Sherry Labs News Story
Fastener Failure Analysis: Fracture Failure of Countersunk Flat Head Screws Case
By: Admin on April 19, 2012
Fastener Failure Analysis, Case Study: Fracture Failure of Countersunk Flat Head Screws appeared in the April/May 2012 publication of Fastener Technology International.
We have the opportunity to view an analysis conducted on behalf of a European appliance manufacturer experiencing fracture failure of countersunk flat head screws. Language barriers across continents can create communication issues. Fortunately, most Engineers are very skilled professionals and communicate very well in English. Despite those pluses, our Metallurgical Engineer Analysts take special care to take into account any nuances in phrasing and nomenclature that vary from what they normally encounter. This is often evident in values of measurement. An example in this case is where torque is expressed in Newton meters (Nm) versus our more familiar value of foot pounds or inch pounds. This is not a major problem as long as we take extra care to be thorough in all communications regarding the analysis we present.
Our job is to ensure the client is comfortable with how we present our findings. Now let's dive into the case presented.
Material to Be Evaluated
Three bags of T40 M8 x 27 flat head, countersunk screws were received for metallurgical analysis. We labeled the samples as Bag-A, Bag-B and Bag-C for reference. The screws were identified by the client as T40 M8 x 27 flat head countersunk screws and were used in the manufacture of clothing washers in three plants. Screws are typically torqued to 28 to 32 Nm at each plant, but one plant reported some failures occurring as low as 2 to 3 Nm. Until recently, 80 million screws had been used in this application with no reported failures. It was
also noted that a new vendor was recently added to provide T40 M8 x 27 screws. Failure issues had not occurred until the new vendor’s product had been introduced.
Our client provided the following material and product requirements. Material grade: C21 grade steel (no specification cited). Hardness, core: 365 to 435 HV. Hardness, surface: 550 to 700 HV. Plating, copper strike layer: 3 to 5 μm. Plating, nickel layer: 5 to 7 μm.
The nonfractured screws had the following approximate dimensions: Length, overall: 27.2 mm. Length, threaded section: 16.5 mm. Diameter, head, maximum: 14.8 mm. Diameter, threaded section: 8 mm; Diameter, screw end: 6 mm.
Visual Examination—Each of the sample screws was examined visually using a stereomicroscope at magnifications of 10X to 70X. Samples from Bag A contained two fractured screw heads identified as Samples A2 and A3, and one screw that had also been removed from the appliance, and was identified as Sample A1, which did not fracture. No defects were observed on Sample A1. However, the screw had been tightened sufficiently to score the base of the screw head exposing the copper plating. Sample A2 fracture exhibited a darkened surface with spots of rust and the surface had a rocky appearance with no evidence of fatigue.
The fracture originated at a thread root. After removing the rust from sample A3, it too exhibited a rocky fractured surface originating at a thread root with no evidence of fatigue. Samples from Bag B were identified as Samples B1 through B5. Examination found no external deficiencies. It was also noted that the design and finish of the B samples were identical to the A1 sample, indicating that they were very likely manufactured by the same vendor.
Samples from Bag C were identified as Samples C1 through C5. Examination of the C samples also revealed no external deficiencies. Two features in particular differentiated the C samples from the A and B samples (see Figure 1, Figure 2 and Figure 3). The tip ends of the screws for Sample C were simply fractured after forming the screw. In contrast, screws from Samples A and B were nicely finished with flat ends perpendicular to the screw length. Another distinguishing feature revealed the head rim on Sample C screws was nearly twice as thick as those on Samples A and B. These characteristics should allow screws from the two vendors to be quickly identified as to their manufacturer.
Chemical Analysis—Chemical analysis was performed on a sample from each bag of screws. For Sample A, the analysis was performed on the fractured screws. The results were compared to the C21 grade as specified by Italian and German specifications.
High manganese content of the samples precluded conformance of the screws to Italian C21 specification per UNI 6922. However, none of the samples conformed explicitly to the German C21 grade relative to DIN 2528. The hydrogen content of the fractured A samples was significantly higher than that of B or C samples. A complete comparative matrix of chemical composition of all samples was submitted with the report.
Scanning Electron Microscopy/Microanalysis—The fractured surfaces of Samples A2 and A3 were examined before and after cleaning to remove oxides with the Scanning Electron Microscopy (SEM). Elemental analysis was performed by standardless Energy Dispersive X-ray Spectroscopy (EDS) in conjunction with SEM. Both Samples A2 and A3 exhibited similar fracture morphologies, as shown in Figure 4. The fractures were primarily intergranular, which is characteristic of a brittle fracture mechanism. In addition, micro-pores and “grain boundary yawning,”(i.e., intercrystalline subsidiary cracks) were evident throughout the fractures. All of these features are characteristic of hydrogen embrittlement, also called hydrogen induced cracking.
All evidence and test results indicate that the screws identified in the A and B bags failed as a consequence of hydrogen embrittlement. Hydrogen embrittlement was confirmed by SEM examination that was characterized by intergranular fracture, micro-pores and grain boundary yawning. In addition, chemical analysis disclosed a significant increase in hydrogen for the failed samples compared to the unused samples. Brittle fractures were also observed with the unused screws of the A and B samples, even though they had not been placed
Significantly, brittle cracking was observed near the head in the thread roots of the unused samples as well as in the head recess. In contrast, no fractures were found with the Bag C screws.
Hydrogen embrittlement often occurs during plating operations when hydrogen is produced on the surface of the part as plating proceeds. Hydrogen can diffuse into the surface and become entrapped at crystallographic dislocations, small pores and other subsurface defects.
Hardened structures are more susceptible to hydrogen embrittlement as are parts with thicker platings. It was noted that the failed screw heads had plating thickness nearly twice that of the bag C samples and that the nickel plating exceeded the maximum plating thickness of the specification.
The samples were similar in composition and were manufactured from medium-carbon steel similar to C21. However, none of the samples met the chemical specifications of the Italian or German C21 grades.
It was also noted that the A and B screws had higher manganese and silicon contents than the Bag C screws. Nonetheless, compositional variation had no significant affect on the failure mechanism.
The T40 M8 x 27 screws failed as a consequence of hydrogen embrittlement. This failure is undoubtedly related to an inadequate post-plating heat treatment, known as “baking,” that releases hydrogen from the matrix and restores ductility. Hydrogen embrittlement was identified with A and B bags.
The vendor of the A and B samples and the vendor of the Bag C samples can be identified by differences in screw design including the screw ends and the shape of the head rims. It was recommended that screws produced by the vendor identified as bag A and B be removed from inventory. It may be possible to salvage these screws, but they will require an extended baking operation to ensure that hydrogen is adequately removed from all samples. The vendor should also institute corrective actions to ensure that future screw shipments are provided with a sufficient baking operation.
It is not possible to predict how many screws may have been improperly processed without exact knowledge of the vendor’s procedures and operations. Susceptible screws may be limited to a single batch that was improperly cycled. It may be that if baking was performed, the furnace was overloaded and a portion of the batch did not have sufficient time at temperature to completely remove the hydrogen.
Good manufacturing practice for nickel-plated parts includes baking for 24 hours at 190°C to 200°C. If using racks or trays, the layer height should not exceed 50 mm. Screws should be baked within four hours of plating. Baking is most effective when performed within one to four hours after plating. After that period, it becomes more difficult to remove hydrogen. Recommended baking time for parts that exceed the four-hour time limit after plating would be 48 to 96 hours. This process should ensure that all parts would be free from hydrogen embrittlement issues.
Care must be taken to never jump to a conclusion before a full analysis is done. In this case, a visual examination of the samples with the naked eye would have concluded that the problem batch would have been the samples in Bag C due to the screw end configuration. This of course turned out to be the problemfree batch due to adequate baking during heat treating.
Whenever a new vendor is selected, it is cost effective and prudent to submit samples to an independent laboratory to conduct inexpensive material qualification certification testing before final selection. This step can save countless dollars and problems going forward.
We welcome the opportunity to provide you a quote for our services or discuss how we can help.
News Tags: failure investigation failure analysis services failure analysis lab fastener failure analysis failure analysis countersunk flat head screw failure fracture failure analysis chemical analysis SEM-EDS analysis hydrogen embrittlement