared to cooling measurements. Measurement of the coefficient of thermal expansion at high-rates has two major advantages, a lack of sufficient time for additional amounts of leucite to crystallize, and insufficient time for the collapse of trapped excess volume to occur via structural relaxation. Since conventional dilatometer measurements are plagued by the occurrence of structural relaxation in the specimen during the measurement itself, investigators at this laboratory developed an experimental laser dilatometer [12,17] that is capable of measuring the thermal expansion of porcelains beyond their glass transition temperatures at rapid heating and cooling rates of 6008C/min. The results of the high-rate thermal expansion measurements made using the laser dilatometer show that the thermal expansion coefficient obtained on first heating was not statistically different from values for succeeding heat and cool runs for all the materials studied. These results are consistent with the hypothesis that, at highrates, there is insufficient time for the relaxation of trapped excess volume to occur. Therefore, high-rate laser dilatometry eliminated structural relaxation, thereby providing a thermal expansion measurement that is free of interference from thermal history effects. For all except the first heating runs, the thermal expansion coefficients determined via the laser dilatometer were somewhat lower than those determined via the conventional dilatometer. For a few of the porcelains (BIO, CII, WIL), this difference was slight, whereas for others (CER, CRY) it was more pronounced. There are several possible reasons for this difference. One possibility is that the different methods of measuring temperature (optical pyrometer vs. thermocouple) and change in length (laser interferometry vs.
S.S. Khajotia et al. / Dental Materials 15 (1999) 390–396
double pushrod LVDT) in the laser dilatometer and conventional dilatometer could have produced systematic differences in the thermal expansion measurements. One known systematic difference between the two systems is the thermal gradient that existed in the laser dilatometry specimens during heating. The infrared heating of the hollow cylindrical specimens was modeled by Finite Element Analysis (FEA) to assess the thermal gradients that might exist during heating [21]. In this FEA analysis, the ends of the cylinders—particularly the end resting on the base mirror—were found to be somewhat cooler than the center portion throughout the laser dilatometer heating run. Having a portion of the specimen at a lower temperature than the bulk would cause the measured thermal expansion to be lower than the true value for the material. It is unlikely that this or similar systematic differences would be the only reason for the discrepancy, however, because the differences were not uniform across all porcelains—the laser and conventional measurements were quite similar for some of the porcelains. Another possible reason for the observed differences between the laser and conventional dilatometry measurements is the inherent difference in expansivity of the lower-density structure of a quenched glass and the higher-density structure of an annealed glass [3]. Since the laser dilatometry specimen is heated and cooled too rapidly for structural relaxation to occur, it is the lower-density (quenched) structure on which expansion is actually measured in the laser dilatometer. In the conventional dilatometer, structural relaxation occurs during the first heating run, so expansion is measured on the higherdensity (annealed) structure. The relevant structure for porcelain–metal thermal compatibility considerations is the lower-density (quenched) structure, because during the rapid cooling of a PFM restoration from the glass transition temperature down to room temperature, the glassy porcelain matrix has insufficient time to rearrange to the most dense configuration. Therefore, it is the thermal expansion coefficient of the lower-density structure that must be compatible with the metal. Thus, if the thermal expansion coefficients of the annealed and quenched structures were indeed different, it would be preferable to measure the expansion coefficient of the quenched (lower-density) structure.
5. Conclusions • Conventional dilatometer measurements exhibited evidence of structural relaxation during the first heating run, as evidenced by the significant difference in thermal expansion coefficients between the first heating and subsequent runs. • Laser dilatometer measurements were not affected by this thermal history effect. • For both conventional and laser dilatometry, thermal expansion coefficients obtained during cooling
395
measurements are slightly higher than those obtained during heating measurements. • Elimination of structural relaxation in dental porcelain can be accomplished by measuring thermal expansion at heating and cooling rates as high as 6008C/min, thereby providing a thermal expansion measurement that is free of interference from thermal history effects.
Acknowledgements The support of this work through NIH/NIDCR Grant No. R01 DE07806, the Dental Materials Section of the Department of Oral Rehabilitation, and the Department of Oral Biology of the Medical College of Georgia is gratefully acknowledged.
References [1] Fairhurst CW, Hashinger DT, Twiggs SW. The effect of thermal history on porcelain expansion behavior. J Dent Res 1989;68:1313– 1315. [2] Kingery WD, Bowen HK, Uhlmann DR. Introduction to ceramics, 2. New York: Wiley, 1976. [3] Touloukian YS, Kirby RK, Taylor RE, Lee TYR. Theory of thermal expansion of solids, Thermophysical properties of matter, 13. New York: Plenum Press, 1978. pp. 3a–16a. [4] Fairhurst CW, Anusavice KJ, Hashinger DT, Ringle RD, Twiggs SW. Thermal expansion of dental alloys and porcelains. J Biomed Mater Res 1980;14:435–446. [5] Mackert Jr JR. Effects of thermally induced changes on porcelain– metal compatibility. In: Preston JD, editor. Perspectives in dental ceramics, Proceedings of the Fourth International Symposium on Ceramics, Chicago: Quintessence, 1988. pp. 53–64. [6] Rekhson SM, Mazurin OV. Stress relaxation in glass and glass-tometal seals. Glass Technol 1977;18:7–14. [7] Rekhson SM. Annealing of glass-to-metal and glass-to-ceramic seals. Part 1. Theory. Glass Technol 1979;20:27–34. [8] Rekhson SM. Annealing of glass-to-metal and glass-to-ceramic seals. Part 2. Experimental. Glass Technol 1979;20:132–143. [9] Dorsch P. Thermal compatibility of materials for porcelain-fused-tometal (PFM) restorations. Ceramic Forum Int/Ber Dt Keram Ges 1982;59:1–5. [10] Mackert Jr JR, Evans AL. Effect of cooling rate on leucite volume fraction in dental porcelains. J Dent Res 1991;70:137–139. [11] Mackert Jr JR, Evans AL. Quantitative X-ray diffraction determination of leucite thermal instability in dental porcelain. J Am Ceram Soc 1991;74:450–453. [12] Twiggs SW, Searle JR, Ringle RD, Fairhurst CW. A rapid heating and cooling rate dilatometer for measuring thermal expansion in dental porcelain. J Dent Res 1989;68:1316–1318. [13] Mackert Jr JR, Butts MB, Fairhurst CW. The effect of the leucite transformation on dental porcelain expansion. Dent Mater 1986;2:32–36. [14] Piche´ PW, O’Brien WJ, Groh CL, Boenke KM. Leucite content of selected dental porcelains. J Biomed Mater Res 1994;28:603–609. [15] Mackert Jr JR, Khajotia SS, Russell CM, Williams AL. Potential interference of leucite crystallization during porcelain thermal expansion measurement. Dent Mater 1996;12:8–12. [16] Mackert Jr JR, Williams AL, Russell CM. Crystallization of leucite in dental porcelains in the temperature range 6508–10008C. J Dent Res 1999;78:127 Abstr. No. 174.
396
S.S. Khajotia et al. / Dental Materials 15 (1999) 390–396
[17] Twiggs SW, Mackert Jr JR, Khajotia SS. High-rate dilatometer. J Dent Res 1997;76:21 Abstr. No. 62. [18] Weinstein M, Katz S, Weinstein AB. Fused porcelain-to-metal teeth. US Patent No. 3,052,982, September 11, 1962. [19] Twiggs SW, Hashinger DT, Morena R, Fairhurst CW. Glass transition temperatures of dental porcelains at high heating rates. J Biomed Mater Res 1986;20:293–300.
[20] Rekhson SM. Thermal stresses, relaxation and hysteresis in glass. J Am Ceram Soc 1993;76:1113–1123. [21] Khajotia SS. The effect of leucite crystallization and thermal history on thermal expansion measurement of dental porcelains (dissertation). Augusta, GA: Medical College of Georgia, 1997.