Biomaterials
Volume 24, Issue 25 , November 2003, Pages 4541-4549

Corrosion behavior of cast titanium with reduced surface reaction layer made by a face-coating method

M. Koike^{, , a, b}, Z. Cai^b, H. Fujii^a, M Brezner^b and T. Okabe<sup>b

a</sup> Nagasaki University Graduate School of Biomedical Sciences, Course of Medical and Dental Sciences, Department of Developmental and Reconstructive Medicine, Division of Removable Prosthodontics and Management of Oral Function, 1-7-1 Sakamoto, Nagasaki 852-8588, Japan
^b Department of Biomaterials Science, Baylor College of Dentistry, The Texas A&M University System Health Science Center, 3302 Gaston Ave., Dallas, TX 75246, USA

Received 22 September 2002;  accepted 22 January 2003. ; Available online 10 April 2003.

1. Abstract

This study characterized the corrosion behavior of cast CP titanium made with a face-coating method. Wax patterns were coated with oxide slurry of Y₂O₃ or ZrO₂ before investing with a MgO-based investment. Three surface preparations were tested: ground, sandblasted, and as-cast. Uncoated castings served as controls. Sixteen-hour open circuit potential (OCP) measurement, linear polarization and potentiodynamic cathodic polarization were performed in an aerated modified Tani-Zucchi synthetic saliva at 37°C. Anodic polarization was conducted in the same deaerated medium. Polarization resistance (R_p) and Tafel slopes were determined. Corrosion current density was calculated for each specimen. Results (n=4) were subjected to nonparametric statistical analysis (
=0.05). Cross sections of cast specimens were examined by optical microscopy. Energy dispersive spectroscopy (EDS) spot analysis was performed at various depths below the surface. The OCP stabilized within several hours for all the specimens. Apparent differences in anodic polarization behavior were observed among the different surfaces. A distinctive wide passive region followed by breakdown was seen on specimens with ground and sandblasted surfaces. There were no significant differences in the corrosion resistance among the control and the two face-coating groups for each group. The Mann-Whitney test showed significantly lower OCP and higher R_p values for ground surfaces. The surface condition significantly affected the corrosion behavior more than the face coating methods. In most cases, specimens with as-cast surfaces exhibited the least corrosion resistance during the potentiodynamic anodic polarization.

Author Keywords: Titanium; Casting; Face-coating; Corrosion behavior;Surface condition

2. Article Outline

1. Introduction

2. Materials and methods

2.1. Preparation of specimens

2.2. Preparation of corrosion evaluation

2.3. Determination of corrosion characteristics

2.4. Examination of cast specimens using an optical microscope

2.5. Examination of cast specimens using scanning electron microscopy (SEM)

3. Results

3.1. Evaluation of corrosion characteristics

3.2. Microstructural examination

3.3. Semi-quantitative EDS analysis

4. Discussion

5. Conclusion

Acknowledgements

References

3. 1. Introduction

Excellent biocompatibility is one of the absolute necessities for a material to be used in vivo. The outstanding biocompatibility of titanium has long been shown in surgical implants and other medical devices [1]. In dentistry, cast titanium has been used in the last two decades for crowns [2], bridges, partial/complete dentures [3, 4 and 5] and metal-ceramic restorations [6]. However, there are still some technical problems to be solved before these dental appliances can be used successfully in clinical practice.

One unique, but often unwelcome, characteristic in investment casting of titanium is the formation of a hardened layer on the surface of the casting due to the reaction between oxides in the investment materials and the molten titanium [7, 8 and 9]. Even when refractories having thermodynamic stability similar to that of the titanium oxide family are used (such as MgO and Al₂O₃), the very strong reducing power of titanium decomposes such common oxides. The oxygen liberated from the oxides diffuses into the surface of the casting to form a hardened layer (150-200 
m thick, depending on the investment used) [9 and 10]. The hardness of this oxygen-rich
-phase stabilized layer known as the "
-case" is as much as three times greater than that in the bulk interior structure [11]. Such an oxygen-rich layer is hard and brittle [12] and is considered detrimental because ductility and fatigue resistance are reduced [13]. This surface layer also prevents strong bonding between titanium and dental porcelain [14].

Several methods have been tested to minimize the formation of this reaction layer, including coating the wax pattern with an oxide that is thermodynamically more stable than the titanium oxides before investing [15]. In a previous study, when a wax pattern was coated with a slurry containing Y₂O₃ powder, the surface hardness (25 
m from the cast surface) of cast commercially pure titanium (CP titanium) was found to be remarkably reduced (280 VHN) compared to the surface hardness of the noncoated cast titanium (530 VHN) [11]. More recently, a study of ZrO₂ coating for CP titanium also showed that the hardness near the cast surface could be reduced (457 VHN) [16], although this method was not as effective as coating with Y₂O₃. Thus, these coating methods appear to be very promising. Since the nature of the surface layer formed from the coating technique could be different from that found in noncoated castings, we were interested in comparing the corrosion behavior of conventionally cast titanium with titanium cast using the coating technique.

Several studies have been published on the corrosion behavior of cast CP titanium with the surface reaction layer intact. Doi et al. [17] examined CP titanium cast into a phosphate-bonded SiO₂ investment. Cyclic potentiodynamic polarization was performed in deaerated 0.9% NaCl solution on three surface conditions: ultrasonically cleaned, sandblasted surface with the reaction layer intact, and ground surface with the reaction layer completely removed. The ultrasonically cleaned surface with the reaction layer showed the least corrosion resistance. The other two conditions showed similar behavior, although the current density on the anodic polarization diagram for the sandblasted surface with the reaction layer was almost ten times larger than that for the ground surface without the reaction layer.

In another study [18], the effect of various surface conditions on the corrosion resistance of cast CP titanium prepared in a phosphate-bonded SiO₂ mold was investigated by anodic polarization in 1% NaCl solution. Anodic polarization was performed on sandblasted CP titanium with the reaction layer, on the acid-treated Si-rich layer, on the ground acid-treated surface, and on the ground bulk surface. Sandblasted specimens with the reaction layer showed a reduced passive region and a considerable increase in the primary passive current density.

In our previous study of CP titanium in artificial saliva (Fusayama's solution) and an acidic saline solution [19], the sandblasted surfaces of cast specimens without the reaction layer showed an increased primary passive current density compared with the ground surfaces. The sandblasted CP titanium specimens with and without the reaction layer had similar polarization diagrams. It was concluded that the surface reaction layer was less influential than surface finishing (ground or sandblasted) on the corrosion behavior of the CP titanium.

More recently, we investigated the corrosion behavior of some titanium alloys [20]. Open-circuit potential (OCP) measurement, linear polarization and potentiodynamic polarization (cathodic and anodic) were performed for these alloy specimens in modified Tani-Zucchi synthetic saliva [21]. Polarization resistance, R_p, was determined from the results of the linear polarization. Based on the results of the anodic polarization diagrams, the current density at passivation, I_pass, and the breakdown potential, E_b, were evaluated. The corrosion current density, I_corr, was also calculated using the polarization resistance and the values of the anodic and cathodic Tafel slope. Regardless of the kind of titanium alloys examined, the measured polarization resistance and corrosion current density revealed that the sandblasted specimens with the reaction layer removed were inferior compared to the specimens with the two other surface conditions.

Since face-coating is an effective method to produce cast titanium with a reduced surface reaction layer and can potentially be used to produce dental prostheses, the corrosion characteristics of cast titanium prepared with the face-coating technique were evaluated in a simulated oral environment. Surface microstructures and compositions of the specimens were examined by scanning electron microscopy (SEM) and energy dispersive spectroscopy (EDS).

4. 2. Materials and methods

2.1. Preparation of specimens

A rod (32 mm in diameter) of CP titanium (CP Ti; ASTM Grade 2, Titanium Industries, Grand Prairie, TX, USA) was machined into 30 g pieces (10.4 mm thick and 29.0 mm in diameter) and used as ingots for casting. A magnesia-based investment material, Selevest CB (Selec Co., Osaka, Japan), was used. Two different oxide powders, Y₂O₃ (1.5 
m; Nihon Yttrium, Tokyo, Japan) and ZrO₂ (<0.5 
m; Aldrich Chemical Company Inc. Milwaukee, WI, USA), were used to coat the wax patterns. The selection of these oxides in the present study was based on their thermodynamic stability at temperatures near the titanium melting temperature [9]. To prepare the coating slurry, 0.7 g methyl cellulose powder (25 CPS, ICN Biomedical, Aurora, OH, USA) was mixed with 34.3 g of deionized water for 15 min using a vacuum mixer (Combination Unit, Model D, Whip Mix, Louisville, KY, USA) to form a methyl cellulose solution. Next, each oxide powder (50 g) was added to the methyl cellulose solution and vacuum mixed for 3 min, followed by vibrating for 2 min using the vibrator on the vacuum mixer. The procedure of mixing the powder with the methyl cellulose solution and the vibration was repeated 5 times in order to produce slurry suitable for coating.

Four square plates (10×10×2 mm) made of an inlay wax (Blue Inlay Casting Wax; Hard-Type I, Class 1, Kerr Co., Romulus, MI, USA) each with a sprue (3 mm long and 2 mm diameter) attached at one of the corners were placed parallel on a runner rod (20 mm long, 5 mm diameter). The runner rod was directly connected to the sprue former. Then, the sprued wax patterns were dipped three times into the oxide/methyl cellulose slurry mentioned above and dried for one hour at room temperature.

Each coated wax pattern was invested with Selevest CB and burnt out following the manufacturer's recommendation (held for 60 min at a peak temperature of 850°C, furnace-cooled and held at 350°C until casting) [22]. Casting was performed in a centrifugal dental titanium casting machine (Ticast Super R, Selec Co., Osaka, Japan). A detailed description of the casting procedure is given elsewhere [23]. For the control specimens, similar castings were made without the oxide-coating steps during investing.

The internal porosity of the cast specimens was radiographically examined using a conventional dental X-ray unit (Dentsply Gendex GX900, Des Plaines, IL, USA) under the following conditions: target film distance, 50 cm; tube voltage, 70 kVp; tube current, 15 mA; exposure, 3 s. Specimens with noticeably large pores were excluded from further testing. The minimum pore diameter detectable by this method was estimated to be 0.05 mm.

2.2. Preparation of corrosion evaluation

The corrosion characteristics of cast titanium were evaluated using open-circuit potential (OCP), linear polarization, potentiodynamic cathodic polarization, and potentiodynamic anodic polarization. Three surface conditions were examined: ground, sandblasted, and as-cast. Ground surfaces were made by grinding the specimens with 1000 grit SiC paper using light pressure in water until a metal luster appeared. Since the as-cast surfaces were not completely flat and included some extruded fins, it was estimated that approximately 60 
m was removed, depending on the kind of specimen. Ground specimens were ultrasonically washed in deionized water for 5 min. Most of the reaction layer remained on the surfaces of these specimens. The sandblasted surfaces were prepared with Al₂O₃ particles (50 
m) using a sand-blaster (Blastmate II System, Ney Dental Inc., Bloomfield, CT, USA) for 15 s [distance between the specimen surface and nozzle: 10 mm; air pressure: approximately 0.59 MPa (85 psi)]. In contrast to the ground specimens, the amount of reaction layer removed by sandblasting was minimal (<10 
m). These specimens were ultrasonically washed in distilled water for 5 min.

A brass nut (Round Thumb Nut, Brass No Finish, Above Board Electronics, Garland, TX, USA) was glued to one of the flat surfaces (10×10 mm) of each specimen using conductive paint (LEIT-C: Conductive Carbon Cement, Electron Microscopy Sciences, Washington, PA, USA) to ensure electrical conduction between the specimen and the brass nut. This assembly was embedded in an epoxy resin disk (Epoxide, Buehler Ltd., Lake Bluff, IL, USA) so that the specimen surface appeared at the bottom of the mount. The brass nut was exposed on the other surface for easy connection to an electrical lead wire. For all the electrochemical tests, a modified Tani-Zucchi solution (MTZ) maintained at 37°C±1°C was used as an electrolyte. This solution (pH=6.7-6.8) consists of 1.5 g each of KCl and NaHCO₃, 0.5 g each of NaH₂PO₄·H₂O and KSCN and 0.9 g of lactic acid in 1000 ml of deionized water [21]. The assembled specimen was placed in a glass corrosion cell and then the glass tip of the salt bridge connected to a saturated calomel electrode (SCE) was placed approximately 1 mm from the center of the specimens. The cell was filled with freshly prepared electrolyte (within 24 h), and a platinum coil (counter electrode) was soaked in the electrolyte; the cell was then tightly sealed.

2.3. Determination of corrosion characteristics

The assembled corrosion cell was connected to a potentiostat (Model 273A; EG&G, Princeton Applied Research, Princeton, NJ, USA) controlled by a personal computer with dedicated software (352 SoftCorr III, EG&G, Princeton Applied Research). The OCP measurement continued up to 16 h after the aerating gas (air+10% CO₂) started to flow through the electrolyte. Linear polarization was then conducted at potentials near the OCP (−8 mV to +8 mV) at an increasing potential rate of 0.1 mV/s. The polarization resistance (R_p: M
 cm²) was calculated from the slope of the electrode potential vs. current density. Subsequently, potentiodynamic cathodic polarization was initiated at the OCP potential and terminated at 300 mV below the OCP value at a rate of 0.167 mV/s in the aerated medium. The cathodic Tafel slope (
_c: V/current density) was calculated as the potential change over one decade (one order of magnitude) decrease in the current density at potentials near the OCP.

OCP measurement determines the corrosion potential of a metal in an electrolyte. Its value can be used to predict long-term service lifetime for metal structures [24]. Linear polarization is a non-destructive test. It is used to determine when a test electrode is at its steady state. Polarization resistance, which is the slope of the linear polarization diagram, is used to estimate the general corrosion rate of the metal. Anodic polarization is a test to characterize the corrosion behavior of a metal and evaluate how effectively a passive film protects a metal from corrosion. Cathodic polarization of a metal was used in combination with anodic polarization to determine Tafel constants and to determine the corrosion current density (I_corr) in the present study.

After these measurements, a mixture of N₂+10% CO₂ gas was introduced into the electrolyte for 2 h and then the potentiodynamic anodic polarization started at 200 mV below the OCP at an increase of 0.167 mV/s up to 2000 mV above the OCP. The anodic Tafel slope (
_a: V/current density) was determined in a similar way. Since the current density increases with an increase in electrode potential, the
_a value is the potential change over one decade increase in current density at potentials near the OCP. Using these three calculated parameters, R_p,
_c and
_a, the corrosion rate, I_corr (A/cm²), was determined for each specimen using the Stern-Geary equation [25]:

Icorr= a c/2.3Rp( a+ c).

Two additional experimental parameters were determined: the current density at which passivation was initiated in the anodic potentiodynamic polarization diagram, I_pass (A/cm²) and the breakdown potential (E_b:V). The breakdown potential is the potential where the current density noticeably increases with increasing potential.

The experimental results were statistically analyzed by a non-parametric method [24] (Kruskal-Wallis test) at a significance level of
=0.05 using computer software (SPSS for Windows, 10.0, Chicago, IL, USA). Since the values appeared not to be statistically different for each surface, the data were further examined using the Mann-Whitney test.

2.4. Examination of cast specimens using an optical microscope

The microstructure near the surfaces of all types of cast specimens tested (three different investing methods and three surface modifications) was examined using an optical microscope (Model Epiphot 200, Nikon Co., Tokyo, Japan). Plate specimens similar to those used for corrosion evaluation were cross sectioned at the center and then electrolytically plated with nickel (approximately 10 
m thick) [26] to retain the edge of the specimens for microstructural examination. The plated specimens were embedded in an epoxy resin so that the cut surface could be seen under the microscope. The embedded specimens were metallographically prepared with a final polish using a 0.02 
m colloidal silica (Blue Colloidal Silica, Allied High Tech Products, Rancho Dominguez, CA, USA).

2.5. Examination of cast specimens using scanning electron microscopy (SEM)

Specimens observed by optical microscopy were also examined using a scanning electron microscope (SEM) (JSM 6300, JEOL, Peabody, MA, USA) equipped with an energy dispersive spectrometer (EDS) (Voyager X-ray microanalysis and digital imaging system, Noran Instruments, Middleton, WI, USA). Amounts of contaminated elements at 5, 10, 20, 30, 40 and 50 
m from the cast surface of each specimen were semi-quantitatively determined by spot analyses using 180 s live time. Two specimens were analyzed for each coating method, and four areas at each depth were randomly selected on each specimen. Results were subjected to
(
z) correction [27].

5. 3. Results

3.1. Evaluation of corrosion characteristics

Table 1 summarizes various corrosion parameters obtained for the three different surfaces tested. Data are included from specimens prepared with both the Y₂O₃ and ZrO₂ coating techniques and from the control specimens. Fig. 1, Fig. 2 and Fig. 3 show typical diagrams of the electrode potential vs. time for the ground, sandblasted, and as-cast specimens. The Kruskal-Wallis test showed no significant differences in OCP values for each of three surfaces among the coated and the control specimens (p=0.225-0.697). The Mann-Whitney test showed that the OCP of the ground specimens was significantly lower than that of the other specimens (p<0.05). This was most evident in the ZrO₂-coated group (Table 1).

Table 1. Corrosion parameters determined in this studya

(7K)

Fig. 1. Change in electrode potential with time for Y₂O₃-coated cast CP titanium.

(8K)

Fig. 2. Change in electrode potential with time for ZrO₂-coated cast CP titanium.

(7K)

Fig. 3. Change in electrode potential with time for noncoated cast CP titanium.

The mean polarization resistance (R_p) values of all the tested specimen groups ranged from 1.30 to 3.20 M
 cm² (Table 1). The Kruskal-Wallis test showed no significant differences for R_p for each of the three surfaces among the coated and the control groups. However, the R_p values of the specimens with ground surfaces were significantly higher than for the specimens with other surfaces (p=0.013).

Fig. 4, Fig. 5 and Fig. 6 show typical potentiodynamic anodic polarization diagrams for the three types of surfaces. From these diagrams, apparent differences in anodic polarization behavior are observed among the surfaces for all the face-coated specimens (Fig. 4, Fig. 5 and Fig. 6). A passive region followed by breakdown and repassivation was seen on specimens with ground and sandblasted surfaces ( Fig. 4 and Fig. 5), but not on the as-cast specimen surfaces. The mean passive regions of the specimens with the ground and the sandblasted surfaces were −6 to 1110 mV and 126 to 1240 mV, respectively. The Kruskal-Wallis test showed no significant differences for any of the face-coated surfaces, but it showed significant differences between the ground and sandblasted surfaces. No significant differences were found in the corrosion current density (I_corr) among the face-coated specimens and the control (p=0.589), and among the three surfaces (p=0.206). The Mann-Whitney test showed significantly higher I_pass values for the sandblasted surfaces compared to the ground surfaces (p<0.01). Therefore, the specimens with the sandblasted surfaces showed lower corrosion resistance.

(7K)

Fig. 4. Potentiodynamic anodic polarization diagram of Y₂O₃-coated cast CP titanium.

(7K)

Fig. 5. Potentiodynamic anodic polarization diagram of ZrO₂-coated cast CP titanium.

(7K)

Fig. 6. Potentiodynamic anodic polarization diagram of noncoated cast CP titanium.

The breakdown potential values ranged from 757 to 1240 mV for all the face-coated titanium. The Kruskal-Wallis test showed no significant differences for any of the face-coated surfaces, but it showed significant differences for two surface conditions (p<0.03). The Mann-Whitney test indicated that the breakdown potentials of the ground specimens were lower than for the sandblasted surfaces. Therefore, the passive region of the ground specimens tended to be narrower than that of the sandblasted specimens.

3.2. Microstructural examination

Typical optical micrographs near the cast surfaces of the noncoated specimens are shown in Fig. 7a-c. The surface layer (approximately 100 
m) resulting from the reaction between the molten titanium and the investment material is seen on an as-cast specimen in Fig. 7a. Note that the surface layer consists of three types of structures: a plate layer on the surface (approximately 5 
m thick), coagulated nodule grains (approximately 25 
m), and elongated rods. Fig. 7b shows a ground specimen from which a majority of the surface reaction layer was removed by grinding. Since at least 60 
m was removed from the cast surface during grinding, the surface structure is very similar to the bulk interior structure. The microstructure at the surface of the sandblasted specimens (Fig. 7c) is similar to that found in the as-cast specimens ( Fig. 7a), since sandblasting removed only a minimal amount of the surface structure. Note that the microstructure near the cast surface of the Y₂O₃-coated specimens (Fig. 8) is similar to that of the interior structure and that no appreciable reaction layer is seen. On the other hand, there was a reaction layer (approximately 50 
m) on the as-cast ZrO₂-coated specimens (Fig. 9) although it was thinner than that on the as-cast specimen (Fig. 7). The surface layer on the ZrO₂-coated specimens consisted of coagulated grains (approximately 25 
m each).

(82K)

Fig. 7. Optical micrographs of area near the cast surfaces of noncoated CP titanium. (a) As-cast specimen, showing the
-case layer with large
-titanium plates. (b) Ground specimen; surface
-case was partially removed. (c) Sandblasted specimen; the
-case layer is similar to Fig. 7a.

(28K)

Fig. 8. Optical micrograph of area near the cast surface of Y₂O₃-coated CP titanium (as-cast specimen). A much reduced
-case is shown. Large dendrites as seen in Fig. 7a are nearly eliminated.

(26K)

Fig. 9. Optical micrograph of area near the cast surface of the ZrO₂-coated CP titanium (as-cast specimen). A reduced
-case layer is shown. Some large dendrites still remain.

3.3. Semi-quantitative EDS analysis

The EDS analysis near the cast surface of the as-cast noncoated specimens revealed that Al and Zr were present to approximately 30 
m below the cast surface. Semi-quantitative analysis revealed that the mean values of Al and Zr at 5 
m from the cast surface were 5 mass % (hereafter, all percentages in mass % will be shown as "%") and 1%, respectively. These concentrations rapidly decreased with depth towards the interior of the specimens. In the ZrO₂-coated specimens, a comparatively high concentration of Zr (16%) was found 5 
m from the surface and gradually decreased to a minimal amount at 40 
m below the surface. As in the as-cast noncoated specimens, approximately 5% Al was found at the cast surface (5 
m), and the concentration was reduced to a minimal level at approximately 30 
m below the surface. In the Y₂O₃-coated specimens, the only elements found at an appreciable level were Ca (approximately 1.5%) and Al (<1%).

6. 4. Discussion

In the present study, the OCP measurements, linear polarization, and cathodic polarization were conducted in aerated artificial saliva. Anodic polarization was performed in a deaerated environment. Saliva in the human oral cavity is generally considered to be an aerated environment. Therefore, OCP measurement, linear polarization and cathodic polarization were conducted in an aerated medium in this study. Based on the mixed-potential theory, a deaerated environment will provide additional information on the anodic polarization of test specimens. However, a deaerated environment will only affect the reduction reaction, not the oxidation reaction. In addition, certain areas in the oral cavity, such as the subgingival region, and areas under the plaque, are in a deareated condition.

The cast titanium specimens in this study prepared with three investment methods had different surface microstructures, depending on the level and type of reaction that took place between the mold and molten titanium. Although the amounts of elements diffused into the structure near the cast surface are different among the types of specimens, the main elements found were Al and Zr or Ca. It is reported that the magnesia-based investment material we used, Selevest CB, contains 67.0% MgO, 28.0% alumina cement, 2.7% Zr powder, and 2.3% other [28]. The main constituent of alumina cement is Al₂O₃ and CaO. Thus, the Al we detected in the noncoated and the ZrO₂-coated specimens must be the result of the decomposition of Al₂O₃ by the reducing action of the molten titanium. Since the standard free energy of formation of Al₂O₃ at a temperature above the melting point of titanium is close to that of the titanium oxide family [9], Al was liberated from Al₂O₃ by the molten titanium and diffused into the titanium that solidified adjacent to the mold surface. Since the concentrations of Al found in the non-coated and ZrO₂-coated specimens are similar, Al diffused into the titanium in these specimens by a similar process. In contrast, the concentration of Zr in the ZrO₂-coated specimens was approximately sixteen-fold higher than in the non-coated specimens. Thus, a portion of the Zr diffusing into the ZrO₂-coated specimens must also be from the metallic Zr in the investment. On the other hand, note that the diffused elements in the surface structure in the Y₂O₃-coated specimens were minimal, indicating that the Y₂O₃ layer formed in the mold by the coating technique was effective at minimizing the surface reactions. This finding is also consistent with the observed specimen microstructures (Fig. 8), which had no appreciable reaction layer.

To prepare specimens for the present study, approximately 60 
m of the surface layer was removed when the specimens were ground. As described above, the depth of the diffused elements was 30 
m at most for the noncoated specimens and approximately 40 
m for the ZrO₂-coated specimens. Therefore, the surface structure contaminated with elements from the mold was totally removed from all the ground specimens. In other words, the surface structure of these specimens was similar to that of the interior of the CP titanium. However, the surface structure of the as-cast and the sandblasted specimens contained Al and Zr, since the surface reaction layer was virtually intact.

An initial increase in the OCPs during the early hours followed by stabilization observed on all the specimens suggested that a protective passive film formed rapidly on the metal surfaces in the artificial saliva and remained stable during the entire immersion period. The OCP values were not significantly different among the face-coated and the noncoated specimens. However, OCP values of the as-cast specimens were significantly higher than those of the ground and sandblasted ones, and the values of the ground specimens were significantly lower than the others (p<0.001). This finding indicates that the as-cast surface was the most stable among the surface conditions tested, which may be attributed to the thicker titanium oxide remaining on the surface. Therefore, the OCP values were not affected by face-coating, but by the surface condition. Even though the surface reaction layers were not totally removed from the sandblasted and as-cast specimens, the contaminated elements in these surface layers apparently did not affect their OCP values.

As for the passive current density, I_pass, the sandblasted specimens exhibited higher values compared to those for ground specimens. A similar observation was made in our previous studies [19 and 20], in which an increase in I_pass was found due to the rough surface. Sandblasting produced increased surface area compared to the surface area of the ground specimens. The I_pass values may also have been higher for the sandblasted specimens due to the reaction layer remaining on the specimen surfaces. The large standard deviations in the I_pass results may have happened because of the different levels of surface roughness and contamination among the specimens.

In this study, specimens having the same surface condition showed similar corrosion behavior regardless of the investment methods used, indicating that similar protective titanium oxide films must form on both face-coated and noncoated specimens. An electron diffraction study on titanium after anodic polarization by Kuphasuk [29] revealed that the TiO₂ (rutile) with tetragonal structure was the main oxide formed on the surface. Koike and Fujii [30] also found the formation of TiO₂ on cast CP titanium in various solutions. They reported that the thickness of the oxide layer depended on the kind of corrosion media. The lower breakdown potential values (E_b) found in the present study for the ground specimens may occur because the oxide film on the ground specimens was thinner. Such a thin titanium oxide film might cause surfaces to break down at lower electrode potentials. In this study, breakdown of the passive film was observed on specimens with ground and sandblasted surfaces. It is well recognized that breakdown of the passive film on titanium surfaces in a physiological environment is due to the pitting of chloride ions incorporated in the passive film [31 and 32].

The other two experimental parameters examined (the polarization resistance (R_p) and corrosion current density (I_corr)) should be negatively correlated to each other [24], and we found such a tendency in our study. Although no statistical significance was found among the surface conditions for I_corr, we found that the R_p values of the ground specimens were significantly the highest among all the conditions.

7. 5. Conclusion

We found that all the corrosion characteristics of the cast titanium examined were similar among the investment methods. Surface condition appears to be the determining factor for the electrochemical behavior of titanium in a simulated oral environment. In particular, it significantly affected the polarization behavior more than the face-coating methods did. Ground surfaces exhibited superior corrosion resistance compared to sandblasted surfaces. In most cases, the specimens with as-cast surfaces exhibited the least corrosion resistance because of their rough surface condition. However, within the oxidation potential range (−58 to +212 mV vs. SCE) in the normal human oral cavity [33], all the face-coating/surface combinations examined in this study showed excellent corrosion resistance.

8. Acknowledgements

A portion of this study was supported by a research grant (R01 DE11787) funded through the National Institutes of Health/National Institute of Dental and Craniofacial Research, Bethesda, MD. The authors gratefully acknowledge editorial assistance with this paper from Mrs. Jeanne Santa Cruz.

9. References

1. J.C. Wataha, Materials for endosseous dental implants. J Oral Rehabil 23 (1996), pp. 79-90. Abstract-MEDLINE  

2. K. Ida, Y. Tani, S. Tsutsumi, T. Togaya, T. Nambu, K. Suese, T. Kawazoe, M. Nakamura and H. Wada, Clinical application of pure titanium crown. Dent Mater J 4 (1985), pp. 191-195. Abstract-MEDLINE  

3. D. Low, H. Best and T. Mori, An experimental model for the assessment of titanium denture casting techniques. Dent Mater J 13 (1994), pp. 103-109. Abstract-MEDLINE  

4. A.J. Sutton and P.M. Rogers, Discoloration of a titanium alloy removable partial denture: a clinical report. J Prosthodont 10 (2001), pp. 102-104. Abstract-MEDLINE  

5. R.R. Wang and A. Fenton, Titanium for prosthodontic applications: a review of the literature. Quintessence Int 27 (1996), pp. 401-408. Abstract-MEDLINE  

6. T. Togaya, M. Suzuki, S. Tsutsumi and K. Ida, An application of pure titanium to the metal porcelain system. Dent Mater J 2 (1983), pp. 210-219. Abstract-MEDLINE  

7. O. Miyakawa, K. Watanabe, S. Okawa, S. Nakano, M. Kobayashi and N. Shiokawa, Layered structure of cast titanium surface. Dent Mater J 8 (1989), pp. 75-185.

8. O. Miyakawa, K. Watanabe, S. Okawa, S. Nakano and M. Kobayashi, Reaction layers of titanium cast in molds containing spinel. J Dent Mater 14 (1995), pp. 560-568.

9. T. Okabe and H. Hero, The use of titanium in dentistry. Cell Mater 5 (1995), pp. 211-230. Abstract-EMBASE  

10. J. Zhang, M. Okazaki and J. Takahashi, Effect of casting methods on surface state of pure titanium casting. J Dent Mater 13 (1994), pp. 221-227.

11. Okabe T, Shimizu H, Woldu M, Brezner M, Carrasco L, Watanabe K. Mechanical properties of titanium cast using yttria face-coating. THERMEC 2000-Proceedings of International Conference on Processing and Manufacturing of Advanced Materials, vol. 117/3, Section A7, Las Vegas, USA, December, CDROM, 2000.

12. M.L. Bauccio, Technical note 9: descaling and special surface treatments. In: R. Boyer, G. Welsch and E.W. Collings, Editors, Materials Properties Handbook: Titanium Alloys, ASM International, Materials Park, OH (1994), p. 1080.

13. Donachie MJ. Titanium: A technical guide. Metals Park, OH: ASM International, 1988. p. 28.

14. Z. Cai, N. Bunce and T. Nunn, M, Okabe, Porcelain adherence to dental cast CP titanium: effects of surface modifications. Biomaterials 22 (2001), pp. 979-986. SummaryPlus | Full Text + Links | PDF (336 K)

15. H. Hashimoto, A. Kuroiwa, K. Wada, Y. Hibino, H. Kouchi, T. Hashimoto, Y. Hasegawa, Y. Ando and Y. Akaiwa, Reaction product on the surface of titanium castings. J Dent Mater 11 (1992), pp. 603-614.

16. Shafer T, Cai Z, Carrasco L, Okabe T. Porcelain bonding to cast titanium made from oxide face-coated patterns. J Dent Res 2002;81:A-261, Abst 2023.

17. H. Doi, T. Yoneyama, M. Kotake and H. Hamanaka, Corrosion resistance on the reacted surface of titanium castings with a phosphate bonded silica investment. J Dent Mater 11 (1992), pp. 672-678.

18. H. Miura and Y. Miyagawa, Electrochemical corrosion behavior of titanium castings. Part 1. Effects of degree of surface polishing and kind of solution. J Dent Mater 15 (1996), pp. 183-295.

19. Z. Cai, H. Nakajima, M. Woldu, A. Berglund, M. Bergman and T. Okabe, In vitro corrosion resistance of titanium made using different fabrication methods. Biomaterials 20 (1999), pp. 183-190. SummaryPlus | Full Text + Links | PDF (153 K)

20. Z. Cai, T. Shafer, I. Watanabe, M.E. Nunn and T. Okabe, Electrochemical characterization of cast titanium alloys. Biomaterials 24 (2003), pp. 213-218. SummaryPlus | Full Text + Links | PDF (208 K)

21. M. Marek, The release of mercury from dental amalgam: the mechanism and in vitro testing. J Dent Res 69 (1990), pp. 1167-1174. Abstract-MEDLINE  

22. Nissin Co. Instruction manual of Selevest CB, C&B Investment for Pure Titanium, 1999.

23. T. Okabe, C. Ohkubo, I. Watanabe, O. Okuno and Y. Takada, The present state of dental titanium casting. J Metals 50 (1998), pp. 24-29.

24. W.S. Tait. An introduction to electrochemical corrosion testing for practicing engineers and scientists, Pair O Docs Publications, Wisconsin (1994).

25. M. Stern and A.L. Geary, Electrochemical polarization. I. A theoretical analysis of the shape of polarization curves. J Electrochem Soc 104 (1957), pp. 56-63.

26. Murphy, Michael, editor. Metal Finishing. 62nd Guidebook and Directory Issue. New York: Elsevier Science Inc., 1994.

27. Goldstein JI, editor. Scanning Electron and X-ray Microanalysis. A text for biologists, materials scientists, and geologists. 2nd ed. New York: Plenum Press, 1992. p. 395-416.

28. Ohyagi S, Iwaki S, Shimakawa S, Nakai F, Togaya T. Properties of an aluminous cement-bonded magnesia investment (Selevest). Proceedings of the Third International Symposium on Titanium in Dentistry, Australia, 1995. p. 186.

29. C. Kuphasuk, Y. Oshida, C.J. Andres, S.T. Hovijitra, M.T. Barco and D.T. Brown, Electrochemical corrosion of titanium and titanium-based alloy. J Prosthet Dent 85 (2001), pp. 195-202. Abstract | PDF (180 K)

30. M. Koike and H. Fujii, In vitro assessment of corrosive properties of titanium as a biomaterial. J Oral Rehabil 28 (2001), pp. 540-548. Abstract-MEDLINE   | Full Text via CrossRef

31. Stansbury EE, Buchanan RA, editors. Fundamentals of electrochemical corrosion. Materials Park, OH: ASM International, 2000. p. 233-43; 277-83; 295-6.

32. R. Schenk, The corrosion properties of titanium and titanium alloys. In: D. Brunette, P. Tengvall, M. Textor and P. Thomsen, Editors, Titanium in medicine: Materials science, surface science, engineering, biological responses and medical applications, Springer, Berlin (2001), pp. 146-170.

33. G.J. Ewers and E.H. Greener, The electrochemical activity of the oral cavity. J Oral Rehabil 12 (1985), pp. 469-476. Abstract-EMBASE | Abstract-MEDLINE  

Corresponding author. Department of Biomaterials Science, Baylor College of Dentistry, The Texas A&M University System Health Science Center, 3302 Gaston Ave., , Dallas, TX 75246, , USA. Tel.: [phone]; fax: [phone]

---