Removal of the alpha-Case Layer

High specific yield strength, good corrosion resistance, excellent fatigue properties, and biocompatibility are the most important properties that make titanium alloys suitable for medical applications. Until recently, TiAl₆V₄ (Ti64) was used for most biomedical applications, but vanadium is now suspected to be toxic to the human body in the long term. Instead, the α + β titanium alloy TiAl₆Nb₇ (Ti67), which has similar properties to Ti64, is now more commonly used.

The use of cast titanium in biomedical applications has been limited by inherent casting problems. Molten titanium is extremely reactive with elements such as nitrogen and oxygen, as well as with those present in the investment casting process. The high affinity of titanium to oxygen in combination with its high solubility quickly results in the formation of titanium oxides and an oxygen-enriched surface layer on the base metal. Additionally, reactions with the mold material may cause the formation of embrittling phases.

The brittle surface layer that has been observed on titanium alloys is termed the "α-case layer". The presence of such a layer promotes crack formation under static or cyclic loading conditions, thus reducing the usefulness of the material. Careful attention has to be paid to avoid the presence of the α-case; this can be achieved before formation, through vacuum metallurgy, or after, by removal of the embrittling layer mechanically or chemically; for example, by acid pickling.

Now, a team of German materials scientists has reported a new method for removing the α-case layer in TI67. Casting of Ti67 specimens was carried out under an inert argon atmosphere, using MgO+Al₂O₃+CaO-based investments as the mold material in a centrifugal casting machine. Post-casting, an acid mixture of 70% HNO₃ + 10% HF and distilled water was used to remove the α-case layer. Cast rectangular samples with dimensions of 10 x 10 x 1 mm³ and sponge specimens with dimensions of 26 x 14.5 x 12mm³ were used to carry out these pickling studies.

A three-layer structure was observed in the cast specimen. The outer layer was identified as an α-case between 5- and 30-μm thick. This fine-grained α-case layer was formed through a reaction between the Ti67 alloy and the aluminum from the investment mold material. Underneath the α-case layer, a coarse acicular Widmanstätten microstructure was detected. The acicular zone had a thickness between approximately 20 and 50 μm and consisted of α -Ti laths formed within a β-Ti matrix. The size of the acicular layer depended on the kind of investment material, the mold temperature, and the casting volume. The third layer was found to be a bulk microstructure with fine plates. This layer, present in the core of the sample, was expected to have the same properties as the bulk Ti67 alloy.

Differences in the properties of the α-case layer, the coarse acicular layer, and the bulk zone were observed by measuring the Vickers microhardness profile across the sample. While the microhardness of the α-case was measured to be around 800 HV, it was found to be 350 HV in the core bulk microstructure. The coarse acicular layer exhibited intermediate hardness values of around 620 HV. This increase in hardness from the core to the surface can be partially attributed to rapid cooling at the mold surface during casting or to an increase in the concentration of dissolved oxygen.

The mass loss observed during the initial acid pickling was linear, but after 60–70 minutes the slope of the curve decreased. This reduction of the pickling speed was considered to be a signal of complete α-case elimination. The percentage of mass loss at this point was around 10% of the sample mass. This unexpectedly high mass loss was due to the additional removal of investment residues from the casting process. After 70 minutes, the surface of the sample was completely cleaned.

The pitting corrosion effect was the predominant mechanism of material removal from the samples during the pickling process. However, a gradual reduction of this effect on the samples’ surfaces was observed during the procedure. After 20 minutes of pickling, the effect of pitting corrosion was high. After 40 minutes, the specimen still was highly affected by pitting corrosion, but the loci of the pitting corrosion marks had begun to be replaced by rounded marks, similar to worm burrows. These “worm marks” grew during the pickling process until they covered the complete surface, after 60–70 minutes. Unlike the pitting cracks, the worm marks did not exhibit sharp edges, which can initiate cracks under mechanical loading conditions. For biomedical applications, pitting corrosion marks and cracks are not acceptable, as they act as nuclei for fatigue crack formation. This can lead to premature structure failure, which makes the material unattractive for use in bio-implants.

These worm marks can also explain the change in the rate of mass loss during the pickling process. Specimens with strong pitting corrosion exhibit more effective surface area to be attacked, and so lose mass faster than the specimens fully covered by worm marks.

In an attempt to increase the speed of the pickling process, ultrasonic vibrations were passed through the acid bath during the cleaning process. These vibration helped to eliminate any non-metallic residues from the surface of the specimens, and also ensured that the acid bath remained homogeneous, to ensure a uniform surface-cleaning effect.

In order to measure the effectiveness of the treatment, the surface roughness of the samples was measured. Nonpickled samples showed arithmetic mean roughness values of about R_a = 5.47 μm and maximum roughness values of R_max = 44.72 μm. After 70 min of pickling the values dropped substantially, to R_a = 1.01 μm and R_max = 8.84 μm, thus showing the potential of this treatment method for manufacturing biomedical implants.

T. Guillen et al., Adv. Eng. Mater. 2009, 11, 680 ; DOI: 10.1002/adem.200900097

More Articles in Advanced Engineering Materials