July 2007

By Tim Dyer
Carpenter Advanced Ceramics
Auburn, CA, USA

Despite common perception, it seems that not all commercial 99.8% pure alumina oxides are alike. While they may have the same chemistry and meet accepted industry specifications for this material, they can behave differently under key operating conditions.

Recent accelerated acid etching tests were performed on samples of this high-purity ceramic from seven different manufacturers. The tests revealed different levels of surface roughness degradation and vacuum integrity after acid etching.

That should be a concern because those characteristics have an important bearing on surface finish, fine grain size, uniform structure, high density, corrosion resistance, polishability and seal quality in applications where alumina components must withstand long-term exposure to acid. 

Pertinent uses include semiconductor processes, a broad range of industrial bearings that are subject to corrosion attack, high-speed rotating equipment, sliding blood valves and desalination operations.

The tests were conducted by Carpenter Advanced Ceramics (CAC), with subcontractor assistance, specifically to quantify the reactions of commercial 99.8% purity aluminas to the semiconductor processes employed in recycling after use in manufacturing equipment.

This approach recognized the fact that ceramic semiconductor process kit components can be recycled many times after use. Acid etching is commonly performed to recycle alumina reactor components when process residues need to be removed. The main objective of the test was to simulate and evaluate the effect of multiple use and recovery cycles on a process kit component.

Test Process

In the testing process, the seven different 998 alumina samples were polished and submerged for 10 to 20 minutes in a relatively aggressive 33/33/33% solution (enough to dissolve glass) of hydrofluoric acid, nitric acid and deionized water. Each ring-shape test sample was 200 mm diameter round by 5 mm thick, diamond polished on both sides to a surface of 5 micro-inches Ra.

To quantitatively measure surface damage and roughness (Ra) caused by the acid, ceramic samples were evaluated using a stylus profilometer.  Vacuum helium leak testing was done using a custom fixture before and after acid exposure. The same seal was used after each leak measurement.

After polishing to a fine finish at CAC, the alumina samples were cleaned in ultrasound and a mild soap solution for 10 minutes, then rinsed and dried. Helium leak testing was conducted to ensure that all samples were leak tight prior to acid etching.

A custom-made aluminum clam shell fixture was employed to test the Helium leak rate on both sides of the sample plate through a sealing length of 77 cm. Viton 747 o-ring seals were used for all tests. Samples were placed into the fixture attached to a Veeco MS50 vacuum Helium leak analyzer.

Helium leak tests were run twice, then the best result of both tests was recorded. Both of the o-rings were cleaned carefully for all tests. Acid-etched samples were professionally cleaned; therefore, they were not touched prior to leak testing or roughness testing.

Surface roughness and helium leak were evaluated for each sample after a total of 10 minutes and 20 minutes of acid exposure. Surface roughness data was based on an average of three measurements taken 2 cm away from the sealing surface.

Test Results

All aluminas tested demonstrated some surface roughness degradation after 20 minutes of concentrated acid exposure. (Figure 1). Two of the seven, however, showed only a slight surface roughening. Their performance in the test was significantly superior to the other five samples.

Figure 1 – Surface roughness of 99.8% purity alumina test samples after accelerated acid etching.

One of the two best performers in the roughness tests also exhibited best results in the helium tests (Figure 2). Another sample, not one of the two leaders in the roughness evaluation, performed nearly as well as the alumina offering the highest leak integrity. Test sample No. 2 demonstrated poor results in the helium tests, relative to all the other test samples.

Alumina sample No. 4 was destroyed by an accidental scratch observed to cross the sealing surface. For that reason, 20-minute etching vacuum Helium leak data was not taken.

Figure 2 – Results of helium leak test on polished alumina samples after immersion in hydrofluoric, nitric acid and deionized water.

Summary

Tests have demonstrated clearly that 99.8% purity aluminas corrode differently in strong HF/HNO₃ acid solutions. Those ceramics that demonstrate superior corrosion resistance, that are best able to resist surface degradation and maintain vacuum integrity, can provide an advantage in component performance in many applications dependent on high purity alumina. 

Obviously, how the alumina is processed can make a difference in performance. Many factors need to be considered because no two manufacturers of 99.8% purity aluminas make their material in precisely the same way. There are different techniques, for example, for balancing silica content and engineering grain boundaries. These and other processes should be investigated.

How much of a difference can processing make when the materials are seemingly identical? Refer to the semiconductor industry for evidence of improvement. The industry for some time has been recycling 99.8% purity alumina reactor components, to remove residues, for an average of three cycles. Then the alumina component has to be replaced.

With an improved grade of alumina, those semiconductor manufacturers have reported acid etching an average of 10 times before alumina part replacement is required.

* * *

Carpenter Advanced Ceramics, a subsidiary of Carpenter Technology Corporation, is a vertically integrated manufacturer of technical ceramics, industrial ceramics and high purity alumina- and zirconia-based structural ceramic components.