A three-step metallographic preparation procedure was developed for titanium and its alloys. Attack polishing is utilized in the third step for optimal results, particularly for imaging alpha-Ti with polarized light. Two-phase a-b alloy specimens and all b alloys are easier to prepare than single-phase a specimens. Kroll's reagent appears to be adequate for most alloys. A modification of Weck's reagent was used for color metallography.

Fig. 1. Polished surface of alpha-Ti, ASTM F67, Grade 2, in the annealed (1038°C) condition showing (left edge) extreme surface damage due to band sawing (modified Weck's reagent, polarized light plus sensitive tint).


Titanium and its alloys have become quite important commercially over the past 50 years due to their low density, good strength-to-weight ratio, excellent corrosion resistance and good mechanical properties. On the negative side, the alloys are expensive to produce.

Titanium, like iron, is allotropic and this produces many heat treatment similarities with steels. Moreover, the influences of alloying elements are assessed in like manner regarding their ability to stabilize the low temperature phase, alpha, or the high temperature phase, beta. Like steels, Ti and its alloys are generally characterized by their stable room temperature phases - alpha alloys, alpha-beta alloys and beta alloys, but with two additional categories: near alpha and near beta.

Titanium and its alloys are more difficult to prepare for metallographic examination than steels. As for all refractory metals, titanium and its alloys have much lower grinding and polishing rates than steels. Deformation twinning can be induced in alpha alloys by overly aggressive sectioning and grinding procedures. It is safest to mount relatively pure Ti specimens, especially those from service in a hydrogen-containing environment, in castable ("cold") resins rather than using hot compression mounting due to the potential for altering the hydride content and morphology. However, these resins must be used in such a way as to minimize the heat of polymerization. Elimination of smearing and scratches during polishing can be difficult.

Early mechanical preparation techniques [1-5] tended to be rather long, involving procedures nearly always incorporating an attack polishing solution in the last step or last two steps. Some of the more commonly used attack polishing solutions are summarized in [6]. The problems associated with obtaining well-prepared surfaces have prompted considerable interest in electropolishing procedures [3-5, 7, 8]. The inherent danger of some of these electrolytes has prompted interest in chemical polishing procedures [9]. Electrolytic and chemical polishing solutions for Ti and Ti alloys are also summarized in [6]. Mechanical polishing methods for titanium and its alloys continued to rely upon these older procedures into the 1970s [10] and 1980s [11]. Perhaps the first publication of a modern approach for preparing titanium was that of Springer and Ahmed [12] in 1984. This was a three-step procedure, assuming that the planar grinding step can be performed with 320-grit SiC paper, which may not always be possible. If the specimens are sectioned using a wafering blade or an abrasive blade of the proper bond strength, which produce a smooth surface with minimal damage, then 320-grit SiC paper may be used. If a rougher surface with greater damage was produced, such as would result from use of a power hacksaw, then grinding must commence with a coarser grit paper in order to remove the damage in a reasonable time. Grinding and polishing rates of Ti are much lower than for many other metals and alloys.

Fig. 2. Appearance of titanium hydride at the inner diameter of a CP Ti tube that broke in service showing the greatest amount of TiH in (a) where a low-viscosity, slow curing epoxy was used with a conductive mounting approach to keep the heat of polymerization below 30ºC. (samples unetched)

Specimen Preparation

Although Ti and its alloys can be readily sectioned using band saws, power hacksaws and similar machine-shop tools, these devices produced a great deal of damage. Figure 1 demonstrates the substantial depth of damage that can be produced when sectioning commercial purity (CP) titanium. If the left edge was chosen for the plane-of-polish, then at least 200µm must be ground away to get through the sectioning damage. This damage will be difficult to remove in rough grinding, as the grinding rate is very low. Consequently, to obtain perfect surfaces, section Ti and its alloys with only laboratory abrasive saws or precision saws using blades designed for metallography (avoid using blades made for production machining).

Strictly speaking, any mounting compound can be used for Ti and its alloys. If specimens of Ti used in applications where hydrogen can be picked up are to be mounted, it is best to use a low-viscosity epoxy resin and a conductive mounting approach to minimize the exotherm during polymerization. If the heat involved in polymerization is substantial, titanium hydrides could be dissolved. Specimens never placed in service are unlikely to contain hydrides, and more freedom of choice in mounting is pos-sible. To minimize the heat of polymerization, wrap aluminum foil, as used in cooking, around a block of steel or copper (a heat sink). Then, glue a phenolic ring form (a cylinder) to the foil to create a mold. Place the specimen inside the ring form and add the epoxy. If a low-viscosity epoxy resin is used, which cures slowly, the exotherm during curing will be <10°C above room tem-perature. If a plastic or silicone rubber mold is used with the same epoxy, the exotherm will be higher. The faster the epoxy cures, the higher the exotherm. Acrylic resins cure in less than 10 minutes and the exotherm is very high - high enough to burn your fingers if you touch the mold while it is curing. That is not "cold" mounting! Mounting of your specimens facilitates specimen identification, simplifies automation and yields far better edge retention than unmounted specimens. But choose a resin that does not produce shrinkage gaps.

Fig. 3. Microstructure of as-hot rolled ASTM F67 Grade 2 CP Ti revealed (a) after the three-step method and (b) after 20 minutes of vibratory polishing after the three-step method. The specimens are in cross-polarized light and are not etched.

To illustrate the effect of mounting temperature, Fig. 2 shows the microstructure of a CP Ti tube 19-mm in diameter with a 1-mm wall thickness that was used in a hydrogen-bearing atmosphere. The tube became plugged and broke in service. The writer cut several rings from the tube and mounted them with different compounds: EpoMet® thermosetting resin using a hot mounting press, EpoKwick® fast-curing (~45 minutes) epoxy, EpoThin® low-viscosity, slow curing (~8 hours) epoxy, and several others, including a cast acrylic resin. The specimen mounted in EpoThin resin contained the most TiH; all others, regardless of the type of resin, contained somewhat less TiH. Interestingly, the interface between the alpha-Ti matrix and the TiH was not as sharp in the specimen mounted with the slow-curing epoxy using a conductive molding approach as the interfaces for all other mounted specimens. The hot-mounted specimen (using a mounting press at 150°C) appears to have at least as much TiH, if not more, than the specimen mounted using a fast-curing epoxy in a plastic mold.

Fig. 4. CP Ti (ASTM F67, Grade 4, longitudinal plane, annealed) prepared using the three-step method followed by etching with Kroll's reagent and viewing with bright field illumination.

a) ID of tube mounted in EpoThin resin using the conductive mounting approach.
b) ID of tube mounted in fast-curing EpoKwick resin with a polymeric mold.
c) ID of a tube mounted in a press at 150°C using EpoMet thermosetting resin.

Fig. 5. CP Ti (ASTM F67, Grade 4, transverse plane, annealed) prepared using the three-step method and tint etched with modified Weck's reagent. The specimen was examined with cross-polarized light and sensitive tint to enhance coloration.

A series of experiments was conducted to develop an improved method to prepare titanium and its alloys. Numerous surfaces were tried with the aim of producing a damage free surface in CP Ti so that good polarized light images can be obtained after the last step. The method developed works best when cutting damage is minimized. Sectioning is a violent process and the vast majority of problems encountered in specimen preparation can be attributed to failure to remove the damage from sectioning. So, the first rule for successful preparation is: introduce the least possible amount of damage in sectioning (which will also produce a good surface finish). Next, mount the specimen for ease of identification and for facilitation of edge retention. Then, commence grinding with the finest possible SiC grit size 240-grit is usually adequate and 320-grit SiC may be used if you are careful in placing the specimens in the specimen holder so that the surfaces are flat and parallel to the SiC paper surface. The next rule is: commence grinding with the finest possible abrasive that will remove the sectioning damage in a reasonable time. Coarse abrasives introduce more damage than fine abrasives. Automated grinding and polishing is highly recommended, not only because it yields superior results compared to hand polishing, but also because the final step employs an attack-polishing agent.

Fig. 6. Alpha phase stabilized at the surface of a heat-treated Ti-3%Cr experimental alloy prepared using the three-step method and etched with Kroll's reagent.

Step 2 utilizes psa (pressure-sensitive adhesive) backed silk cloth and 9-µm diamond abrasive. Setting the platen speed at about 100 rpm, placing the syringe tip at the center of the cloth and slowly pulling the tip towards the cloth periphery, charges the cloth with diamond in paste form. This deposits a concentric track of diamond on the cloth. Turn off the polisher and rub the paste into the cloth surface. Then, add a petroleum-based lubricant and commence polishing at 150 rpm, 6 lbs (27 N) load using "contra" rotation. In this approach, the head rotates clockwise while the platen rotates counterclockwise. Every 30 seconds, squirt a small amount of 9-µm diamond suspension onto the cloth to keep the cutting rate high. Continue polishing for 10 minutes. After 10 minutes, clean the specimens and the holder and change the surface.

For step 3, I use a psa-backed, napped MicroCloth® pad (synthetic suede) with the same load, rpm, time and rotation direction using MasterMet® colloidal silica as the abrasive. Mix 5 parts colloidal silica with 1 part hydrogen peroxide (30% concentration - avoid skin contact) as the attack-polish agent. Contra rotation works best when the head speed is under 100 rpm. The machines used for these experiments have a 60-rpm head speed. This helps to keep the abrasives on the cloth surface. If the head and platen both rotate in the same direction (called "complementary" rotation), centrifugal force throws the abrasive and the lubricant off the surface almost as fast as you add it. A rule for polishing is: keep the polishing surfaces uniformly covered with abrasive and lubricant to minimize smearing, pullout and deformation. After step 3, clean the specimen and holder. I stop adding any abrasive at least 20 seconds before the 10 minute polishing cycle ends. With 10 seconds remaining, direct the water jet onto the polishing cloth to clean both the cloth and the specimens. Colloidal silica is more difficult to remove from specimens than other abrasives. Table 1 summarizes the three-step preparation method.

Fig. 7. Microstructure of Ti-6%Al-2%Sn-4%Zr-2%Mo-0.1% Si after (a) alpha-beta forging at 954ºC and alpha-beta annealing at 969ºC and (b) after beta forging at 1038ºC and beta annealing at 1024ºC. The specimens were prepared using the three-step method and etched with Kroll's reagent.

After step 3, CP Ti can be examined as polished with crossed polarized light to observe the grain structure. Figure 3a shows an example of ASTM F67 Grade 2 CP Ti examined after the 3-step preparation procedure. This is an as-rolled specimen and it contains some mechanical twins. If the specimen is placed on a vibratory polisher, using only colloidal silica (no attack polishing agent), better coloration can be obtained although no further detail is detected, as shown in Fig. 3b.

Examination of CP Ti is actually more effective with polarized light in the as-polished condition, when using a properly prepared specimen, than with bright field illumination after etching. Figure 4 shows the microstructure of CP Ti in bright field after etching with Kroll's reagent. The grain structure is reasonably well delineated, but details are not as good as using polarized light on an as-polished specimen. Color etching with a modification of Weck's reagent also produces better grain structure development than Kroll's reagent (Fig. 5). Weck's reagent for Ti contains: 100mL water, 50mL ethanol and 2g NH4F·HF. This composition will produce white "butterfly-shaped" artifacts in the color image, which can be eliminated using only 25mL ethanol. Etch by immersion until the surface is colored, usually about 15-25 seconds. Coloration is enhanced with examination using polarized light and a sensitive tint filter. It is often helpful to move slightly off the crossed position.

Fig. 8. Basket weave alpha-beta microstructure of as-cast Ti-4%Zr annealed at 800°C after etching with modified Weck's reagent and viewed with polarized light plus sensitive tint.

Other cloths can be used for the final polishing step, e.g., MasterTex® and ChemoMet® cloths. The three-step method works very well on the alpha-beta alloys and the beta alloys. Alumina suspension works nearly as well as colloidal silica.

A few variants of the attack polishing solution have been tried. Leonhardt [14] uses a mixture of 150mL colloidal silica, 150mL water, 30mL H2O2 (30%), 1-5mL HF and 1-5mL HNO3. Results with this attack polishing additive to the abrasive were equiva-lent to the one used. Buchheit [6] added 5mL of a 20% aqueous CrO3 solution to 30mL of an alumina slurry. To try this, but us-ing colloidal silica instead, 10mL of the 20% CrO3 solution was added to 75mL of colloidal silica. This also produced excellent results. When using these attack polishing solutions, care must be taken in handling, mixing and using these additives as they contain very strong oxidizers and acids. Avoid physical contact with the ingredients and the prepared attack polishing abrasives.

Fig. 9. Microstructure of a laser weld in Ti-6%Al-4%V etched with modified Weck's reagent and viewed with polarized light plus sensitive tint.

Microstructures

Quality control laboratories frequently check lots of titanium for the presence of an alpha case at the surface due to oxygen pick-up during heat treatment. Oxygen is an alpha stabilizer and the case is detrimental to machining, mechanical properties and service life. Good edge retention is important for this work and mounting is necessary. Edge retention is highly dependent upon elimination of shrinkage gaps between the specimen and the mount. EpoMet resin gives superb results but requires a mounting press. Of the cast resins, epoxy works best. The three-step method, despite step 3 being 10 minutes on a napped cloth, gives perfect results using either EpoMet resin or an epoxy resin. The specimens are perfectly flat coming into step 3. As long as the pressure is kept at 6 lbs, and not lower, flatness is not impaired. Figure 6 illustrates alpha case in an experimental Ti alloy prepared using the three-step method.

Alpha-beta alloys respond perfectly to the three-step method, as they are easier to prepare than the alpha alloys. Figure 7 shows the microstructure of an alpha-beta alloy, Ti6242, after alpha-beta forging and alpha-beta annealing compared to the same alloy after beta forging and beta annealing. The beta transus temperature for this alloy is 995°C ± 15°C. Forging and annealing below the beta transus results in a fine grained alpha-beta microstructure (primary alpha and transformed beta) while forging and annealing above the beta transus results in a coarse grained basket weave alpha-beta microstructure.

Fig. 10. Microstructure of beta alloys prepared using the three-step method: a) Ti-5%V-3%Al-3%Cr-3%Sn (beta transus is ~760°C); and, b) Ti-3%Al-8%V-6%Cr-4%Mo-4% Zr, called Beta C (beta transus is 730°C). Both etched with Krolls.

Modified Weck's reagent can also be used with alpha-beta alloys with good results. As an illustration, Fig. 8 shows the microstructure of as cast and heat-treated Ti-4%Zr while Fig. 9 shows the microstructure of a laser weld in Ti-6%Al-4%V. Both were etched in modified Weck's reagent and are viewed in polarized light plus sensitive tint.

Beta alloys can also be prepared easily with the three-step method. Figure 10 illustrates the microstructure of two beta alloys, Ti-5333 and Beta C.

Conclusions

A three-step procedure was developed and found to be quite successful for preparing titanium and titanium alloys. Use of an attack-polish additive in step 3 is required to obtain good results with CP titanium and alpha Ti alloys. Most two-phase Ti alloys can be satisfactorily prepared without using an attack-polishing additive, although results were better when it was used. Polarized light is very effective for examining the microstructure of alpha-Ti. Color etching can be used to reveal the microstructure of alpha and alpha-beta alloys. Kroll's reagent works well for alpha-beta and beta alloys. IH