Failure Analysis of a Stainless Steel MembraneSpeaker - Ed. Ciecko
PhD Candidate - Purdue University
This project is an example of a TAP (Technical Assistance Program) which is a joint program between the State of Indiana and various University departments. TAP programs are designed to assist industry with "small-scale" problems where University facilities and expertise can offer assistance. The "typical" TAP projects takes about 10- 15 man-hours for a student or team of students/faculty to complete.
This specific example related to a failure occurring in an industrial filter cartridge for organic chemicals. Within a cartridge, the chemical is pumped into the ID of an array porous tubes, and filtered chemical discharged through an exit port in the outer cartridge wall. Each tube is composed of a "shell" of 316L, 24% porous Stainless Steel produced by Powder metallurgy, and is coated with a membrane of Titanium Oxide which is the "functional" filter. At the end each stainless tube a wrought 316L stainless fitting is welded which joins with the manifold for distribution of incoming solution. These welds were failing without warning after some period of time (usually greater than two years). The environment is acidic (4.5 PH) and 80 - 108C when the filter is in use, but is cycled to basic (10.5 PH) ambient temperature during a purge cycle of the filter every 6-7 hours.
Students assigned to this project performed the following: 1) Microstructural Analysis; 2) Qualitative Fracture Analysis (with aid of SEM); and 3) Detailed microstructural study of the welds.
Several important discoveries were made:
1) After two years of service the TiO2 coating was virtually gone, rendering the tube disfunctional for its intended purpose.
2) Porosity in the P/M stainless shell increases during service form the 24% initial value to about 35%. (This was determined via a displaced oil/air weight density determination).
3) Fracture is occurring just to the P/M side of the weld, as an entirely brittle fracture. Drop tests confirmed that new P/M stainless shells were ductile, while used ones had become brittle.
4) Shrinkage during welding was leading to both reduction of volume, and large (frequently sharp) porosity on the P/M side of the weld. Sharp porosity at this point is very effective as a crack starter leading to the failures.
5) Service conditions were leading the loss of grain boundry material by some etching phenomenon. This is also contributing to the failures.
6) Given the loss of TiO2 occurs prior to weld failure, this device should not be expected to last more than 2 years.
Parties interested in the TAP program can learn more at:
www.ecn.purdue.edu/TAP