Use of x-ray photoelectron spectroscopy and cyclic polarization to evaluate the corrosion behavior of six nickel-chromium alloys before and after porcelain-fused-to-metal firing

Use of x-ray photoelectron spectroscopy and cyclic polarization to evaluate the corrosion behavior of six nickel-chromium alloys before and after porcelain-fused-to-metal firing Michael D. Roach, MS,a John T. Wolan, PhD,b Doug E. Parsell, PhD,c and Joel D. Bumgardner, PhDd Mississippi State University, Mississippi State, and University of Mississippi Medical Center, Jackson, Miss. Statement of problem. Nickel-chromium casting alloys rely on a surface oxide layer for corrosion resistance to the oral environment. Porcelain-fused-to-metal (PFM) firing procedures may alter the surface oxides and corrosion properties of these alloys. Changes in alloy corrosion behavior affect metal ion release and therefore local and/or systemic tissue responses. Purpose. The aim of this study was to evaluate changes in alloy surface oxides and electrochemical corrosion properties after PFM firing. Material and methods. The electrochemical corrosion behavior of 6 commercial nickel-chromium alloys was evaluated in the as-cast/polished and PFM fired/repolished states. Surface chemistries of the alloys were analyzed by x-ray photoelectron spectroscopy. Results. Results indicated an increase in corrosion rates after PFM firing and repolishing for alloys containing 14% to 22% Cr and 9% to 17% Mo. This increase in corrosion rates was attributed to a decrease, caused by the PFM and repolishing process, in the Cr and Mo levels in the surface oxides of these alloys. The PFM firing and repolishing process did not alter the corrosion behavior of the alloys containing lower levels of Cr and Mo and/or Be additions in their bulk composition. These alloys exhibited low levels of Cr and Mo surface oxides in both test conditions. Si particles became embedded in the surfaces of the fired alloys during repolishing and may have contributed to the changes in surface oxides and the corrosion behavior of some alloys. Conclusion. The effects of PFM firing and repolishing on Ni-Cr dental casting alloy surface oxides and corrosion properties appear to be alloy dependent. (J Prosthet Dent 2000;84:623-34.)

CLINICAL IMPLICATIONS Ni-Cr dental casting alloys used in porcelain-fused-to-metal restorations rely on a protective Cr-Mo surface oxide for corrosion resistance. The porcelain firing procedures used may reduce the amount of Cr and Mo in the surface oxides and increase the corrosion rate for some alloys. Increased corrosion of these alloys may be of concern because of the potential development of hypersensitivity reactions to Ni, Cr, Be, and other metal ion corrosion products released over the lifetime of the restoration.

Presented at the International Association for Dental Research annual meeting, Vancouver, BC, Canada, March 1999. Supported by a Research Initiation Grant from Mississippi State University. Work was conducted in partial fulfillment of the requirements for the degree of Master of Science in Biological Engineering in the Department of Agricultural and Biological Engineering, University of Mississippi Medical Center (M. R.). aResearch Engineer, Department of Restorative Dentistry and Biomaterials, School of Dentistry, University of Mississippi Medical Center. bAssistant Professor, Chemical Engineering, Mississippi State University. cAssistant Professor, Department of Restorative Dentistry and Biomaterials, School of Dentistry, University of Mississippi Medical Center. dAssociate Professor, Department of Agricultural and Biological Engineering, Mississippi State University. DECEMBER 2000

N

ickel-chromium dental casting alloys were developed as an alternative to gold-based alloys for partial dentures and crowns in part because of the superior properties of Ni-Cr alloys in porcelain-fused-to-metal (PFM) applications. Ni-Cr alloys now account for the majority of prosthetic restorations used clinically.1,2 Their modulus of elasticity is approximately twice that of Au-based alloys, which allows for a decrease in cross-sectional thickness of the casting and more space for the porcelain veneer.1,2 The thermal coefficients of expansion of these alloys are similar to those of porcelain veneers, which prevent cracking during heating and cooling processes.3,4 Melting temperatures of 1400°C to 1450°C are common for these alloys; thus, they can withstand the temperatures used during PFM firing.5 THE JOURNAL OF PROSTHETIC DENTISTRY 623

THE JOURNAL OF PROSTHETIC DENTISTRY

ROACH ET AL

Table I. Comparison of manufacturers’ reported and EDS-determined alloy bulk compositions (wt%) Alloy

Neptune* Litecast§ Rexalloy* Vera Bond II¶ Vera Bond¶ Biobond II#

(MAN) (EDS) (MAN) (EDS) (MAN) (EDS) (MAN) (EDS) (MAN) (EDS) (MAN) (EDS)

Code

Ni

Cr

Mo

Nb

V

Al

Ga

Fe

Be

Other

A

63.0 60.0 68.5 65.5 67.0 64.7 74.5 72.7 76.9 76.3 80.7 80.6

22.0 20.4 15.5 14.7 14.0 12.8 11.5 11.1 12.6 12.4 13.3 14.2

9.0 12.0 14.0 16.9 8.0 9.2 3.5 7.3 5.0 7.0 — 0.2

† 5.7 — 1.0 — 0.7 — 7.4 — 1.1 — 0.6

— — — — — — — — — — 4.0 4.0

† — 1.0 2.0 — 0.6 — 1.5 2.9 2.9 — 0.4

— — — — 7.5 7.0 — — — 0.3 — —

† 2.0 — — — 5.1 — — — — — —

— ‡ — ‡ — ‡ — ‡ 2.0 ‡ 1.8 ‡

Ti, Y

B C D E F

Si, Mn — — 0.5 Co —

EDS = Energy-dispersive spectrometry; MAN = manufacturer. *Jeneric/Pentron Inc, Wallingford, Conn. †Manufacturer indicated presence but did not specify amount. ‡Could not be detected by EDS. §Ivoclar Williams North America Inc, Amherst, N.Y. ¶Aalba Dent Inc, Cordelia, Calif. #Dentsply International Inc, York, Pa.

Previous studies have addressed the effects of high temperatures used during PFM firing on the physical properties of Ni-Cr alloys. Goodall and Lewis6 found that the interdendritic structures became coarse on exposure to heat treatments at temperatures used during porcelain firing. Winkler et al7 and Marinello et al8 demonstrated a decrease in hardness of Ni-Cr alloys as a result of heat treatments. This softening of the alloy was attributed to homogenization in composition. Morris9 found a statistically significant reduction in strength after porcelain firing in a Ni-Cr-Be alloy and a high-strength Ni-Cr alloy without Be. Few studies have addressed the effects of PFM processes on the corrosion and surface properties of these alloys. De Micheli and Riesgo,10 as well as Meyer,11 investigated the effects of heat treatments simulating porcelain firing on the corrosion of Ni-Cr alloys in artificial saliva and reported no detectable changes in the anodic corrosion behavior of the alloys. However, these studies focused on only alloy equilibrium and oxide breakdown potentials. On the other hand, several studies that examined the surfaces of these alloys reported changes in their color and in the elements present.12-15 In particular, Be increased in the surfaces of Be-containing alloys after high-temperature, non-repolishing treatments.12,14 Previous studies have shown that variations in alloy bulk compositions affect the formation and composition of the surface oxide layer of as-cast alloys.16-21 These studies indicated that alloys containing a high Cr content (16%-27%) and a minimum of 6% Mo exhibited a uniform oxide layer that provided corrosion resistance. Conversely, alloys containing lower levels of Cr and Mo and/or Be additions did not exhibit homogeneous sur624

face oxide layers and were susceptible to accelerated corrosion.17,18, 22-24 It also has been suggested that alloys containing Nb may exhibit enhanced corrosion resistance because of the element’s low solubility in physiologic solutions.25 Lewis26 reported minor decreases in bulk Cr levels of a Ni-Cr alloy after 5 repeated castings but concluded that this change was not likely to affect corrosion properties. Although there are few case studies on tissue reactions resulting from the use of Ni-Cr dental alloys, concerns remain about the potential carcinogenicity of and/or hypersensitivity to Ni, Be, and other metal ion corrosion products released over the lifetime of the restoration.17,18,27-32 However, no known cases exist of Ni-Cr dental casting alloys causing cancer. Ni-Cr alloys rely on a surface oxide layer for corrosion resistance to the oral environment. The PFM cycle may affect the composition or distribution of these surface oxides and therefore affect the corrosion properties of these alloys. Such a change in the corrosion properties would affect metal ion release and therefore local and/or systemic tissue responses. In this study, the corrosion properties of 6 Ni-Cr alloys in artificial saliva in the ascast/polished and porcelain fired/repolished (fired) states (as-cast and fired, respectively) were evaluated by cyclic polarization techniques. The results were compared with the surface chemistry of the alloys as determined by x-ray photoelectron spectroscopy (XPS).

MATERIAL AND METHODS Alloy preparation Six commercial Ni-Cr dental casting alloys used in PFM applications were selected to represent a range of available compositions, including high and low Cr and VOLUME 84 NUMBER 6

ROACH ET AL

Mo content as well as Be additions. The bulk compositions for the alloys are listed in Table I. These alloys were flame cast using a 50/50 oxygen-propane gas mixture into 10 × 5-mm cylinders and 20 × 4-mm disks by a private dental laboratory (Oral Tech, Pearl, Miss.). The cast alloys were polished to simulate clinical conditions through wet grinding with a series of SiC papers (180, 300, 600, and 1500 grit), rough polishing with “Green-coarse” rubber wheels, and fine polishing with Fabulustre (GFC, Carlstadt, N.J.) and a soft bristle brush. After polishing, the cylinders and disks were cleaned ultrasonically in acetone for 5 minutes and rinsed with distilled water before corrosion tests and surface analyses. After testing, alloys were repolished to a bright and shiny surface as previously outlined. The alloys then were subjected to a general porcelain firing cycle. This cycle included degassing at 1010°C under vacuum holding for 5 minutes, opaque firing at 980°C under vacuum and cooling in air, body firing at 970°C under vacuum and cooling in air, and glaze firing at 980°C and cooling in air (UMMC Dental Lab, Jackson, Miss.). The alloys then were repolished. Because the manufacturers did not suggest a specific method for repolishing, the procedure used in this study was as follows: (1) wet grinding through a series of SiC paper (300, 600, and 1500 grit); (2) rough polishing with “Green-coarse” rubber wheels; and (3) fine polishing with Fabulustre (GFC) and a soft bristle brush. The samples were ultrasonically cleaned in acetone and rinsed in distilled water before testing.

X-ray photoelectron spectroscopy analysis The chemical composition of the surface layer of the alloys in the as-cast and fired states was determined by using XPS. Briefly, XPS uses X rays to cause the ejection of electrons, known as photoelectrons, from atoms. The photoelectrons are separated according to their energies in an electron analyzer, detected, and recorded. Because the energy of each ejected photoelectron is characteristic of the atom from which it came, the resulting electron energy spectrum may be used to identify the elements present. The chemical state of the atom (for example, oxide, carbide, or chloride) results in small known shifts in the energy of the characteristic photoelectron. These shifts may be used to identify the chemical state of the atoms present. It should be noted that photoelectrons that interact with other electrons and/or atoms before leaving the sample lose their characteristic energies; hence, only photoelectrons from the first few layers of atoms from the surface of the sample escape with their characteristic energies and are collected for analyses. The experiments were conducted with a Physical Electronics Model 1600 surface analysis system (Eden Prairie, Minn.) with Mg K radiation (1253.6 eV) at 300 W. The analysis area on each specimen was a cirDECEMBER 2000

THE JOURNAL OF PROSTHETIC DENTISTRY

cle approximately 800 µm in diameter. XPS spectra were referenced to the peak position of adventitious C. Adventitious C comes from the saturated, hydrocarbonlike species introduced from the laboratory environment and is almost always present on sample surfaces. This peak has a well-known binding energy of 284.8 eV and is used to correct any drift in peak positions in the collected spectra.33 Survey spectra were obtained by using a pass energy of 46.95 eV. High-resolution spectra were acquired by using a pass energy of 23.5 eV. Quantitative analysis was based on an estimate of the near-surface composition, which was obtained from the peak areas and atomic sensitivity factors, assuming that the region was homogeneous.33 Results of this analysis must be considered semiquantitative because elemental distributions (geometrical or matrix effects) can have a large influence on calculated values.34

Corrosion testing Corrosion tests were run for the as-cast and fired samples. An EG&G Model 6310 Electrochemical Impedance Analyzer (EG&G Princeton Applied Research, Princeton, N.J.) was used with a test cell as specified in ASTM G5.35 Triplicate cyclic polarization tests were conducted for each alloy in an artificial saliva solution (1.5 g/L KCL, 9.0 g/L NaCl, 0.5 g/L KSCN, 0.5 g/L NaH2PO4-H2O, 0.9 g/L lactic acid; pH=4.6). This artificial saliva solution was modified from previous artificial saliva solutions.36-38 Within the solution, the chloride levels were increased and the pH decreased to increase the aggressiveness of the electrolyte.39 All alloy test samples were allowed to reach a steady open circuit potential (Ecorr) for a period of 1 hour. Cyclic polarization tests then were initiated at 100 mV versus standard calomel electrode (SCE) below Ecorr and scanned at a rate of 0.1667 mV/s in the positive direction until a threshold current density of 1 × 10–3 A/cm2 was reached.40 The scan then was reversed back to the Ecorr of the alloys. From these curves, the open-circuit potentials (Ecorr) and corrosion rates (icorr) of the alloys were determined with the EG&G software, M352 version 2.05 (EG&G Princeton Applied Research). The oxide layer breakdown potentials of the alloys (Ebr) were estimated graphically. Hysteresis in the corrosion curves was used to characterize alloy susceptibility to pitting and crevice corrosion.40

Statistical analysis Levene’s homogeneity of variance tests were performed on all corrosion parameters. If the variance was homogeneous, corrosion parameters were tested for statistical significance using analysis of variance (ANOVA) at P≤.05. When differences were detected, Duncan’s multiple range comparison tests were used to identify statistical differences between alloy corro625

THE JOURNAL OF PROSTHETIC DENTISTRY

ROACH ET AL

Fig. 1. Representative XPS survey spectra from alloy E (76 Ni, 12.4 Cr, 7 Mo, 3 Al, 1.1 Nb, 2.0 Be). Peaks for alloy’s major components are shown in (A) as-cast and polished state and (B) PFM fired and repolished state.

sion properties within a specific condition. Student t tests were used to determine significant differences between the as-cast and fired states of the individual alloys at P≤.05. If the variance for a corrosion parameter was not homogeneous, a mixed variance analysis was performed to determine differences between the alloys and between the test conditions.

RESULTS SEM/energy-dispersive spectrometry and XPS analyses The alloy manufacturers’ reported compositions and the energy-dispersive spectrometry (EDS) semiquantitative compositions are listed in Table I. Of the as-cast alloy compositions in Table I, alloys A, B, and F were similar to the manufacturers’ reported composition. Alloy C contained significant amounts of Fe in addition to the reported composition. Alloy D contained significant amounts of Nb and larger amounts of Mo than indicated by the manufacturer. Analysis of alloy E also indicated significant amounts of Nb. With the SEM in the backscattered imaging mode, the Be-containing alloys exhibited typical dendritic microstructures with the Ni-Be eutectic phase present in the interdendritic areas. XPS survey spectra were used to determine the elemental components present on the alloy surfaces and 626

to help identify their chemical states. Ni, Cr, and O (and, if present in the alloy composition, Mo, Be, and other elements) were shown to be present on the ascast test alloy surfaces (Fig. 1). For all of the as-cast alloys, the O 1s peak exhibited a broad asymmetric shape indicating the presence of both metal oxides and hydroxides on the alloy surfaces (Fig. 2, A). A complex Ni 2p peak was present on all alloy surfaces with binding energies that correlated with metallic Ni and NiO. A complex asymmetric Cr 2p peak indicating the presence of Cr2O3 as well as metallic Cr also was identified. All alloys containing Mo exhibited a complex Mo 3d peak composed of metallic Mo and MoO3 (Fig. 2, B). A metallic Be peak was present on the Be-containing alloys, although it was somewhat masked by the Ni 3s peak (Fig. 1). Alloys A and D exhibited complex Nb 3d peaks indicative of metallic Nb, NbO, and Nb2O5. The XPS spectra for alloy C contained a Ga feature probably because of an oxide, although the chemical state could not be positively determined. After the alloys were fired and repolished, XPS scans demonstrated the presence of Ni, Cr, and their oxides on the surfaces of the alloys. However, the peaks for Mo, Be, and other metallic elements were reduced or difficult to detect (Fig. 1). The Mo 3d peak was detected on all Mo-containing alloys, though at a VOLUME 84 NUMBER 6

ROACH ET AL

THE JOURNAL OF PROSTHETIC DENTISTRY

Fig. 2. Representative high-resolution XPS spectra for O 1s and Mo 3d peaks on alloys before and after PFM firing. A, Alloy A (60 Ni, 20 Cr, 12 Mo, 6 Nb, 2 Fe) exhibited reduction in metal oxide shoulder on O 1s peak after PFM firing. Alloy E (76 Ni, 12.4 Cr, 7 Mo, 3 Al, 1.1 Nb, 2.0 Be) exhibited similar metallic oxides before and after PFM firing. B, Alloy A exhibited reduction of MoO3 present in Mo 3d peak after PFM firing. Alloy E did not demonstrate this reduction, although less MoO3 was present in as-cast polished state (n = 2).

much lower intensity than in the as-cast state for alloys A, B, and C. Be was not detected on any of the fired, Be-containing alloy surfaces but may have been masked by the Ni 3s peak (Fig. 1). As in the as-cast condition, alloys A and D exhibited Nb 3d peaks correlating to NbO and Nb2O5 in the fired condition. After firing and repolishing, the metallic oxide shoulder of the O 1s peaks for alloys A, B, and C decreased (Fig. 2, A). Little, if any, change in the shape of the metallic oxide shoulder was observed for alloys D, E, and F in the 2 states (Fig. 2, A). The O peaks on all alloys exhibited a slight shift toward the O 1s reference for SiO2 after firing and repolishing. A Si 2p peak indicating the presence of SiO2 also was present on all alloys after porcelain firing (Fig. 1). This Si 2p peak was detected only on the surfaces of alloy F in the as-cast condition. Figure 3 provides a comparison of the relative surface concentrations of elements on alloy surfaces. Alloys A, B, and C exhibited decreases in Ni, Cr, Mo, and O levels in the fired state compared with the ascast state (Fig. 3, A through C). Nb levels also decreased for alloy A in the fired state (Fig. 3, A). Relative levels of Ni, Cr, and O (and, if present in the alloy, Mo) were very similar in the 2 test conditions for alloys D, E, and F (Fig. 3, D through F). All alloys DECEMBER 2000

exhibited an increase in surface Si levels in the fired state compared with the as-cast state.

Corrosion analyses The open-circuit potentials (Ecorr), oxide breakdown potentials (Ebr), and corrosion current densities (icorr), as determined through cyclic polarization (cp) testing, are listed in Table II. Representative curves for alloys A, C, and F are provided in Figure 4 to show the range of corrosion behavior of the as-cast and fired alloys. No statistical differences were found between the Ecorr values of the alloys in the as-cast state (Table II). For the as-cast condition, alloys A, B, C, and D exhibited lower icorr values than alloys E and F (Table II). Alloys A and B demonstrated the statistically highest Ebr values, followed by alloys C and D, and finally, alloy E. Alloy F did not demonstrate a passive region; therefore, no breakdown potential was recorded (Fig. 4, A). Alloys A and B demonstrated little or no hysteresis between the forward and reverse scans (Fig. 4, A). Alloys C, D, E, and F exhibited large hystereses, indicating a greater susceptibility to pitting and crevice corrosion than alloys A and B (Fig. 4, A). In the fired condition, no statistical differences were identified between the Ecorr or icorr values of the alloys 627

THE JOURNAL OF PROSTHETIC DENTISTRY

ROACH ET AL

A

B

C

D

E

F Fig. 3. Comparison of XPS semiquantitative surface and oxide composition data before and after PFM firing for all alloys tested. Adventitious C on surfaces of alloys makes up surface percentage not represented. All alloys exhibited enhancement of Si on surfaces after PFM firing and repolishing. A, B, and C demonstrated reduction in metallic oxide species present after firing. D, E, and F exhibited similar metallic oxides before and after firing (n = 2).

(Table II). Alloys A and B demonstrated the statistically highest Ebr values, followed by alloys C and D, and finally, alloys E and F (Table II). Alloys A and B demonstrated little or no hysteresis between the forward and reverse scans (Fig. 4, B). Alloys C, D, E, and F demonstrated larger hysteresis, indicating a greater susceptibility to pitting and crevice corrosion than alloys A and B (Fig. 4, B). A comparison of the Ecorr values of the individual alloys in the as-cast and fired conditions revealed no statistical differences (Table II). The corrosion rate for the non-Be alloy (B) in the fired state was statistically higher than in the as-cast state (Table II). No signifi628

cant differences were detected between the Ebr potentials of the individual alloys in the 2 conditions, except for alloy F (Table II). After PFM firing, alloy F exhibited a slightly defined Ebr that was not observed in the as-cast condition (Fig. 4, B). No significant changes in the hysteresis behavior between the as-cast and fired conditions of the alloys were noted (Fig. 4).

DISCUSSION Few studies have addressed the effects of high temperatures that were reached during PFM firing on the corrosion and surface properties of Ni-Cr dental alloys. These alloys rely on surface oxides for corrosion VOLUME 84 NUMBER 6

ROACH ET AL

THE JOURNAL OF PROSTHETIC DENTISTRY

Table II. Comparison of corrosion parameters for Ni-Cr alloys in as-cast and fired conditions icorr (µA/cm2)

Ecorr (mV vs SCE) Alloy

A B C D E F

As-cast

–122.3 –162.6 –206.3 –169.0 –183.7 –173.3

± ± ± ± ± ±

76.3A 31.3A 30.0A 22.7A 37.1A 9.2A

Fired

–150.0 –183.0 –199.7 –181.4 –171.7 –156.6

± ± ± ± ± ±

66.4B 34.7B 2.22B 64.9B 17.3B 13.7B

As-cast

0.07 0.06 0.20 0.12 4.92 5.58

± ± ± ± ± ±

0.03C 0.04C# 0.05C 0.05C 2.14D 2.48D

Ebr (mV vs SCE) Fired

0.45 0.54 2.1 2.8 5.53 3.44

± ± ± ± ± ±

0.38E 0.20E# 1.7E 4.4E 3.14E 0.66E

As-cast

592.5 ± 7.9F 595.7 ± 2.1F 78.0 ± 33.9G 47.0 ± 3.6H –47.0 ± 11.8I None

Fired

614.7 633.3 38.0 65.7 –64.3 –118.5

± ± ± ± ± ±

54.4J 19.3J 15.6K 64.4K 7.2L 9.2L

Duncan’s statistical groupings in the as-cast and fired conditions are noted with superscript capital letters. #Values were determined to be statistically different between the as-cast and fired conditions.

resistance in the oral environment. Temperatures reached during firing may alter the composition of the surface oxides on these alloys, which may in turn alter their corrosion behavior and host tissue reactions. The aim of this study was to evaluate and compare the surface chemistry and cyclic polarization corrosion behavior of Ni-Cr alloys in an artificial saliva solution before and after porcelain firing. The electrolyte used in this study was based on an artificial saliva solution developed by Tani and Zucchi36 and modified by Marek and Topfl.37 This artificial saliva solution has been further modified by Pfeiffer and Schwicherath38 to lower the pH from 6.7 to 4.6 through the removal of sodium bicarbonate. In this study, NaCl was added to this electrolyte to simulate a “worst case” electrolyte for testing as recommended by Dr Guersten at Hanover Medical College, Hanover, Germany. Marek39 reported that electrolytes with increased chloride levels are much more aggressive than natural saliva, but their use nonetheless may be justified. Although the test electrolyte used contained a higher chloride content than that of normal artificial saliva, its use was justified because understanding the corrosion behavior of the alloys under worst-case conditions is important. The composition of the alloys, as reported by the manufacturers and as determined in this study, were, for the most part, comparable. Because composition analyses were not conducted on the ingots received from the manufacturers, differences observed in the compositions of the alloys may be due to the loss of elements during the casting process. Lewis26 did report decreases in some of the elements of a Ni-Cr alloy, particularly C, Al, Mo, and Be, after 5 successive remelting/casting procedures; however, only minor losses in Cr levels were observed. He concluded that the decreases in C, Al, Mo, and Be may affect mechanical properties of the alloy, although losses in Cr are not likely to decrease corrosion properties. Lewis’s rationale for evaluating changes in alloy composition after 5 successive remelting/casting procedures was not clear, because manufacturer recommendations DECEMBER 2000

generally indicate that recasting of the alloys is not recommended. Nevertheless, the differences in the reported and determined compositions of the as-cast alloys were considered relatively minor and not attributed to the loss of elements during the casting process. Ni-Cr dental casting alloys rely on a Cr and Mo surface oxide layer for corrosion resistance in the oral environment. The oxides of Ni, Cr, Mo, Nb, and other alloying elements on the surfaces of the alloys in the ascast state are developed because of the rapid O uptake from the atmosphere after polishing.20,21 The lower corrosion rates and higher breakdown potentials of the as-cast, non-Be–containing alloys were attributed to Cr, Mo, and, if present, Nb-rich surface oxides compared with the Be-containing alloys, which exhibited reduced Cr and Mo surface oxide levels. Nb surface oxides may be particularly effective in enhancing corrosion resistance because of their low solubility in physiological solutions.25 Indeed, the corrosion rates of the as-cast, Be-containing alloys were approximately 10 to 100 times that of the as-cast, non-Be–containing alloys. The surfaces of Be-containing alloys were reported to be nonhomogeneous with some areas rich in Cr and Mo oxides and other areas depleted.17 The NiBe eutectic phase in these alloys is responsible for the development of this segregated surface oxide and reduction in Cr and Mo surface oxides, particularly over the eutectic.17 Thus, these data suggest that, in the as-cast state, non-Be–containing alloys with 13% to 20% Cr, 9% to 17% Mo, and/or greater than 5% Nb may form surface oxide layers capable of providing corrosion resistance. Alloys containing Be may not form Cr and Mo surface oxides, regardless of the bulk Cr and Mo content, which provide the best corrosion resistance. These results are in agreement with previous studies.17,18,22,24,27,30 The surfaces of the Ni-Cr alloys after porcelain firing also demonstrated complex oxides composed of the major elements present in the alloys’ bulk. Because PFM firing takes place only under a slight vacuum (720 mm Hg), it is easily assumed that atmospheric O is still present in the oven. Thus, the high temperatures 629

THE JOURNAL OF PROSTHETIC DENTISTRY

ROACH ET AL

A

B Fig. 4. Representative cyclic polarization curves of alloys A (60 Ni, 20 Cr, 12 Mo, 6 Nb, 2 Fe), C (65 Ni, 13 Cr, 9 Mo, 7 Ga, 5 Fe), and F (81 Ni, 13 Cr, 4 V, 2 Be) in artificial saliva (pH=4.6) at room temperature. A, In as-cast polished state and B, after PFM firing and repolishing (n = 3).

reached during porcelain firing also allowed for the formation of metal oxides. Visual color changes and compositional differences of the surfaces of the alloys have been observed as a result of PFM firing.12 In this study, the alloys were repolished after firing, which also 630

allowed for the uptake of atmospheric O by newly exposed surface components. Thus, the surface oxides that developed on these alloys were a consequence of both PFM firing and the repolishing procedures. In general, alloys A, B, and C demonstrated a VOLUME 84 NUMBER 6

ROACH ET AL

THE JOURNAL OF PROSTHETIC DENTISTRY

Fig. 5. Representative SEM photomicrograph of Si particle embedded in surface of alloy A during repolishing after PFM firing. Note crevice surrounding embedded particle.

decrease in the relative amount of Ni, Cr, and Mo oxides present after firing compared with their as-cast conditions (Figs. 2 and 3, A, B, and C). Corrosion rates for these alloys in the fired condition were approximately 10 times that of their as-cast conditions, although only the increase in corrosion rate for alloy B was statistically significant (Table II). There was no difference in the breakdown potentials of these alloys in the 2 test conditions. The firing/repolishing process affected only the relative levels of components of the surfaces of the alloys, not the chemical state of the components. Surface oxides act, in part, as a nonconductive barrier or resistor to electron flow between the bulk alloy elements and the electrolyte.40 Changes in the relative levels of surface components would change the oxide resistance of the alloys and hence their corrosion rates. Changes in the thickness of the surface oxide layers also would affect their corrosion resistance. Unfortunately, the capabilities of the surface analysis instrumentation used in this study did not permit an estimation of the thickness of the oxides of these alloys. However, because the presence of both metallic and oxide constituents in the XPS scans of the alloys in both test conditions suggests similarly thin surface oxide layers, oxide thickness may not be as significant as oxide composition in this case. Thus, the small reductions in the relative Cr and Mo surface oxide levels of the high Cr and Mo alloys (A, B, and C) resulted in increases in their corrosion rates. On the other hand, alloys D, E, and F exhibited little, if any, change in relative amounts of Ni, Cr, Mo, DECEMBER 2000

and Nb oxides between the as-cast and fired conditions (Figs. 2 and 3, D, E, and F). Because the Cr, Mo, and Nb surface oxide levels for these alloys were similar in the 2 test conditions, it is not surprising that their corrosion rates in the 2 conditions were similar. The breakdown potentials of these alloys, except for alloy F, were also similar in the 2 test conditions. A low breakdown potential, similar to the Be-containing alloy E, was just discernable for alloy F in the fired condition. The high corrosion rates and low breakdown potentials of the Be-containing alloys in both test conditions were attributed to the low Cr and Mo levels in their surface oxide layers in both test conditions. Hence, the firing and repolishing procedures used reduced Cr and Mo surface oxide levels of the alloys containing 13% to 20% Cr and 9% to 17% Mo to levels similar to the Be and/or low Cr- and Mocontaining alloys. This reduction in Cr and Mo oxides resulted in an increase in the corrosion rates for the high Cr and Mo alloys to rates similar to the Becontaining and/or low Cr and Mo alloys. Baran13,20 showed through x-ray diffraction and Auger electron spectroscopy that Cr2O3 becomes the dominant metallic oxide formed on these alloys after high-temperature oxidation treatments. In this study, even though Cr2O3 became the dominant metallic oxide present in alloys A, B, and C after PFM firing, it was still at lower relative concentrations than in the ascast condition. This reduction in surface Cr oxide levels may be due, in part, to the presence of Si on the surfaces of the fired alloys, because oxide levels are rel631

THE JOURNAL OF PROSTHETIC DENTISTRY

ative to each other. If this was the case, however, decreases in relative oxide levels would have been observed for all alloys studied. Because the levels of Cr oxides on alloys D, E, and F were very similar in the 2 test conditions (indeed, an increase in Cr oxide levels was observed for alloy E after firing), the presence of Si on the fired alloy surfaces cannot be the sole factor. The temperatures reached during firing (970°C1010°C) are similar to those that allow for the formation of chromium carbides (510°C-817°C) and the sensitization of stainless steels.40 It is possible that the firing process allowed for the formation of chromium carbides in the Ni-Cr alloy microstructures, which in turn would have reduced surface Cr oxide levels. Previous data from Ni-Cr alloys not repolished after firing exhibited a metal carbide shoulder on the leading edge of the C peak.12 Repolishing the alloys may have removed the chromium carbides and reduced surface Cr levels. Changes in Ni-Cr alloy microstructures because of firing processes have been reported but not correlated to changes in alloy surface oxide composition.6-8 In alloys D, E, and F, NiO remained the dominant metallic oxide present even after firing and repolishing. The NiO dominance on these alloys was attributed to lower Cr and higher Ni concentrations in the alloy bulk than in alloys A, B, and C (Table I). The loss of Mo surface oxides on alloys A, B, and C is most likely due to the loss of MoO3. MoO3 is reported to be volatile at temperatures above 700°C. 15,26 Because typical porcelain veneer firing temperatures are greater than 700°C, MoO3 would be vaporized during the firing process. Roach et al12 and Baran13 also reported an absence of MoO3 on Ni-Cr alloys after high-temperature oxidation treatment without repolishing. Repolishing would have removed the oxide developed during the firing protocol, at least partially exposed fresh native alloy elements, and thus allowed for the uptake of O by any Mo remaining on the surfaces of the alloys. However, because most of the surface Mo was volatized and diffusion of Mo would be limited by its large size, it may be hypothesized that the surfaces of the alloys became depleted in Mo. Hence, even on repolishing, these alloys may not be capable of reforming Mo oxide levels at prefired levels. The decrease in the MoO3 portion of the Mo 3d peak was not as evident in alloys D, E, and F. These alloys, however, contain less Mo in the bulk composition and, as a result, form less MoO3, even in the as-cast condition. Therefore, the loss of MoO3 in their surfaces after PFM firing would be less. Roach et al12 and Diaz-Arnold et al14 detected a large BeO peak when Be-containing Ni-Cr alloys were exposed to high-temperature oxidation treatments and not repolished. In this study, however, Be was not detected in the XPS scans after PFM firing and repolishing. Because of its small size, Be was reported to 632

ROACH ET AL

diffuse to the surface and form BeO during high-temperature oxidation treatments.14 The repolishing then removed the fired BeO-containing surface oxides. Because the firing process caused Be to diffuse to the surface, and because Be then was removed during repolishing, it may be hypothesized that the near-surface layers of the alloys became depleted in Be. A decrease in the relative amount of Be in these surfaces would cause an increase in the relative amounts of other surface oxides. Indeed, a slight increase in the relative amount of Cr oxide was observed for alloy E in the fired condition compared with its as-cast condition. Thus, the repolishing procedure may provide some benefit by reducing Be content on the surfaces of these alloys and, hence, the release of Be corrosion products. A large Si 2p peak indicating the presence of surface SiO2 was present on all alloys after firing. Only alloy F exhibited this Si peak in the as-cast condition. Because softening of these alloys was reported to occur after PFM firing processes, SiC particles most likely became embedded in the alloy surfaces from the SiC paper and polishing wheels during repolishing.7,9 To investigate this theory, the alloy surfaces were examined by SEM/EDS. Figure 5 shows an embedded Si particle in the surface of alloy A after firing and repolishing. Such particles were sparsely distributed on the surfaces of the alloy. Therefore, it appears that the reported softening of Ni-Cr alloys after firing may provide an opportunity for SiC particles to become embedded in the alloy surfaces during repolishing. The heat produced during the repolishing procedure may allow for the formation of SiO2 on the surfaces of the embedded Si particles. In addition, Si often is added to the bulk of the Ni-Cr alloy to reduce oxidation during casting.15,26 Only alloys B and F were known to contain trace amounts of Si. The temperatures reached during PFM firing may have caused Si in these alloys to migrate to the surface, where it would rapidly react with O to form SiO2. This may have further contributed to the total SiO2 peak exhibited in the XPS spectra for these alloys. Therefore, both the Si particles and Si in the alloy compositions contributed to the large Si XPS peak observed on the alloys in the fired state. Although alloy manufacturers recommend repolishing the alloys after firing, they do not specify a particular protocol. A clinical dental laboratory, Oral Tech (Pearl, Miss.), reported blasting the surfaces of fired Ni-Cr alloys with alumina powder or glass beads and/or using Si-containing grinding and polishing wheels to remove excess oxides present. Hence, the SiC papers and Si-containing grinding wheels used to repolish the alloys were similar to those used in commercial clinical practices. These results raise questions regarding alloy repolishing procedures that may provide opportunities for Si VOLUME 84 NUMBER 6

ROACH ET AL

particles to become embedded in the alloy surfaces after PFM firing. The presence of Si particles on the surfaces of the fired alloys is very much a concern with respect to the results of this study. Either the surface chemistry and corrosion results are misleading because of the Si factor, or the firing process caused changes in the microstructures of some alloys that resulted in changes in their surface oxides and hence in their corrosion properties. Because changes in surface oxides and corrosion behavior were not consistent for all alloys, it is likely that the Si particles were not the only factor that affected surface oxide composition and corrosion behavior. Although the microstructures of the Ni-Cr alloys were not evaluated in this study, previous reports have indicated changes in their microstructures because of PFM firing processes.6-8 This evidence, along with the changes in alloy surface oxides as observed in this study, provides support for the hypothesis that PFM firing causes changes in alloy microstructures that result in altered surface oxides and corrosion behavior for some Ni-Cr alloys. Additional research using a repolishing technique without SiC and including microstructural analysis is underway to test this hypothesis.

CONCLUSIONS The corrosion rates of Ni-Cr alloys with 14% to 22% Cr and 9% to 17% Mo were increased because of the PFM processes and repolishing procedures used in this study. The increase in corrosion rates corresponded well with the reduced Cr and Mo levels in the surface oxides of the fired alloys. The surface oxide composition and the corrosion behavior of alloys with lower Cr and Mo levels and/or Be additions were not significantly affected by the PFM firing and repolishing process. Therefore, the effects of PFM firing and repolishing on Ni-Cr alloy surface oxides and corrosion properties were dependent on the alloy composition. Softening of the alloys due to PFM firing may provide opportunities for repolishing particles to become embedded in alloy surfaces, to act as initiation sites for accelerated corrosion processes, and to increase Ni-Cr alloy corrosion product release to levels greater than previously suspected. Hence, the potential for the development of hypersensitivity and other tissue reactions to increased metal ion loads remains a concern. We acknowledge Dr Steven D. Gardner’s technical contributions and expertise with a portion of the x-ray photoelectron spectroscopy data. We also thank Dr Miroslav Marek at the Georgia Institute of Technology and Dr E. Douglas Rigney at the University of Alabama at Birmingham for the kind remarks they made and helpful discussions they contributed to while we prepared this manuscript.

DECEMBER 2000

THE JOURNAL OF PROSTHETIC DENTISTRY

REFERENCES 1. Leinfelder KF. An evaluation of casting alloys used for restorative procedures. J Am Dent Assoc 1997;128:37-45. 2. Craig RG. Restorative dental materials. 9th ed. St Louis: Mosby-Year Book Inc; 1993. p. 415-49. 3. Tucillo JJ, Cascone PJ. Biocompatibility of dental materials. Composition and properties of noble and precious metal alloys. In: CRC series in biocompatibility. Boca Raton, FL.: CRC Press; 1982. p. 19-36. 4. Weinstein M, et al. Fused porcelain-to-metal teeth and porcelain covered metal-reinforced teeth. US Patents 3,052,982 and 3,052,983, 1962. 5. Huget EF, Vermilyea SG. Base metal dental and surgical alloys. In: Smith DC, Williams DF, editors. Biocompatibility of dental materials, vol IV. Boca Raton, FL: CRC Press; 1982. p. 37-50. 6. Goodall TG, Lewis AJ. The metallography of heat treatment effects in a nickel-base casting alloy. A preliminary report. Aust Dent J 1979;24:235-7. 7. Winkler S, Morris HF, Monteiro JM. Changes in mechanical properties and microstructure following heat treatment of a nickel-chromium base alloy. J Prosthet Dent 1984;52:821-7. 8. Marinnello CP, Lüthy H, Schärer P. Influence of heat treatment on the surface texture of an etched cast nickel-chromium base alloy: an evaluation by profilometric records. J Prosthet Dent 1986;56:431-5. 9. Morris HF. Veterans Administration Cooperative Studies Project No. 147/242. Part VII: the mechanical properties of metal ceramic alloys as cast and after simulated porcelain firing. J Prosthet Dent 1989;61:160-9. 10. De Micheli SM, Riesgo O. Electrochemical study of corrosion in Ni-Cr dental alloys. Biomaterials 1982;3:209-12. 11. Meyer JM. Corrosion resistance of nickel-chromium dental casting alloys. Corrosion Sci 1977;17:971-82. 12. Roach M, Gardner S, Parsell D, Bumgardner JD. Comparison of nickelchromium dental alloys before and after porcelain firing. Proc Inst Biol Eng 1998;1:B27-B31. 13. Baran GR. Oxide compounds on Ni-Cr alloys. J Dent Res 1984;63:1332-4. 14. Diaz-Arnold AM, Keller JC, Wightman JP, Williams VD. Bond strength and surface characterization of a Ni-Cr-Be alloy. Dent Mater 1996;12:58-63. 15. Gregory DO. Nickel-chromium alloys in casting. Miss Dent Assoc J 1982;38:18-20. 16. Baran GR. The metallurgy of Ni-Cr alloys for fixed prosthodontics. J Prosthet Dent 1983;50:639-50. 17. Bumgardner JD, Lucas LC. Surface analysis of nickel-chromium dental alloys. Dent Mater 1993;9:252-9. 18. Covington JS, McBride MA, Slagle WF, Disney AL. Quantization of nickel and beryllium leakage from base metal casting alloys. J Prosthet Dent 1985;54:127-36. 19. Covington JS, McBride MA, Slagle WF, Disney AL. Beryllium localization in base metal dental casting alloys. J Biomed Mat Res 1985;19:747-50. 20. Baran GR. Auger chemical analysis of oxides on Ni-Cr alloys. J Dent Res 1984;63:76-80. 21. Kear BH, Pettit FS, Fornwalt DE, Lemaire LP. On the transient oxidation of a Ni-15Cr-6Al alloy. Oxid Metals 1971;3:557-69. 22. Meyer JM. The corrosion of dental Ni-Cr alloys: an in vitro evaluation. In: Hildebrand HF, Champy M, editors. Biocompatibility of Co-Cr-Ni alloys. New York: Plenum Press; 1988. p. 305-20. 23. Pourbaix M. Electrochemical corrosion of metallic biomaterials. Biomaterials 1984;5:122-34. 24. Sarkar NK, Greener EH. In vitro corrosion resistance of new dental alloys. Biomat Med Devices Artif Org 1973;1:121-9. 25. Kovacs P, Davidson JA. Chemical and electrochemical aspects of the biocompatibility of titanium and its alloys. In: Brown SQ, Lemons JE, editors. Medical applications of titanium and its alloys: the material and biological issues. Ann Arbor: ASTM STP 1272; 1996. p. 163-78. 26. Lewis AJ. Changes in composition of a nickel-base partial denture casting alloy upon fusion and casting. Aust Dent J 1975;20:14-8. 27. Geis-Gerstorfer J, Pässler K. Studies on the influence of Be content on the corrosion release and mechanical properties of Ni-25Cr-10Mo alloys. Dent Mater 1993;9:177-81. 28. Haberman AL, Pratt M, Storrs FJ. Contact dermatitis from beryllium in dental alloys. Contact Dermatitis 1993;28:157-62. 29. NIOSH. Pocket guide to chemical hazards. NIOSH Publications. Washington, DC: US Government Printing Office; 1994.

633

THE JOURNAL OF PROSTHETIC DENTISTRY

30. Pan J, Geis-Gerstorfer J, Thierry D, Leygraf C. Electrochemical studies of the influence of beryllium on the corrosion resistance of Ni-25Cr-10Mo cast alloys for dental applications. J Electrochem Soc 1995;142:1454-8. 31. Wiltshire WA. Nickel and cobalt based alloys for resin-bonded prostheses. Quintessence Dent Technol Yearbook 1989;13:153-60. 32. Vilaplana J, Romaguera C, Grimalt F. Occupational and non-occupational allergic contact dermatitis from beryllium. Contact Dermatitis 1992;26:295-8. 33. Moulder F, Stickel WF, Sobol PE, Bomben KD. Handbook of x-ray photoelectron spectroscopy. Eden Prarie, MN: Perkin-Elmer Corp; 1992. 34. Hoflund, GB. Spectroscopic techniques: x-ray photoelectron spectroscopy, auger electron spectroscopy, ion scattering spectroscopy. In: Rivire JC, Myhra S, editors. Handbook of surface and interface analysis: methods in problem solving. New York: Marcel Dekker; 1998. p. 57. 35. ASTM G5-94. Standard reference method for making potentiostatic and potentiodynamic anodic polarization measurements. Annual book of ASTM standards, wear and erosion; metal corrosion. Vol 3.02. West Conshohocker, Pa.: American Society for Testing and Materials; 1997. p. 54-64. 36. Tani G, Zucchi F. Electrochemical measurement of the resistance to corrosion of some commonly used metals for dental prosthesis. [Italian] Minerva Stomatol 1967;16:710-3. 37. Marek M, Topfl E. Electrolytes for corrosion testing of dental alloys [abstract]. J Dent Res 1986;65:301.

ROACH ET AL

38. Pfeiffer P, Schwicherath H. Nickel release of dental alloys as a function of the pH value of the corrosive solution. [German] Dtsch Zahnarztl Z 1991;46:753-6. 39. Marek M. The corrosion of dental materials. In: Scully JC, editor. Treatise on materials science and technology. Corrosion: aqueous processes and passive films. Vol. 23. London: London Academic Press; 1983. p. 331-93. 40. Fontana MG. Corrosion engineering. 3rd ed. New York: McGraw-Hill Books; 1986. p. 39-152, 482-502. Reprint requests to: DR JOEL D. BUMGARDNER ABE, BOX 9632 MISSISSIPPI STATE UNIVERSITY MISSISSIPPI STATE, MS 39762 FAX: (662)325-8536 E-MAIL: [email protected] Copyright © 2000 by The Editorial Council of The Journal of Prosthetic Dentistry. 0022-3913/2000/$12.00 + 0. 10/1/111496

doi:10.1067/mpr.2000.111496

Bound volumes available to subscribers Bound volumes of The Journal of Prosthetic Dentistry are available to subscribers (only) for the 2000 issues from the publisher at a cost of $92.00 ($106.00 international) for Vol. 83 (January-June) and Vol. 84 (July-December). Shipping charges are included. Each bound volume contains a subject and author index, and all advertising is removed. Copies are shipped within 30 days after publication of the last issue in the volume. The binding is durable buckram with the journal name, volume number, and year stamped in gold on the spine. Volumes 81 and 82 are also available. Payment must accompany all orders. Contact Mosby, Subscription Customer Service, 6277 Sea Harbor Dr, Orlando, FL 32887, or call 800-654-2452 or 407-345-4000. Subscriptions must be in force to qualify. Bound volumes are not available in place of a regular Journal subscription.

634

VOLUME 84 NUMBER 6