Corrosion and biocompatibility testing of palladium alloy castings

dental materials Dental Materials 17 (2001) 7±13

www.elsevier.com/locate/dental

Corrosion and biocompatibility testing of palladium alloy castings M. Syverud*, J.E. Dahl, H. Herù 1, E. Morisbak NIOM, Scandinavian Institute of Dental Materials, P.O. Box 70, N-1305 Haslum, Norway Received 17 February 1999; revised 18 February 2000; accepted 22 February 2000

Abstract Objectives: The biocompatibility of palladium±copper alloys has been questioned in the literature. The intention of the present work was to study: (a) the release of ions in an immersion test from a copper-containing alloy, Option w (79% Pd, 10% Cu, 9% Ga, 2% Au), compared with an alloy without Cu, IS85 (82% Pd, 6% Ga, 3.5% Sn, 3.5% In, 2.5% Ag, 2.5% Au); (b) the effect of oxide ®lms produced by preoxidation on corrosion resistance; and (c) the possibility of biologically synergetic effects of ions released. Methods: Specimens of both alloys were cast, rubbed and ultrasonically cleaned. Metallographic specimens were prepared after (a) casting and (b) preoxidizing treatment at approximately 10008C and studied by SEM and EDS. Immersion tests were carried out in a solution of 0.1 mol/l of NaCl and 0.1 mol/l of lactic acid at 378C for 7 days. The alloy specimens were tested in the following three steps: (1) as preoxidized; (2) after subsequent removal of a 0.1 mm thick layer by grinding; and (3) after an additional removal of approximately 0.1 mm by grinding. The test solutions were analyzed by means of ICP to record the amounts of ions that had leached out from the alloy specimens. The biocompatibility was studied by cell culture tests and the HET-CAM method. Test solutions were prepared by dissolving PdCl2 and CuCl2 to appropriate concentrations. Results: The metallographic investigations revealed moderate segregations in interdendritic regions and grain boundaries. After preoxidation in air a zone of oxidation from 25 up to 200 mm thickness for Option and from 5 to 10 mm for IS85 was found. Oxidation was observed along a rim for both alloys and for Option also along interdendritic positions. The oxides were seen as small, dark spots ,1 mm in a metallic matrix. These results indicate that: (1) the oxidation of the copper-containing palladium alloy is far more severe than that of the alloy with no copper; and (2) the elemental release from these oxides is substantially larger than that from the corrosion of the metallic structure. The results of the cell culture testing showed that Cu was most toxic, followed by Cu 21 1 Pd 21 (1:2), based on the determination of the concentration that caused 50% cytotoxicity. The HET-CAM testing showed Cu 21 1 Pd 21 (1:2) to have the highest irritation score. Signi®cance: The copper-containing Pd alloy developed a 0.1 mm thick rim with small oxide particles in a metallic matrix during preoxidation. If this rim is not completely removed, signi®cantly more Cu, Ga and Pd ions have been shown to leach into the test solution than from the alloy structure. No synergetic effect of Cu and Pd ions was observed in cultured cells, while the mixture Pd 21 1 Cu 21 (1:2) was most irritating to mucous membrane as evaluated by the HET-CAM method. q 2001 Academy of Dental Materials. Published by Elsevier Science Ltd. All rights reserved. Keywords: Palladium alloy casting; Corrosion; Biocompatibility

1. Introduction Palladium-based alloys have been used extensively in prosthetic dentistry in recent years because of their low price compared with gold and because of the many attractive properties from a dental point of view. The biocompatibility of palladium±copper±gallium alloys, however, has been questioned in the literature. Complaints from patients presenting diffuse symptoms have been reported by Cai et al. [1] in a review of the biocompatibility of palladium, its alloys and * Corresponding author. Tel.: 147-675-12200; fax: 147-675-91530. 1 Deceased. E-mail address: [email protected] (M. Syverud).

compounds. These alloys released Pd, Cu and Ga ions [2] which have the potential of being toxic in cell culture studies [3±5]. Cytotoxic concentrations of these ions based on LC50values and the MTT-test [6] were in the range 200±340 mM [4]. Concern about possible adverse biological effects has been particularly noticeable in Germany and their Ministry of Health [7] recommended that manufacturers of Pd±Cu alloys should be required to document that no adverse biological effects are associated with this type of alloy. Apparently, Pd±Cu alloys are considered to be worse in this respect than other dental palladium alloys. Such a relationship is expected to arise from a higher release rate of ions or from the synergistic effect between ions leaching out from these alloys increasing the biological response.

0109-5641/01/$20.00 + 0.00 q 2001 Academy of Dental Materials. Published by Elsevier Science Ltd. All rights reserved. PII: S 0109-564 1(00)00033-6

8

M. Syverud et al. / Dental Materials 17 (2001) 7±13

Ag

Au

Ga

In

Pd

Sn

Cu

Three sets with three specimen in each set were tested after the following stages of preparation for both alloys (Table 3): (1) as preoxidized; (2) after subsequent removal of a 0.1 mm thick surface layer by grinding; and (3) after the removal of additional 0.1 mm by grinding.

2.5 ±

2.5 2.0

6.0 9.0

3.5 ±

82.0 79.0

3.5 ±

± 10.0

2.2. Immersion tests

Table 1 Compositions of alloys in wt% Alloys

IS85 Option

Element

The hypothesis of a higher release rate of ions is partly supported by immersion tests in 0.1 mol/l lactic acid and sodium chloride (pH ˆ 2.3 and 4.2) carried out by Pfeiffer and Schwickerath [2]. Their results indicate that alloys with copper, gallium and/or tin are more prone to corrosion than other palladium alloys. This tendency is supported by the recent work of Schwickerath et al. [8] using the same aqueous solution. They found that far more ions leached out from a Pd alloy with 11% Cu and 8% Ga than from a Pd alloy with 6.5% Sn, 6% Ga and 5% Cu. Cai et al. [1] suggested that in Pd±Cu±Ga alloys a lamellar structure consisting alternately of Pd and Pd2Ga is responsible for its high corrosion rate by formation of local galvanic cells. It should be noted that in these studies specimens with a surface ground after the preoxidation treatment were applied, i.e. without an oxide ®lm on the surface. The release of ions from dental palladium alloys with such oxide ®lms has been found to be substantially increased [9]. The aim of the present work was to carry out immersion tests in order to: 1. compare the corrosion resistance of a Pd±Cu±Ga alloy with a palladium alloy without copper; 2. study the effect of oxide ®lms produced by preoxidation on corrosion resistance; 3. study structural reasons for any major observed differences revealed in the corrosion properties; 4. study the possibility of biologically synergetic effects of ions released from the alloys.

Immersion tests of each set with three specimens at a time were carried out for 7 days at 378C in a solution of 0.1 mol/l of lactic acid and 0.1 mol/l of NaCl, largely following the procedure in ISO 1562 (1993), Annex A, but for differences in the grinding procedure prior to testing as indicated above. The pH value of the test solution was 2.3. An additional experiment to study the release at pH ˆ 7 was carried out by using only 0.1 mol/l of NaCl and no lactic acid on both alloys applying one set of three specimens as in the other immersion experiments. These specimens were tested only in the preoxidized condition without grinding. The solutions from the immersion tests were sent in sealed polystyrene containers for analyses by inductively coupled plasma (ICP-OES) (ARL 3580, Thermo-Optek, Ecublens, Switzerland), (carried out by Shef®eld Analytical Services Limited, Shef®eld, UK). Three aliquots of each solution which had been exposed to three specimens were analyzed. The results from control samples, prepared with known amounts of Pd, Ga and Cu ions in aqueous solutions by addition of salts, gave results within the range 0 to 220% of the prepared compositions. 2.3. Metallographic investigations

2. Materials and methods

Metallographic specimens were cut from the sprues as well as from the cast plates for both alloys before and after the preoxidation treatment. The preparation followed standard procedures. The structures were studied by scanning electron microscopy (SEM) (Philips XL 30, Eindhoven, The Netherlands), chie¯y using back-scattered electrons to image the structures of the specimens (BSE) and energy dispersive spectroscopy (EDS) (EDAX DX4-i, Mahwah, NJ, USA). Six EDS spot analyses were carried out on each characteristic type of area.

2.1. Alloy composition and specimen preparation

2.4. Biocompatibility investigations

The compositions of the two conventional palladium alloys applied are given in Table 1. The alloy Option is copper containing, the alloy IS85 is not. The specimens for each of the two alloys with dimensions 10 £ 32 £ 1.5 mm 3 were cast in an induction-heated, centrifugal casting machine (Induktherm HFS-3 Vac, Linn Elektronik, Hirschboch, Germany), rubbed carefully with a blunt instrument taking care not to remove the oxide ®lm and ultrasonically cleaned. All specimens were preoxidized in air (the Option alloy: 10108C for 5 min, the IS85 alloy: 10358C for 5 min) and bench-cooled according to manufacturer's instructions.

Palladium and copper solutions were prepared by dissolving PdCl2 (Aldrich Chemical, Steinheim, Germany) in 1 ml 1 M HCl and diluted with distilled water and by dissolving CuCl2 (Sigma Chemical, St. Louis, USA) in distilled water. The stock solutions were ®ltered through a Millipore ®lter with a pore size of 0.2 mm, and the Pd and Cu concentrations were determined by the ICP-OES method (Shef®eld Analytical Services Limited, Shef®eld, UK). The stock solutions were diluted with distilled water and mixed to desired concentrations (0.1, 0.5 and 1.0 mM) of either Pd 21, Cu 21, Pd 21 1 Cu 21 (3:1) or Pd 21 1 Cu 21 (1:2) which were related

M. Syverud et al. / Dental Materials 17 (2001) 7±13

9

Fig. 1. SEM (BSE) micographs of the alloys as cast. u± ±u 30 mm. (a) Option; (b) IS85.

Fig. 2. SEM (BSE) micrographs of the preoxidized alloys. Cross-sections of sprues. u± ±u 0.2 mm. (a) Option; (b) IS85.

to the concentrations obtained in the immersion studies. Solutions of Ga were not obtained as precipitation occurred in the medium, and thus had to be excluded from the testing.

medium was then aspirated and the formazan dye was extracted in 0.1 ml of 2-propanol, acidi®ed with 0.04 M HCl. The amount of formazan was measured at 570 nm on a multiwell scanning spectrophotometer (Multiskan EX, Labsystems, Helsinki, Finland) and calculated as a percentage of the control (eight parallels for each test solution).

2.5. Cell culture test L-929 Mouse ®broblasts (Bio Whittaker, Verviers, Belgium) were maintained in stationary culture in Eagle's minimum essential medium with Earle's salts, MEM (Gibco BRL, Paisley, UK) supplemented with 5% (v/v) fetal bovine serum (FBS, Gibco BRL, UK) and containing 100 IU/ml penicillin and 100 mg/ml streptomycin (Gibco BRL, UK). The cells were harvested using 0.5 mg/ml trypsin and 0.2 mg/ml EDTA (Gibco BRL, UK) in Earle's balanced salt solution (Gibco BRL, UK), then plated in 96-well tissue culture plates (Costar, Cambridge, USA) at 150 000 cells per ml, 0.1 ml in each well. The cells were then incubated for 24 h at 378C in 5% CO2 before the metal ion solutions were added, 10 ml per well, and further incubated for 24 h. The effect of the metal ion solutions on the cells was assessed by determining MTT±formazan production [6]. Active mitochondria reduce the tetrazolium MTT to its blue, insoluble, formazan form. After incubation with the metal ions, 20 ml MTT (Sigma Chemical St. Louis, USA) dissolved in phosphate buffered saline to a concentration of 5 mg/ml, were added to each well and incubated for 4 h. The

2.6. HET-CAM method The HET-CAM procedure was highly modi®ed from a previously published method [10]. Eight-day-old embryonated hens' eggs were obtained from the National Institute of Public Health, Oslo, Norway. The eggs were incubated at 378C until testing on day nine. The chorioallantoic membrane was accessed by removal of the shell membrane above the air cell, using a dental drill saw blade and forceps. 300 ml of the test solution was applied to the membrane which was then examined using photomacroscope (Wild M400, Wild Heerbrugg, Heerbrugg, Switzerland). Each substance was tested on three eggs and the experiment repeated once. Irritation of the choriollantoic membrane was scored for hemorrhage, coagulation and lysis over a period of 5 min. Photomicrographs were taken prior to application and at 30 s, 2 and 4 min. An average irritation score (IS) was calculated based on the results of both experiments according to Kalweit et al. [10].

10

M. Syverud et al. / Dental Materials 17 (2001) 7±13 Table 2 EDS analyses wt% Alloy

Element Au

IS 85, as cast Bright area Average 2.7 St. dev. 0.2 Dark area Average 2.5 St. dev. 0.4 Option, as cast Bright area Average 2.0 St. dev. 0.4 Dark area Average 1.5 St. dev. 0.5

Fig. 3. SEM (BSE) micrographs of preoxidized Option. (a) u± ±u 30 mm; (b) u± ±u 3 mm.

3. Results 3.1. Metallography No oxidation was observed at a magni®cation of 200 £ after casting and cleaning for either Option or IS85 (Fig. 1a and b). SEM micrographs (BSE) of the structures in crosssections of the sprues of the two alloys after subsequent preoxidation are displayed at a low magni®cation (26 £ ) in Fig. 2a and b. It can be seen that Option is partly oxidized along a rim approximately 100 mm thick. At a higher magni®cation (200 £ ) in Fig. 3a, it is shown that besides an oxidized surface layer 10±15 mm thick, there are

Fig. 4. SEM (BSE) micrograph of preoxidized IS85. u± ±u 1.5 mm.

Pd

Cu

Ga

O

Ag

Sn

In

85.3 1.1

± ±

5.2 1.8

0.9 0.4

1.8 0.2

1.9 0.4

2.1 0.9

83.9 0.8

± ±

8.2 1.0

1.2 0.3

2.1 0.5

1.6 0.5

0.5 0.9

76.9 0.4

11.3 0.6

9.4 0.6

0.4 0.5

± ±

± ±

± ±

74.5 0.8

10.2 1.3

13.0 0.6

0.9 0.6

± ±

± ±

± ±

oxidized regions around pore holes in interdendritic positions. Small, dark spots ,1 mm can be seen in between white areas in all oxidized regions (Fig. 3b). No oxidation was observed on IS85 at the low magni®cation (26 £ ) in Fig. 2b, whereas at a magni®cation of 4000 £ , an oxidized surface zone of approximately 5 mm (Fig. 4) could be observed on this alloy. In contrast to the alloy Option, no interdendritic oxidation was detected. The pattern of numerous small, dark spots in a web of material with a white appearance in the oxidized regions was also observed for this alloy. The SEM micrograph (BSE) of the structure of the alloy Option prior to preoxidation in Fig. 2 displays a bright matrix with gray, interdendritic areas. The results of the EDS analyses of these two regions are given in Table 2. The difference in their average composition is moderate. The gray, interdendritic areas display small, dark laths. The bright matrix is somewhat richer in Pd and has a lower Ga content than the gray areas. After preoxidation the same tendencies and nearly the same values were found for the compositions of these types of areas. The dark, oxidized interdendritic and surface rim areas have about the same composition as the gray areas, but an oxygen content of only 3.6 wt% as measured by EDS. In the other areas investigated the oxygen content was always lower than 1 wt%. The structure of IS85 as cast and prior to preoxidation is shown in Fig. 1b. The matrix consists of light gray grains surrounded by dark gray areas. Inside the grains many small, bright particles can be seen. This structure did not change noticeably with the preoxidation treatment. The dark gray areas are higher in Ga than the bright grains also for this alloy (Table 2). EDS analyses of the oxidized rim of alloy IS85 showed an oxygen content of about 1.5 wt% and an alloy content similar to that of the matrix.

M. Syverud et al. / Dental Materials 17 (2001) 7±13

Sn

In

0.2 0.0

0.7 0.3

4.0 0.6

,0.03 ±

0.06 0.01

0.2 0.06

0.1 mol/l of NaCl and a pH value of 7. It can be seen that the amounts have decreased by factors varying between 10 and 300 considering both alloys and all elements in question. In the cell culture studies, Cu 21 was most toxic based on the determinations of the concentration that caused 50% cytotoxicity (TC value ˆ 0.1 mM), followed by Pd 21 1 Cu 21 (1:2) (TC50 ˆ 0.25 mM), Pd 21 (TC50 ˆ 0.3mM) and Pd 21 1 Cu 21 (3:1) (TC50 ˆ 0.35 mM) (Fig. 5). The HETCAM testing showed a dose dependent increase in irritation for Pd 21, Cu 21, Pd 21 1 Cu 21 (3:1) and Pd 21 1 Cu 21 (1:2) (Table 5). For equivalent concentrations, the Pd 21 1 Cu 21 (1:2) had a higher IS value than Pd 21 1 Cu 21 (3:1).

,0.06 ±

0.08 0.01

0.11 0.04

4. Discussion

± ±

± ±

± ±

± ±

± ±

± ±

± ±

± ±

± ±

Table 3 Immersion test results (test solution: 0.1 mol/l NaCl 1 0.1 mol/l lactic acid. Released ions in mg/cm 2 7 days pH ˆ 2.3. Procedures for casting and testing as described in Annex A in ISO 1562. Preoxidized according to manufacturer's instructions) Alloy

Element Au

Cu

Ga

Pd

IS85, as preoxidized Average ,0.02 ± 10.0 0.2 St. dev. ± ± 5.0 0.1 0.1 mm surface layer removed by grinding Average ,0,03 ± 1.0 0.7 St. dev. ± ± 0.4 0.2 0.2 mm surface layer removed by grinding Average ,0.04 ± 0.3 0.4 St. dev. ± ± 0.1 0.3 Option, as preoxidized Average ,0.02 52 101 28 St. dev. ± 8 12 11 0.1 mm surface layer removed by grinding Average ,0.03 3 5 4 St. dev. ± 2 4 1 0.2 mm surface layer removed by grinding Average ,0.04 0.4 0.5 1.4 St. dev. ± 0.2 0.3 0.3

11

Ag

3.2. Corrosion The results of the ICP analyses of the lactic acid test solutions (pH ˆ 2.3) for Option and IS85 are given in Table 3. The values for each element are given in mg/cm 2 per 7 days. It can be seen that: 1. Both alloys with oxide ®lms from preoxidation release considerably more ions than the ground specimens without oxides. 2. The relationship for released Pd, Cu and Ga from preoxidized Option is approximately 1:2:4. 3. Much larger amounts of ions are leached out from Option than from IS 85, especially after preoxidation. A direct comparison of the two alloys is complicated by the fact that the alloying elements differ in species and concentrations. Table 4 shows the amount of released ions from both alloys in the oxidized condition in a solution with only

The Pd content in IS85 is 82 wt%, whereas the bright and the dark areas have been found to have a slightly higher Pd content. A similar deviation in Pd content has been observed for Option, except that the EDS measurements for this alloy were a few wt% lower than the content given by the manufacturer. Provided the information from the manufacturers about the Pd content is correct, this indicates some small errors in the EDS measurements for Pd. The interdendritic segregations are small in both alloys, but nevertheless the small, dark lamellae in Option indicate that the interdendritic regions are partly eutectic. Brantley et al [11] assume the dark lamellae to be Pd2Ga. Based on the Pd±Ga phase diagram [12] the formation of a eutectic during solidi®cation is possible. By X-ray diffraction Mezger [13] and Khabliyev et al. [14] have observed a secondary bcc phase in addition to the fcc matrix. It is not clear, however, whether this bcc phase is a part of the eutectic observed. Furthermore, the alloy Option also contains 10 wt% Cu in addition to 9 wt% Ga. Such a ternary alloy may contain phases not present in the associated binary alloys. After preoxidation both alloys displayed an oxidized rim consisting of a mixture of small (,1 mm) dark spots in a white matrix. Because of their small size, they cannot be analyzed by EDS in the scanning electron microscope with any reasonable accuracy. This oxidation rim was not present after casting. This observation is in agreement with the work of Hautaniemi et al. [15] and Brantley et al. [16]. X-ray

Table 4 Immersion test results (test solution: 0.1 mol/l NaCl. Released ions in mg/cm 2 7 days; pH ˆ 7. Procedures for casting and testing as described in Annex A in ISO 1562. Preoxidized according manufacturer's instructions Alloys

Element Ag

Au

Ga

In

Pd

Sn

Cu ±

IS 85 as preoxidized

Average St. dev.

,0.02

,0.02

0.6 0.3

,0.02

,0.02

,0.02

Option as preoxidized

Average St. dev.

±

,0.02

0.3 0.2

±

,0.02

±

0.4 0.1

12

M. Syverud et al. / Dental Materials 17 (2001) 7±13

Fig. 5. Cytotoxicity evaluated as the amount of formazan produced relative to the controls for Cu 21, Pd 21, Pd 21 1 Cu 21 (1:2), Pd 21 1 Cu 21 (3:1). Bars indicate one standard deviation based on eight parallels for each test solution.

diffraction studies by Hautaniemi et al [15] indicated the existence of a CuGa2O4 phase in an fcc matrix of unoxidized matrix. The presence of an fcc alloy matrix harmonizes quite well with the low, average oxygen content of 3.6 wt% in this rim measured in the present work by EDS. In addition, the Pd±Cu±Ga alloy displayed an oxidation along interdendritic positions, in contrast to the Cu-free alloy. Rapid diffusion along grain boundaries could contribute to this oxidation pattern. The holes observed in the center of the internally oxidized regions may have been formed during metallographic preparation if the material becomes brittle in these regions during the preoxidising treatment. These structural features are likely to have an effect on the corrosion properties. The dissolution rate for the Pd± Cu±Ga alloy in the preoxidized condition was found to be about 100 times larger than after the removal of a 0.2 mm thick surface layer by grinding. The trend in this result is in agreement with recent German observations [9]. The metallographic SEM studies of the cross-sections show that all Table 5 The irritation score (IS) for the test solutions. The HET-CAM test Concentration (mM)

0.1 0.5 1.0

Pd 21

0.0 1.5 4.8

Cu 21

0.8 3.7 3.2

[Pd 21 1 Cu 21] (3:1)

(1:2)

0.0 0.5 4.2

4.5 8.5 10.5

oxidized regions have then been removed. The removal of only 0.1 mm resulted in a slightly higher release rate than 0.2 mm. The same tendency could be observed for the IS85 alloy, although not to the same extent. The reason for this effect of the oxide layer could be that the release of ions from an oxide layer is a dissolution process which does not require oxidation of the metal. Alternatively, if the oxide layer is porous, the area of the surface exposed to the test solution is increased. Furthermore, the release rate was found to be pH dependent, since far fewer ions were detected in a solution of only NaCl (pH ˆ 7) compared with a pH ˆ 2.3 in a solution of NaCl 1 lactic acid. This pH-effect is at least partly explained by the equation: MeO 1 H2 O ! Me21 1 2OH2 where Me is the metal. The observed relationship that more ions were found to leach out from Option than from IS85 can only partly be explained by the compositions. The latter alloy contains more Pd (85%) and less Ga (6%) than Option (79% Pd and 9% Ga) in wt%. Furthermore, the thickness of the oxide ®lm for IS85, however, is less than half of that for Option, and there is no oxidation in the interdendritic areas. The concentration of Cu 21 and Pd 21 causing 50% lethality in the cell cultures concurred earlier ®ndings [8,17]. Since biocompatibility of Pd±Cu alloys have been of concern, a study of possible synergetic effects in biological systems is of particular interest. The results from the cell culture studies were somewhat dif®cult to interpret in that

M. Syverud et al. / Dental Materials 17 (2001) 7±13

mixture Pd 21 1 Cu 21 (1:2) was less toxic than Cu 21 and more toxic than Pd 21, whereas the mixture Pd 21 1 Cu 21 (3:1) was the least toxic test solution. A possible synergetic effect of Cu 21 and Pd 21 in cultured cells could thus not be found. What we observed could be indicative of an antagonistic effect from the ion of the lowest concentration, but the biological rationale for this is uncertain. The mixture Pd 21 1 Cu 21 (1:2) was the most irritating to mucous membrane as evaluated by the HET-CAM method. In this test method, the occurrence of adverse vascular changes in the chorioallantoic membrane of enbryonated chicken eggs after exposure to a test chemical indicates the potential of the chemical to damage mucous membranes [10,18]. The method has been thoroughly evaluated and found to be an acceptable alternative to in vivo irritation tests [19±22]. The results also showed that there was little difference in the potential for irritating mucous membrane among the other solutions tested. The biocompatibility testing showed that the mixture of Pd 21 and Cu 21 in the ratio 1:2 was the most irritating and also elicited cytotoxic effect on cultured cells. A corresponding ratio between Pd 21 and Cu 21 was found in extract from the preoxidized Pd±Cu±Ga alloy. In addition these alloys release Ga-ions which have the potential of being cytotoxic. Acknowledgements The authors express their sincere thanks to Dr Alf Wennberg for valuable discussions. References [1] Cai Z, Chu X, Bradway SD, Papazouglu E, Brantley WA. On the biocompatibility of high-palladium dental alloys. Cells Mater 1995;5:357±68. [2] Pfeiffer P, Schwickerath H. Palladiumionenabgabe von Palladiumlegierungen in MilchsaÈure-/Kochsalz-lùsung. Dtsch ZahnaÈrz Z 1994;49:616±8. [3] Leirskar J. On the mechanisms of cytotoxicity of silver and copper amalgams in a cell culture system. Scand J Dent Res 1974;82:74±81. [4] Wataha JC, Hanks CT, Craig RG. The in vitro effect of metal cations on eukaryotic cell metabolism. J Biomed Mater Res 1991;25:1133± 49. [5] Wataha JC, Nakajima H, Hanks CT, Okabe T. Correlation of cytotoxicity with elemental release from mercury- and gallium-based dental alloys in vitro. Dent Mater 1994;10:298±303. [6] Edmondson JM, Armstrong LS, Martinez AO. A rapid and simple MTT-based spectrophotometer assay for determining drug sensitivity in monolayer cultures. J Tissue Cult Meth 1988;11:15±7.

13

[7] Legierungen in der zahnaÈrztlichen Therapie. Eine Informationsschrift des Bundesgesundheitsamtes. Druckerei Hellmich, K.G., 1993 [8] Schwickerath H, Pfeiffer P, Nachtweyh K. Lùslichkeit von Edelmetallegierungen nach Mehrfachverguss. Dtsch ZahnaÈrz Z 1998;53:215±8. [9] Institut der Deutschen ZahnaÈrzte. VertraÈglichkeit von Dentallegierungen unter besonderer BeruÈcksichtigung ªalternativerº verfahren zur Diagnostic. Cologne: Deutscher ZahnaÈrzte Verlag, 1998 (p. 22). [10] Kalweit S, Besoke R, Gerner I, Spielmann H. A national validation project of alternative methods to the Draize rabbit eye test. Toxicol in Vitro 1990;4:702±6. [11] Brantley WA, Cai Z, Wu Q, Carr AB, Mitchell JC. Room temperature aging of Pd±Cu±Ga dental alloys. Cells Mater 1995;3:261270. [12] Massalski TB, editor. 2nd ed.. ASM binary alloys phase diagrams, vol. 2. USA: ASM International, 1990. pp. 1836±7. [13] Mezger PR. Corrosion behaviour of dental casting alloys. Doctoral dissertation, Katholike Universitet, Nijmegen. Nijmegen, The Netherlands: Druckerij Benda b v., 1989. [14] Khabliyev SZB, Sakhanskaja IN, Cheremnykh VG, Litvinof VS. Phase transformations in a high-duty alloy of palladium and copper and gallium. Phys Met Metall 1979;47:187±9. [15] Hautaniemi JA, Juhanoj JT, Suoninen EJ, Yli-Urpo AUO. Oxidation of four palladium-rich ceramic fusing alloys. Biomaterials 1990;11:62±72. [16] Brantley WA, Cai Z, Carr AB, Mitchell JC. Metallurgical structures of as-cast and heat-treated high-palladium dental alloys. Cells Mater 1993;3:103±14. [17] Schedle A, Samorapoompichit P, Rausch-Fan XH, Franz A, FuÈreder W, Sperr WR, Sperr W, Ellinger A, Slavicek R, Boltz-Nitulescu G, Valent P. Response of L-929 ®broblasts, human gingival ®briblasts, and human tissue mast cells to various metal cations. J Dent Res 1995;74:1513±20. [18] Spielmann H. HET-CAM test. In: O'Hara S, Aterwill CK, editors. Toxicity testing protocols, Methods in molecular biology, vol. 43. Totowa, NJ: Humana Press, 1995. p. 199±204. [19] Spielmann H, Kalweit S, Liebsch M, Wirnsberger T, Gerner I, Bertram-Neis E, Krauser K, Kreiling R, Miltenburger HG, Pape W, Steiling W. Validation study of alternatives to the Draize eye irritation test in Germany: cytotoxicity testing and HET-CAM-test with 136 industrial chemicals. Toxicol in Vitro 1993;7:505±10. [20] Gettings SD, Dipasquale LC, Bagley DM, Casterton PL, Chudkowski M, Curren RD, Demetrulias JL, Feder PI, Galli CL, Gay R, Glaza SM, Hintze KL, Janus J, Kutz PJ, Lordo RA, Marenus KD, Moral J, Muscatiello M, Pape WJW, Renskers KJ, Roddy MT, Rozen MG. The CTFA evaluation of alternatives program: An evaluation of in vitro alternatives to the Draize primary eye irritation test. (Phase II) Oil/water emulsions. Food Chem Toxicol 1994;32:943±76. [21] Gettings SD, Lordo RA, Demetrulias J, Feder PI, Hintze KL. Comparison of low-volume Draize and in vitro eye irritation test data. I. Hyrdoalcoholic formulations. Food Chem Toxicol 1996;34:737±49. [22] Spielmann H, Liebsch M, Kalweit S, Moldenhauer F, Wirnsberger T, HolzhuÈtter H-G, Schneider B, Glaser S, Gerner I, Pape WJW, Kreiling R, Krauser K, Miltenburger H-G, Steiling W, Luepke NP, MuÈller N, Kreuzer H, MuÈrmann P, Spengler J, Bertram-Neis E, Siegemund B, Wiebel FJ. Results of a validation study in Germany on two in vitro alternatives to the Draize eye irritation test, the HET-CAM test and the 3T3 NRU cytotoxicity test. ATLA 1996;24:741±858.