- Email: [email protected]
Effect of soldering on metal-porcelain bond strength in repaired porcelain-fused-to-metal castings
Effect of soldering on metal-porcelain bond strength in repaired porcelainfused-to-metal castings Daniel F. Galindo, DDS,a Carlo Ercoli, DDS,b Gerald N. Graser, DDS, MS,c Ross H. Tallents, DDS,d and Mark E. Moss, DDS, PhDe University of Rochester Eastman Dental Center, Rochester, N.Y. Statement of problem. Chemical bonding plays a major role in the adherence between metal and porcelain. The formation of an oxide layer on solder material has not been described in the literature. It is unknown whether the application of solder negatively affects the bond strength between porcelain and metal. Purpose. This in vitro study assessed the effect of solder on the bond strength between metal and porcelain. Material and methods. Forty 20 × 6 × 0.5 mm patterns were divided into test (20) and control (20) groups. Test samples were perforated and repaired with solder, and 2 layers of opaque and dentin porcelain subsequently were applied on all samples. The samples were subjected to a 3-point flexural test on a screw-driven mechanical testing machine at a crosshead speed of 0.5 mm/min. Failure type (adhesive vs cohesive) was quantified by digitizing photographs of test and control samples. Three samples in each group also were examined with an SEM coupled with an x-ray energy-dispersive spectroscopy apparatus (SEM/EDS). Means and standard deviations of loads at failure, sample thickness, and surface area covered with porcelain were calculated, and data were analyzed with Student t test (P≤.05). Results. The mean fracture load for test samples was significantly greater than for control samples (P=.0038). Test samples also were significantly thicker (mean thickness difference 0.14 mm) (P=.0001). When the data were controlled for thickness by using a multiple linear regression analysis, no significant difference was found (P=.68). Test samples had a greater surface area covered with opaque porcelain (P=.0006) as determined by visual inspection. Conclusion. In this study, soldered and nonsoldered samples did not show any significant difference in porcelain-to-metal bond strength. Visual analysis revealed a significant difference in the amount of porcelain remaining on the fracture surface of the test and control samples; a complete quantitative elemental analysis with SEM/EDS is in progress. (J Prosthet Dent 2001;85:88-94.)
CLINICAL IMPLICATIONS Soldering may negatively affect porcelain-metal bond strength. The formation of an oxide layer on solder material and the mechanical properties of the porcelain-solder bond have not been described. In this study, soldered and nonsoldered samples did not show any significant difference in porcelain-to-metal bond strength.
S
oldering is defined as the joining of metal components by a filler material, or solder, which is fused to each of the parts being joined.1 Technically, soldering is performed at temperatures below 425°C and braFunded by the Greater New York Academy of Prosthodontics’ 1998 Student Grant Program. aAssistant Professor, Department of Prosthodontics and Operative Dentistry, University of Connecticut School of Dental Medicine; former postgraduate prosthodontic student, Division of Prosthodontics, University of Rochester Eastman Dental Center. bAssistant Professor, Division of Prosthodontics. cProfessor and Program Director, Division of Prosthodontics. dProfessor, Divisions of Orthodontics, Temporomandibular Joint Disorders, and Prosthodontics. eAssistant Professor, Division of Oral Sciences. 88 THE JOURNAL OF PROSTHETIC DENTISTRY
zing at temperatures above 425°C. In dentistry, the latter procedure is commonly called soldering.2 In fixed prosthodontics, solder material is used to add proximal contacts, repair casting voids, and connect retainers and/or pontics in fixed partial dentures (FPDs).3 Solders are alloys usually composed of gold, silver, and copper. Small quantities of tin, zinc, indium, and phosphorus are included to modify the fusion temperature and flow qualities.4 The microstructure of soldered joints has been described in several textbooks.2,5 In a properly fused joint, there is a well-defined boundary between the solder and the soldered parts. In porcelain-fused-to-metal (PFM) restorations, VOLUME 85 NUMBER 1
GALINDO ET AL
metal-ceramic bonding is obtained through: (1) a chemical bond between metallic oxides and porcelain, (2) mechanical interlocking, and (3) van der Waals forces.2,5-7 It has been shown that chemical bonding plays a major role in the overall adherence between metal and porcelain. Adherence is thought to be promoted by the continuity of electronic structure across the metal-metal oxide interface and the metal oxideporcelain interface through metallic, ionic, and covalent bonds.8 Reactive oxides form in the most superficial layer of the casting and chemically bond to the porcelain.9 The composition of the oxide layer is specific to each alloy.10 High-palladium alloys (those containing approximately 75 wt% palladium) commonly are used as an alternative to gold-based alloys. The more recently introduced Pd-Ga alloys have beneficially lower hardness compared with previously marketed Pd-Cu-Ga alloys.11 However, Pd-Ga alloys exhibit less porcelain adherence when compared with Pd-Cu-Ga alloys and Au-Pd alloys. Although the porcelain adherence and metalceramic bond strength of Pd-Ga alloys have been investigated,12-15 the formation of an oxide layer in the area of the solder joint, which would allow metalporcelain chemical bonding, has not been described in the literature. Moreover, the tests used to evaluate the interfacial bond strength have known limitations; thus, true metal-porcelain bond strength values have not been obtained.16-20 It is unknown whether the application of solder to a metal casting would negatively affect the overall bond strength between the porcelain and the metal. The purpose of this in vitro study was to evaluate the influence of soldering on metalceramic bond strength.
MATERIAL AND METHODS Forty 20 × 6 × 0.5 mm patterns were cut from sheets of acrylic resin (Temporary Splint Material, Buffalo Dental Mfg Co, Brooklyn, N.Y.) and divided evenly into test and control groups. The test patterns were perforated in the center with a round carbide bur (HP-2, S.S. White Burs, Inc, Lakewood, N.J.) that was mounted on a commercial electric rotary tool and a drill press stand. To ensure a consistent position of the defect, 2 diagonals were drawn on the rectangular pattern, and the hole was made at the intersection of tile lines (Fig. 1). The perforation measured 2 mm in diameter, simulating a void extending all the way through a casting. The patterns in the control group remained intact. The plastic patterns were invested in groups of 10 (5 test and 5 control) in a carbon-free phosphatebonded investment (Hi-Temp Investment, Whip Mix Corp, Louisville, Ky.).21 Samples were cast by the same investigator who used a multiorifice propane-oxygen torch, a centrifugal casting machine, and a Pd-Ga alloy JANUARY 2001
THE JOURNAL OF PROSTHETIC DENTISTRY
Fig. 1. Schematic representation of test sample with simulated void.
Table I. Composition (wt%) of the alloy and solder used in this investigation
Protocol* HFWC*
Au
Pd
Ag
In
Ga
Ru
Li
6 45
75.2 12.5
6.5 41.5
6 1
6 —
<1 —
<1 —
HFWC = High-fusing white ceramic. *Williams-Ivoclar, Amherst, N.Y.
(Protocol, Williams-Ivoclar, Amherst, N.Y.) (Table I). Castings were divested, and metal irregularities removed with a carbide bur. Test samples were grit blasted with 50 µm aluminum oxide particles at a pressure of 75 psi around the perforation to facilitate solder flow. Soldering was performed according to the protocol described by Shillingburg et al.1 Platinum foil, which acts as a matrix over which solder can flow, was adapted on the undersurface of the test sample and affixed to the casting with sticky wax. The sample was placed over a fresh mix of carbon-free phosphate-bonded investment (Hi-Temp Investment) until the investment set. A No. 2 pencil was used to outline the area and limit the solder flow around the perforation. The casting was placed on a tripod and warmed slightly, and a small quantity of flux was placed into the void. A square of high-fusing white ceramic solder (HFWC, Williams-Ivoclar) (Table I) was placed over the hole. The casting was heated until solder flowed in the area. All test samples were soldered by the same investigator with a single-orifice propane-oxygen torch. After soldering, the flash was removed, and the area was inspected with an optical microscope (Meiji Techno Model BM, Tokyo, Japan) at 10× to determine whether the void had been completely filled. Test and control samples were finished with mounted aluminum oxide stones (75 µm grit, Shofu Dental Corp, Menlo Park, Calif.) (Figs. 2 through 4). The patterns then were grit blasted with 50 µm aluminum oxide particles at a pressure of 75 psi. This procedure was standardized with a jig that maintained the 89
THE JOURNAL OF PROSTHETIC DENTISTRY
Fig. 2. SEM photomicrograph of finished control sample before grit blasting. (Original magnification ×157.)
GALINDO ET AL
Fig. 4. Same sample as in Figure 3. Use of backscattered electron imaging allows visualization of solder (right) and alloy (left).
Fig. 5. Schematic representation of porcelain application over test and control samples.
Fig. 3. SEM photomicrograph of solder area in finished test sample before grit blasting. (Original magnification ×157.)
pattern 5 cm from the tip of the grit blaster. The air abrasion was considered complete when no bur marks were visible on the metal surface with magnification loupes (×3.5). Patterns were steam cleaned and oxidized according to the manufacturer’s instructions (Table II). The surface for porcelain application was determined by calculating and averaging the buccal surface area of the maxillary teeth (Fig. 5).22 Two layers of opaque porcelain and 2 layers of dentin porcelain (Vita Omega Metal Ceramics, Bad Säckingen, Germany) were applied for a total porcelain thickness of 1.5 mm. Porcelain thickness was standardized with a custom-made silicone mold for a final thickness range of 1.3 to 1.5 mm. Porcelain was condensed on an electric vibrator (Ray Foster Dental Equipment, Huntington Beach, Calif.) for 5 seconds; during the condensation procedure, excess water was 90
removed with tissue paper. Opaque and dentin porcelain application and glazing were performed according to the manufacturer’s recommendations (Table II). The specimens were subjected to a 3-point flexural test on a screw-driven testing machine (4204, Instron Corp, Canton, Mass.) at a crosshead speed of 0.5 mm/min, and the loads at failure were recorded.15,23 Sample thickness was recorded to the nearest l µm at 3 points in the central area of each sample with a Mitutoyo caliper (Mitutoyo Mfg Co Ltd, Tokyo, Japan) (error ≤ 0.001 mm). Thickness measurements were performed by the same investigator. Intraexaminer reliability was established in a pilot study (r=0.99). The thickness of the platinum foil (25 µm) was subtracted from the reading for test samples. Metal-porcelain failure was classified as cohesive (within porcelain; porcelain remaining on metal sample) or adhesive (at metal-porcelain interface; no opaque porcelain remaining on metal sample). Failure type was determined by digitizing the test and control specimens over graph paper with an IBM compatible computer (Adobe Photoshop 4.0, Adobe Systems, VOLUME 85 NUMBER 1
GALINDO ET AL
THE JOURNAL OF PROSTHETIC DENTISTRY
Table II. Oxidation and firing cycles for opaque and dentin porcelain and glaze bake Cycle
Oxidation† Opaque porcelain‡ Dentin procelain‡ Glaze
Final temperature (C°)
Drying time (minutes at 500°C)*
1010 950 910 925
Heating rate (°C/min)
— — 6.0 —
— 150 68 141
Hold (min)
Vacuum (min)
5.0 1.0 1.0 —
5.0 3.0 6.0 —
*Oven muffle temperature. Metal-porcelain samples were dried in front of the open muffle. †According to the recommendation of the alloy manufacturer (Table I). ‡Vita Omega Metal Ceramics, Bad Säckingen, Germany.
Table III. Means and standard deviations* (in parentheses) of fracture loads, thickness, and surface area covered with opaque porcelain for control and test samples Measurement
Fracture load (N) Sample thickness (mm) Number of opaqued squares Percentage cohesive failure
Test
33.62 0.60 10.50 30.00
(12.85) (0.10) (8.14) (23.26)
Control
23.85 0.46 2.70 7.71
(4.49) (0.04) (3.48) (9.95)
P
.0038 .0001 .0006 .0006
*Means and standard deviations computed to 2 decimal places for consistency in the statistical analysis.
Inc, San Jose, Calif.) and scanner (UMAX Astra 2400, 600 × 1200 dpi, UMAX Technologies Inc, Fremont, Calif.). All images were equally magnified (×10.5) on the computer screen and examined by the unaided eyes of 2 investigators. Each graph paper square (1 × 1 mm) was considered “opaqued” (cohesive failure) when at least half of the square was observed to be filled with opaque porcelain. Two investigators examined all metal samples. When in disagreement, the investigators examined the specimens together and came to a consensus. The number of “opaqued” squares was counted and expressed as a percentage of the total surface available for porcelain application. Three samples in each group also were examined with an SEM (Model S/240, Leo Microscopy Inc, Thornwood, N.Y.) coupled with an x-ray energy-dispersive spectroscopy (EDS) apparatus (Model QX 2000, Oxford Instruments Inc, Concord, Mass.). All images were obtained in both secondary and backscattered electron emission modes to visualize the boundaries between solder and alloy. Means and standard deviation of loads at failure, sample thickness, and surface area covered with porcelain were calculated, and the data were analyzed with the Student t test (P≤.05). In a 3-point flexural test, metal thickness can affect the force required to bend the metal sample and cause porcelain failure. Therefore, failure loads in test and control groups were controlled for thickness by using a multiple linear regression analysis.24 Specifically, thickness was used as a covariant in the assessment of the relationship between porcelain failure and test/control group status. This statistical model estimated the independent impact of each facJANUARY 2001
tor (thickness and group status) while accounting for the other.
RESULTS Mean values and standard deviations for fracture load, sample thickness, number of opaqued squares, and percentage of surface area covered with opaque porcelain are presented in Table III. For all specimens tested, except 1, failure was characterized by a progressive separation of the entire porcelain layer from the metal sample. This separation always started at 1 end of the sample and propagated toward the center as the applied force increased. This pattern of fracture initiation and propagation was expected because of the stress concentration that exists at porcelain termination sites.20 The mean fracture load of the test samples was greater than that of the control samples (P=.0038). Test samples were thicker (0.14 mm difference) than control samples (P=.0001) because of the undetected overflow of solder between the undersurface of the casting and the platinum foil. When the mean fracture loads in both groups were controlled for thickness values, no differences could be determined (P=.68). Most metal samples showed areas of adhesive failure (no opaque porcelain) alternating with areas of cohesive failure (opaque porcelain present) (Figs. 6 and 7). The percentage area covered with opaque porcelain was significantly greater in the test group than in the control group (P=.0006). Examination with SEM revealed distinct patterns of porcelain adherence to both test and control samples. The control samples (Fig. 8) mostly had areas of adhesive failure (no opaque porcelain) alternating with areas of cohesive failure (opaque porcelain present), 91
THE JOURNAL OF PROSTHETIC DENTISTRY
GALINDO ET AL
Fig. 6. Digitized photograph of control sample after porcelain failure.
Fig. 8. Scanning electron microphotograph at center of control sample in Figure 6. Isolated patches of porcelain on metal sample indicate mostly areas of adhesive failure (no opaque porcelain) alternating with areas of cohesive failure (opaque porcelain present). (Original magnification ×39.)
Fig. 7. Digitized photograph of test sample after porcelain failure.
whereas the test samples (Fig. 9) were covered with a layer of porcelain (cohesive failure), especially on the solder-repaired void. EDS analysis confirmed a greater amount of porcelain, represented by the greater presence of silicon, on the test samples compared with the controls (Figs. 10 and 11).
DISCUSSION Efficiency and predictability are important in the production of PFM restorations. Solder materials routinely are used to repair casting defects and to join pontics and retainers in FPD fabrication. It is essential that the addition of solder material not compromise the mechanical properties and longevity of PFM restorations; otherwise, costly remakes may be necessary. There have been no investigations into the effect and influence of soldering on the metal-porcelain bond. Papazoglou et a112 determined that, when compared with gold-palladium and Pd-Cu-Ga highpalladium alloys, Pd-Ga high-palladium alloys (such as Protocol) exhibited a lower fraction of adherent porcelain on the metal sample after loading in biaxial tension to flexural failure. In the same study, the debonded porcelain of the fractured Pd-Ga alloy specimens generally was covered by a dark gray layer, which was assumed to be the metal oxide. The findings of this study agree with those of Papazoglou et al. 12 Visual observation of 92
Fig. 9. Scanning electron microphotograph at center of test sample in Figure 7. Metal sample covered with layer of porcelain indicates cohesive failure (opaque porcelain present). Adherence of porcelain to round void repaired with solder. (Original magnification ×39.)
control specimens showed a dark gray layer covering most of the debonded porcelain. This contrasted with the test group findings, where oxide rarely was present on the debonded porcelain. The metal surface of the test samples had a greater area fraction of adherent porcelain when compared with the control samples (Figs. 6 and 7). Papazoglou et al 13 also showed that metalporcelain failure usually occurs in a mixed mode, with areas of adhesive failure between the metal and the metal oxide and areas of cohesive fracture in the porcelain. The mode of failure observed in VOLUME 85 NUMBER 1
GALINDO ET AL
THE JOURNAL OF PROSTHETIC DENTISTRY
Fig. 10. EDS analysis of fracture surface at center of control sample in Figure 6. Analysis indicates moderate amount of silica.
Fig. 11. EDS analysis of fracture surface at center of test sample in Figure 7. Analysis indicates greater amount of silica when compared with control sample.
this investigation followed the pattern described by these authors. Although examination of the fractured metalporcelain samples by visual inspection, digital imaging, optical microscope, and SEM provided information about the mode of fracture, analysis with SEM/EDS provided elemental composition information (Figs. 10 and 11). This study did not detect any significant difference in metal-porcelain bond strength when a hole 2 mm in diameter was repaired with solder material. However, limits of this in vitro study include the use of a 3-point flexural test and the use of a single-size defect. Several methods are available to test the metal-porcelain bond; they include the shear test, 3-point flexural test, 4-point flexural test, and biaxial bending. All of these tests have known limitations that prevent true metal-porcelain bond strength values from being obtained.16-20 Because the 3-point flexural test is commonly used, it was selected to allow comparison with previous studies.15,23 The size of the defect was arbitrarily determined to be the maximum surface defect that could be soldered without having to remake the casting. Porcelain failure type (cohesive or adhesive) was determined microscopically by examining the percentage of the metal surface covered with residual opaque porcelain. All samples showed different combinations of adhesive and cohesive porcelain failure (Figs. 6 and 7). In some areas of the samples, it was difficult to quantify visually the presence or absence of opaque porcelain due to combined microscopic fractures within the opaque layer and/or porcelain. Test samples, however, showed a statistically greater surface area of residual opaque porcelain covering the metal surface. This finding suggests that the soldered samples had higher, yet not statistically significant,
metal-porcelain bond strengths when compared with the control samples. It is unknown whether larger defects or the use of crown-shaped samples would have influenced the results. Future research with quantitative SEM/EDS analysis12 is recommended to obtain more accurate information about the percentage of cohesive porcelain failure.
JANUARY 2001
CONCLUSIONS This study investigated the influence of solder material on metal-porcelain bond strength. Within the limitations of this study, it was determined that the application of solder material does not decrease the overall metal-porcelain bond strength. Although visual analysis showed a significant difference in residual porcelain between test and control samples, a complete quantitative elemental analysis with SEM/EDS is indicated. We acknowledge the generous support and collaboration of Ivoclar North America, Inc.
REFERENCES 1. Shillingburg HT Jr, Hobo S, Whitsett LD, Jacobi R, Brackett SE. Fundamentals of fixed prosthodontics. 3rd ed. Chicago: Quintessence Books; 1997. p. 509-35. 2. Craig RG. Restorative dental materials. 10th ed. St Louis: Mosby-Year Book Inc; 1997. p. 397-404, 485-99. 3. Rosenstiel SF, Land MF, Fujimoto J. Contemporary fixed prosthodontics. 2nd ed. St Louis: Mosby-Year Book Inc; 1995. p. 562-78. 4. Ryge G. Dental soldering procedures. Dent Clin North Am 1958;2:747-57. 5. Anusavice KJ. Phillips’ science of dental materials. 10th ed. Philadelphia: WB Saunders; 1996. p. 583-630. 6. Gavelis JR, Lim SB, Guckes AD, Morency JD, Sozio RB. A comparison of the bond strength of two ceramometal systems. J Prosthet Dent 1982;48:424-8. 7. Bagby M, Marshall SJ, Marshall GW Jr. Metal ceramic compatibility: a review of the literature. J Prosthet Dent 1990;63:21-5. 8. Anusavice KJ, Horner JA, Fairhurst CW. Adherence controlling elements in ceramic-metal systems. I. Precious alloys. J Dent Res 1977;56:1045-52. 9. McLean JW. The science and art of dental ceramics. Vol. 1. In: The nature
93
THE JOURNAL OF PROSTHETIC DENTISTRY
10.
11. 12. 13.
14.
15. 16. 17. 18. 19.
GALINDO ET AL
of dental ceramics and their clinical use. 15th ed. Chicago: Quintessence Publishing Inc; 1979. p. 55-114. Mackert JR Jr, Ringle RD, Parry EE, Evans AL, Fairhurst CW. The relationship between oxide adherence and porcelain-metal bonding, J Dent Res 1988;67:474-8. Carr AB, Brantley WA. New high-palladium casting alloys: 1. Overview and initial studies. Int J Prosthodont 1991;4:265-75. Papazoglou E, Brantley WA, Carr AB, Johnston WM. Porcelain adherence to high-palladium alloys. J Prosthet Dent 1993;70:386-94. Papazoglou E, Brantley WA, Mitchell JC, Cai Z, Carr AB. New high-palladium casting alloys: studies of the interface with porcelain. Int J Prosthodont 1996;9:315-22. Papazoglou E, Brantley WA, Johnston WM, Carr AB. Effects of dental laboratory processing variables and in vitro testing medium on the porcelain adherence of high-palladium casting alloys. J Prosthet Dent 1998;79:514-9. Lorenzana RE, Chambless LA, Marker VA, Staffanou RS. Bond strengths of high-palladium content alloys. J Prosthet Dent 1990;64:677-80. DeHoff PH, Anusavice KJ, Hathcock PW. An evaluation of the four-point flexural test for metal-ceramic bond strength. J Dent Res 1982;61:1066-9. Zeng K, Odén A, Rowcliffe D. Flexural tests on dental ceramics. Int J Prosthodont 1996;9:434-9. Ban S, Anusavice KJ. Influence of test method on failure stress of brittle dental materials. J Dent Res 1990;69:1791-9. Papazoglou E, Brantley WA. Porcelain adherence vs force to failure for palladium-gallium alloys: a critique of metal-ceramic bond testing. Dent Mater 1998;14:122-9.
20. Anusavice KJ, DeHoff PH, Fairhurst CW. Comparative evaluation of ceramic-metal bond tests using finite element stress analysis. J Dent Res 1980;59:608-13. 21. Herø H, Syverud M. Carbon impurities and properties of some palladium alloys for ceramic veneering. Dent Mater 1985;1:106-10. 22. Wheeler RC. A textbook of dental anatomy and physiology. 4th ed. Philadelphia: WB Saunders; 1965. p. 125-247. 23. Schaffer SP. An approach to determining the bond strength of ceramometal systems. J Prosthet Dent 1982;48:282-4. 24. Kleinbaum DG, Kupper LL, Muller, KE. Applied regression analysis and other multivariate methods. 2nd ed. Boston: PWS-Kent Publishing Co; 1988. p. 102-23. Reprint requests to: DR CARLO ERCOLI EASTMAN DENTAL CENTER UNIVERSITY OF ROCHESTER 625 ELMWOOD AVE ROCHESTER, NY 14424 FAX: (716)244-8772 E-MAIL: [email protected] Copyright © 2001 by The Editorial Council of The Journal of Prosthetic Dentistry. 0022-3913/2001/$35.00 + 0. 10/1/112429 doi:10.1067/mpr.2001.112429
New product news The January and July issues of the Journal carry information regarding new products of interest to prosthodontists. Product information should be sent 1 month prior to ad closing date to: Dr. Glen P. McGivney, Editor, UNC School of Dentistry, 414C Brauer Hall, CB #7450, Chapel Hill, NC 27599-7450. Product information may be accepted in whole or in part at the discretion of the Editor and is subject to editing. A black-and-white glossy photo may be submitted to accompany product information. Information and products reported are based on information provided by the manufacturer. No endorsement is intended or implied by the Editorial Council of The Journal of Prosthetic Dentistry, the editor, or the publisher.
94
VOLUME 85 NUMBER 1
CLINICAL IMPLICATIONS Soldering may negatively affect porcelain-metal bond strength. The formation of an oxide layer on solder material and the mechanical properties of the porcelain-solder bond have not been described. In this study, soldered and nonsoldered samples did not show any significant difference in porcelain-to-metal bond strength.
S
oldering is defined as the joining of metal components by a filler material, or solder, which is fused to each of the parts being joined.1 Technically, soldering is performed at temperatures below 425°C and braFunded by the Greater New York Academy of Prosthodontics’ 1998 Student Grant Program. aAssistant Professor, Department of Prosthodontics and Operative Dentistry, University of Connecticut School of Dental Medicine; former postgraduate prosthodontic student, Division of Prosthodontics, University of Rochester Eastman Dental Center. bAssistant Professor, Division of Prosthodontics. cProfessor and Program Director, Division of Prosthodontics. dProfessor, Divisions of Orthodontics, Temporomandibular Joint Disorders, and Prosthodontics. eAssistant Professor, Division of Oral Sciences. 88 THE JOURNAL OF PROSTHETIC DENTISTRY
zing at temperatures above 425°C. In dentistry, the latter procedure is commonly called soldering.2 In fixed prosthodontics, solder material is used to add proximal contacts, repair casting voids, and connect retainers and/or pontics in fixed partial dentures (FPDs).3 Solders are alloys usually composed of gold, silver, and copper. Small quantities of tin, zinc, indium, and phosphorus are included to modify the fusion temperature and flow qualities.4 The microstructure of soldered joints has been described in several textbooks.2,5 In a properly fused joint, there is a well-defined boundary between the solder and the soldered parts. In porcelain-fused-to-metal (PFM) restorations, VOLUME 85 NUMBER 1
GALINDO ET AL
metal-ceramic bonding is obtained through: (1) a chemical bond between metallic oxides and porcelain, (2) mechanical interlocking, and (3) van der Waals forces.2,5-7 It has been shown that chemical bonding plays a major role in the overall adherence between metal and porcelain. Adherence is thought to be promoted by the continuity of electronic structure across the metal-metal oxide interface and the metal oxideporcelain interface through metallic, ionic, and covalent bonds.8 Reactive oxides form in the most superficial layer of the casting and chemically bond to the porcelain.9 The composition of the oxide layer is specific to each alloy.10 High-palladium alloys (those containing approximately 75 wt% palladium) commonly are used as an alternative to gold-based alloys. The more recently introduced Pd-Ga alloys have beneficially lower hardness compared with previously marketed Pd-Cu-Ga alloys.11 However, Pd-Ga alloys exhibit less porcelain adherence when compared with Pd-Cu-Ga alloys and Au-Pd alloys. Although the porcelain adherence and metalceramic bond strength of Pd-Ga alloys have been investigated,12-15 the formation of an oxide layer in the area of the solder joint, which would allow metalporcelain chemical bonding, has not been described in the literature. Moreover, the tests used to evaluate the interfacial bond strength have known limitations; thus, true metal-porcelain bond strength values have not been obtained.16-20 It is unknown whether the application of solder to a metal casting would negatively affect the overall bond strength between the porcelain and the metal. The purpose of this in vitro study was to evaluate the influence of soldering on metalceramic bond strength.
MATERIAL AND METHODS Forty 20 × 6 × 0.5 mm patterns were cut from sheets of acrylic resin (Temporary Splint Material, Buffalo Dental Mfg Co, Brooklyn, N.Y.) and divided evenly into test and control groups. The test patterns were perforated in the center with a round carbide bur (HP-2, S.S. White Burs, Inc, Lakewood, N.J.) that was mounted on a commercial electric rotary tool and a drill press stand. To ensure a consistent position of the defect, 2 diagonals were drawn on the rectangular pattern, and the hole was made at the intersection of tile lines (Fig. 1). The perforation measured 2 mm in diameter, simulating a void extending all the way through a casting. The patterns in the control group remained intact. The plastic patterns were invested in groups of 10 (5 test and 5 control) in a carbon-free phosphatebonded investment (Hi-Temp Investment, Whip Mix Corp, Louisville, Ky.).21 Samples were cast by the same investigator who used a multiorifice propane-oxygen torch, a centrifugal casting machine, and a Pd-Ga alloy JANUARY 2001
THE JOURNAL OF PROSTHETIC DENTISTRY
Fig. 1. Schematic representation of test sample with simulated void.
Table I. Composition (wt%) of the alloy and solder used in this investigation
Protocol* HFWC*
Au
Pd
Ag
In
Ga
Ru
Li
6 45
75.2 12.5
6.5 41.5
6 1
6 —
<1 —
<1 —
HFWC = High-fusing white ceramic. *Williams-Ivoclar, Amherst, N.Y.
(Protocol, Williams-Ivoclar, Amherst, N.Y.) (Table I). Castings were divested, and metal irregularities removed with a carbide bur. Test samples were grit blasted with 50 µm aluminum oxide particles at a pressure of 75 psi around the perforation to facilitate solder flow. Soldering was performed according to the protocol described by Shillingburg et al.1 Platinum foil, which acts as a matrix over which solder can flow, was adapted on the undersurface of the test sample and affixed to the casting with sticky wax. The sample was placed over a fresh mix of carbon-free phosphate-bonded investment (Hi-Temp Investment) until the investment set. A No. 2 pencil was used to outline the area and limit the solder flow around the perforation. The casting was placed on a tripod and warmed slightly, and a small quantity of flux was placed into the void. A square of high-fusing white ceramic solder (HFWC, Williams-Ivoclar) (Table I) was placed over the hole. The casting was heated until solder flowed in the area. All test samples were soldered by the same investigator with a single-orifice propane-oxygen torch. After soldering, the flash was removed, and the area was inspected with an optical microscope (Meiji Techno Model BM, Tokyo, Japan) at 10× to determine whether the void had been completely filled. Test and control samples were finished with mounted aluminum oxide stones (75 µm grit, Shofu Dental Corp, Menlo Park, Calif.) (Figs. 2 through 4). The patterns then were grit blasted with 50 µm aluminum oxide particles at a pressure of 75 psi. This procedure was standardized with a jig that maintained the 89
THE JOURNAL OF PROSTHETIC DENTISTRY
Fig. 2. SEM photomicrograph of finished control sample before grit blasting. (Original magnification ×157.)
GALINDO ET AL
Fig. 4. Same sample as in Figure 3. Use of backscattered electron imaging allows visualization of solder (right) and alloy (left).
Fig. 5. Schematic representation of porcelain application over test and control samples.
Fig. 3. SEM photomicrograph of solder area in finished test sample before grit blasting. (Original magnification ×157.)
pattern 5 cm from the tip of the grit blaster. The air abrasion was considered complete when no bur marks were visible on the metal surface with magnification loupes (×3.5). Patterns were steam cleaned and oxidized according to the manufacturer’s instructions (Table II). The surface for porcelain application was determined by calculating and averaging the buccal surface area of the maxillary teeth (Fig. 5).22 Two layers of opaque porcelain and 2 layers of dentin porcelain (Vita Omega Metal Ceramics, Bad Säckingen, Germany) were applied for a total porcelain thickness of 1.5 mm. Porcelain thickness was standardized with a custom-made silicone mold for a final thickness range of 1.3 to 1.5 mm. Porcelain was condensed on an electric vibrator (Ray Foster Dental Equipment, Huntington Beach, Calif.) for 5 seconds; during the condensation procedure, excess water was 90
removed with tissue paper. Opaque and dentin porcelain application and glazing were performed according to the manufacturer’s recommendations (Table II). The specimens were subjected to a 3-point flexural test on a screw-driven testing machine (4204, Instron Corp, Canton, Mass.) at a crosshead speed of 0.5 mm/min, and the loads at failure were recorded.15,23 Sample thickness was recorded to the nearest l µm at 3 points in the central area of each sample with a Mitutoyo caliper (Mitutoyo Mfg Co Ltd, Tokyo, Japan) (error ≤ 0.001 mm). Thickness measurements were performed by the same investigator. Intraexaminer reliability was established in a pilot study (r=0.99). The thickness of the platinum foil (25 µm) was subtracted from the reading for test samples. Metal-porcelain failure was classified as cohesive (within porcelain; porcelain remaining on metal sample) or adhesive (at metal-porcelain interface; no opaque porcelain remaining on metal sample). Failure type was determined by digitizing the test and control specimens over graph paper with an IBM compatible computer (Adobe Photoshop 4.0, Adobe Systems, VOLUME 85 NUMBER 1
GALINDO ET AL
THE JOURNAL OF PROSTHETIC DENTISTRY
Table II. Oxidation and firing cycles for opaque and dentin porcelain and glaze bake Cycle
Oxidation† Opaque porcelain‡ Dentin procelain‡ Glaze
Final temperature (C°)
Drying time (minutes at 500°C)*
1010 950 910 925
Heating rate (°C/min)
— — 6.0 —
— 150 68 141
Hold (min)
Vacuum (min)
5.0 1.0 1.0 —
5.0 3.0 6.0 —
*Oven muffle temperature. Metal-porcelain samples were dried in front of the open muffle. †According to the recommendation of the alloy manufacturer (Table I). ‡Vita Omega Metal Ceramics, Bad Säckingen, Germany.
Table III. Means and standard deviations* (in parentheses) of fracture loads, thickness, and surface area covered with opaque porcelain for control and test samples Measurement
Fracture load (N) Sample thickness (mm) Number of opaqued squares Percentage cohesive failure
Test
33.62 0.60 10.50 30.00
(12.85) (0.10) (8.14) (23.26)
Control
23.85 0.46 2.70 7.71
(4.49) (0.04) (3.48) (9.95)
P
.0038 .0001 .0006 .0006
*Means and standard deviations computed to 2 decimal places for consistency in the statistical analysis.
Inc, San Jose, Calif.) and scanner (UMAX Astra 2400, 600 × 1200 dpi, UMAX Technologies Inc, Fremont, Calif.). All images were equally magnified (×10.5) on the computer screen and examined by the unaided eyes of 2 investigators. Each graph paper square (1 × 1 mm) was considered “opaqued” (cohesive failure) when at least half of the square was observed to be filled with opaque porcelain. Two investigators examined all metal samples. When in disagreement, the investigators examined the specimens together and came to a consensus. The number of “opaqued” squares was counted and expressed as a percentage of the total surface available for porcelain application. Three samples in each group also were examined with an SEM (Model S/240, Leo Microscopy Inc, Thornwood, N.Y.) coupled with an x-ray energy-dispersive spectroscopy (EDS) apparatus (Model QX 2000, Oxford Instruments Inc, Concord, Mass.). All images were obtained in both secondary and backscattered electron emission modes to visualize the boundaries between solder and alloy. Means and standard deviation of loads at failure, sample thickness, and surface area covered with porcelain were calculated, and the data were analyzed with the Student t test (P≤.05). In a 3-point flexural test, metal thickness can affect the force required to bend the metal sample and cause porcelain failure. Therefore, failure loads in test and control groups were controlled for thickness by using a multiple linear regression analysis.24 Specifically, thickness was used as a covariant in the assessment of the relationship between porcelain failure and test/control group status. This statistical model estimated the independent impact of each facJANUARY 2001
tor (thickness and group status) while accounting for the other.
RESULTS Mean values and standard deviations for fracture load, sample thickness, number of opaqued squares, and percentage of surface area covered with opaque porcelain are presented in Table III. For all specimens tested, except 1, failure was characterized by a progressive separation of the entire porcelain layer from the metal sample. This separation always started at 1 end of the sample and propagated toward the center as the applied force increased. This pattern of fracture initiation and propagation was expected because of the stress concentration that exists at porcelain termination sites.20 The mean fracture load of the test samples was greater than that of the control samples (P=.0038). Test samples were thicker (0.14 mm difference) than control samples (P=.0001) because of the undetected overflow of solder between the undersurface of the casting and the platinum foil. When the mean fracture loads in both groups were controlled for thickness values, no differences could be determined (P=.68). Most metal samples showed areas of adhesive failure (no opaque porcelain) alternating with areas of cohesive failure (opaque porcelain present) (Figs. 6 and 7). The percentage area covered with opaque porcelain was significantly greater in the test group than in the control group (P=.0006). Examination with SEM revealed distinct patterns of porcelain adherence to both test and control samples. The control samples (Fig. 8) mostly had areas of adhesive failure (no opaque porcelain) alternating with areas of cohesive failure (opaque porcelain present), 91
THE JOURNAL OF PROSTHETIC DENTISTRY
GALINDO ET AL
Fig. 6. Digitized photograph of control sample after porcelain failure.
Fig. 8. Scanning electron microphotograph at center of control sample in Figure 6. Isolated patches of porcelain on metal sample indicate mostly areas of adhesive failure (no opaque porcelain) alternating with areas of cohesive failure (opaque porcelain present). (Original magnification ×39.)
Fig. 7. Digitized photograph of test sample after porcelain failure.
whereas the test samples (Fig. 9) were covered with a layer of porcelain (cohesive failure), especially on the solder-repaired void. EDS analysis confirmed a greater amount of porcelain, represented by the greater presence of silicon, on the test samples compared with the controls (Figs. 10 and 11).
DISCUSSION Efficiency and predictability are important in the production of PFM restorations. Solder materials routinely are used to repair casting defects and to join pontics and retainers in FPD fabrication. It is essential that the addition of solder material not compromise the mechanical properties and longevity of PFM restorations; otherwise, costly remakes may be necessary. There have been no investigations into the effect and influence of soldering on the metal-porcelain bond. Papazoglou et a112 determined that, when compared with gold-palladium and Pd-Cu-Ga highpalladium alloys, Pd-Ga high-palladium alloys (such as Protocol) exhibited a lower fraction of adherent porcelain on the metal sample after loading in biaxial tension to flexural failure. In the same study, the debonded porcelain of the fractured Pd-Ga alloy specimens generally was covered by a dark gray layer, which was assumed to be the metal oxide. The findings of this study agree with those of Papazoglou et al. 12 Visual observation of 92
Fig. 9. Scanning electron microphotograph at center of test sample in Figure 7. Metal sample covered with layer of porcelain indicates cohesive failure (opaque porcelain present). Adherence of porcelain to round void repaired with solder. (Original magnification ×39.)
control specimens showed a dark gray layer covering most of the debonded porcelain. This contrasted with the test group findings, where oxide rarely was present on the debonded porcelain. The metal surface of the test samples had a greater area fraction of adherent porcelain when compared with the control samples (Figs. 6 and 7). Papazoglou et al 13 also showed that metalporcelain failure usually occurs in a mixed mode, with areas of adhesive failure between the metal and the metal oxide and areas of cohesive fracture in the porcelain. The mode of failure observed in VOLUME 85 NUMBER 1
GALINDO ET AL
THE JOURNAL OF PROSTHETIC DENTISTRY
Fig. 10. EDS analysis of fracture surface at center of control sample in Figure 6. Analysis indicates moderate amount of silica.
Fig. 11. EDS analysis of fracture surface at center of test sample in Figure 7. Analysis indicates greater amount of silica when compared with control sample.
this investigation followed the pattern described by these authors. Although examination of the fractured metalporcelain samples by visual inspection, digital imaging, optical microscope, and SEM provided information about the mode of fracture, analysis with SEM/EDS provided elemental composition information (Figs. 10 and 11). This study did not detect any significant difference in metal-porcelain bond strength when a hole 2 mm in diameter was repaired with solder material. However, limits of this in vitro study include the use of a 3-point flexural test and the use of a single-size defect. Several methods are available to test the metal-porcelain bond; they include the shear test, 3-point flexural test, 4-point flexural test, and biaxial bending. All of these tests have known limitations that prevent true metal-porcelain bond strength values from being obtained.16-20 Because the 3-point flexural test is commonly used, it was selected to allow comparison with previous studies.15,23 The size of the defect was arbitrarily determined to be the maximum surface defect that could be soldered without having to remake the casting. Porcelain failure type (cohesive or adhesive) was determined microscopically by examining the percentage of the metal surface covered with residual opaque porcelain. All samples showed different combinations of adhesive and cohesive porcelain failure (Figs. 6 and 7). In some areas of the samples, it was difficult to quantify visually the presence or absence of opaque porcelain due to combined microscopic fractures within the opaque layer and/or porcelain. Test samples, however, showed a statistically greater surface area of residual opaque porcelain covering the metal surface. This finding suggests that the soldered samples had higher, yet not statistically significant,
metal-porcelain bond strengths when compared with the control samples. It is unknown whether larger defects or the use of crown-shaped samples would have influenced the results. Future research with quantitative SEM/EDS analysis12 is recommended to obtain more accurate information about the percentage of cohesive porcelain failure.
JANUARY 2001
CONCLUSIONS This study investigated the influence of solder material on metal-porcelain bond strength. Within the limitations of this study, it was determined that the application of solder material does not decrease the overall metal-porcelain bond strength. Although visual analysis showed a significant difference in residual porcelain between test and control samples, a complete quantitative elemental analysis with SEM/EDS is indicated. We acknowledge the generous support and collaboration of Ivoclar North America, Inc.
REFERENCES 1. Shillingburg HT Jr, Hobo S, Whitsett LD, Jacobi R, Brackett SE. Fundamentals of fixed prosthodontics. 3rd ed. Chicago: Quintessence Books; 1997. p. 509-35. 2. Craig RG. Restorative dental materials. 10th ed. St Louis: Mosby-Year Book Inc; 1997. p. 397-404, 485-99. 3. Rosenstiel SF, Land MF, Fujimoto J. Contemporary fixed prosthodontics. 2nd ed. St Louis: Mosby-Year Book Inc; 1995. p. 562-78. 4. Ryge G. Dental soldering procedures. Dent Clin North Am 1958;2:747-57. 5. Anusavice KJ. Phillips’ science of dental materials. 10th ed. Philadelphia: WB Saunders; 1996. p. 583-630. 6. Gavelis JR, Lim SB, Guckes AD, Morency JD, Sozio RB. A comparison of the bond strength of two ceramometal systems. J Prosthet Dent 1982;48:424-8. 7. Bagby M, Marshall SJ, Marshall GW Jr. Metal ceramic compatibility: a review of the literature. J Prosthet Dent 1990;63:21-5. 8. Anusavice KJ, Horner JA, Fairhurst CW. Adherence controlling elements in ceramic-metal systems. I. Precious alloys. J Dent Res 1977;56:1045-52. 9. McLean JW. The science and art of dental ceramics. Vol. 1. In: The nature
93
THE JOURNAL OF PROSTHETIC DENTISTRY
10.
11. 12. 13.
14.
15. 16. 17. 18. 19.
GALINDO ET AL
of dental ceramics and their clinical use. 15th ed. Chicago: Quintessence Publishing Inc; 1979. p. 55-114. Mackert JR Jr, Ringle RD, Parry EE, Evans AL, Fairhurst CW. The relationship between oxide adherence and porcelain-metal bonding, J Dent Res 1988;67:474-8. Carr AB, Brantley WA. New high-palladium casting alloys: 1. Overview and initial studies. Int J Prosthodont 1991;4:265-75. Papazoglou E, Brantley WA, Carr AB, Johnston WM. Porcelain adherence to high-palladium alloys. J Prosthet Dent 1993;70:386-94. Papazoglou E, Brantley WA, Mitchell JC, Cai Z, Carr AB. New high-palladium casting alloys: studies of the interface with porcelain. Int J Prosthodont 1996;9:315-22. Papazoglou E, Brantley WA, Johnston WM, Carr AB. Effects of dental laboratory processing variables and in vitro testing medium on the porcelain adherence of high-palladium casting alloys. J Prosthet Dent 1998;79:514-9. Lorenzana RE, Chambless LA, Marker VA, Staffanou RS. Bond strengths of high-palladium content alloys. J Prosthet Dent 1990;64:677-80. DeHoff PH, Anusavice KJ, Hathcock PW. An evaluation of the four-point flexural test for metal-ceramic bond strength. J Dent Res 1982;61:1066-9. Zeng K, Odén A, Rowcliffe D. Flexural tests on dental ceramics. Int J Prosthodont 1996;9:434-9. Ban S, Anusavice KJ. Influence of test method on failure stress of brittle dental materials. J Dent Res 1990;69:1791-9. Papazoglou E, Brantley WA. Porcelain adherence vs force to failure for palladium-gallium alloys: a critique of metal-ceramic bond testing. Dent Mater 1998;14:122-9.
20. Anusavice KJ, DeHoff PH, Fairhurst CW. Comparative evaluation of ceramic-metal bond tests using finite element stress analysis. J Dent Res 1980;59:608-13. 21. Herø H, Syverud M. Carbon impurities and properties of some palladium alloys for ceramic veneering. Dent Mater 1985;1:106-10. 22. Wheeler RC. A textbook of dental anatomy and physiology. 4th ed. Philadelphia: WB Saunders; 1965. p. 125-247. 23. Schaffer SP. An approach to determining the bond strength of ceramometal systems. J Prosthet Dent 1982;48:282-4. 24. Kleinbaum DG, Kupper LL, Muller, KE. Applied regression analysis and other multivariate methods. 2nd ed. Boston: PWS-Kent Publishing Co; 1988. p. 102-23. Reprint requests to: DR CARLO ERCOLI EASTMAN DENTAL CENTER UNIVERSITY OF ROCHESTER 625 ELMWOOD AVE ROCHESTER, NY 14424 FAX: (716)244-8772 E-MAIL: [email protected] Copyright © 2001 by The Editorial Council of The Journal of Prosthetic Dentistry. 0022-3913/2001/$35.00 + 0. 10/1/112429 doi:10.1067/mpr.2001.112429
New product news The January and July issues of the Journal carry information regarding new products of interest to prosthodontists. Product information should be sent 1 month prior to ad closing date to: Dr. Glen P. McGivney, Editor, UNC School of Dentistry, 414C Brauer Hall, CB #7450, Chapel Hill, NC 27599-7450. Product information may be accepted in whole or in part at the discretion of the Editor and is subject to editing. A black-and-white glossy photo may be submitted to accompany product information. Information and products reported are based on information provided by the manufacturer. No endorsement is intended or implied by the Editorial Council of The Journal of Prosthetic Dentistry, the editor, or the publisher.
94
VOLUME 85 NUMBER 1