A discussion of detailed destructive metallurgical evaluations of a specific lubrication erosion failure, which the authors have conducted in order to analyze and characterize the failures.

Material science is the oldest of all science and engineering disciplines. It is not an overstatement to say that materials science has transformed human history. The Bronze Age and the Iron Age defined the rise and fall of complete civilizations. The economic might of the industrial revolution was built on the Age of Steel.

During an inspection (to determine reports of higher than normal vibration) of a high-speed gearbox that was driving a compressor, one of the gears was found to have unusual vibration characteristics. Visual examination of the opened compressor box revealed that the low-speed (LS) helical pinion exhibited pit-type defects on each of its teeth. Review of the failed component revealed a somewhat repetitive type of damage at one end of the teeth only (Figure 1). Each tooth showed what appeared to be one defect at a similar location 3600 around the pinion. Each defect was located within ~0.5 inch of the end of the helical tooth. The defects were observed on the coast side of the teeth only. Visual examination of the mating gear revealed no evidence of similar damage.

Figure 1: Photographs of the LS Pinion. (a) View of the LS Pinion in its gearbox location, arrows point to tooth defects (b) Closer view of LS Pinion defects – red arrows (c) Pinion after removal from gear—ready for shipment to DST (d) View of HS Pinion showing possible similar defects (red arrows) on teeth (e) View of replacement LS Pinion after 200 hours operation—arrow point to possible defect (f) Pinion rotation direction, damage on coast flanks.

The failed LS pinion is shown in place and prior to disassembly Figure 1a. Figure 1b shows a close up view of the pitting-type damage observed on each tooth. As a result of these defects, the component was removed from the gearbox for further evaluation, as shown in Figure 1c. It is important to note again at the outset that the damage is on the pinion coast flanks, as Figure 1f shows. In order to characterize the damage, we conducted a detailed metallurgical laboratory failure analysis of the damaged pinion. In this regard, Drive Systems Technology, Inc. (DST) utilized the services of a large local shop, Chalmers & Kubeck (C&K), to assist with the handling, measurement, and sectioning of this component.

After receipt of the LS pinion shaft, we conducted several dimensional checks and a wet magnetic inspection (MPI) of the pinion teeth. In order to facilitate our further investigation, the complete 360-degree toothed section of the pinion was subsequently removed. After that initial sectioning, several of the damaged teeth were individually removed by water jet cutting and metallurgically evaluated.

A brief review of the history of the machine revealed that it had been manufactured in operation for approximately 22 months prior to discovery of the damaged pinion. Throughout the service time for this compressor, there were no problems reported with this machine or signs of gear damage. As previously noted, the machine was shut down as a result of high vibration.

Review of the photographs that we initially received in regards to the pinion failure indicated that the high-speed (HS) pinion used in the gearbox might also have some similar indications located at one end of its teeth as the LS pinion. These are pointed out in Figure 1d by the red arrows on the photograph, but were not part of the subject evaluation at this time. After the subject pinion was removed, a new, identical pinion was installed. This pinion was examined after approximately 200 hours of operation and was also found to have some evidence of similar surface damage as indicated in Figure 1e. The frosted areas shown in this Figure are located in an area that appears to be identical with that of the failed LS pinion, which is the subject of this investigation. This can be seen by comparing the defect characteristics and locations shown in Figure 1a with Figure 1e. We were unable to obtain a detailed drawing of the pinion but we were provided with some basic design details as follows:

a. Contains 47 teeth.
b. Rotates at 21,142 rpm.
c. Manufactured from AISI 4340e “Aircraft quality” steel.
d. Material is through hardened to 320 – 360 BHN (Brinell hardness number). These values represent core hardness values prior to nitriding. We converted these to R/C 33 – 36 using standard hardness conversion tables.
e. Part is nitrided to 0.023 – 0.018 inch in depth. No definition of the term “nitride case depth” was given (nitride depth is typically specified as “Total Case Depth”).
f. Surface hardness requirement is 84 – 88 Rockwell 15N. These values convert to R/C 47 – 55, 4
g. There was some concern that a stray electrical current was present in the system which caused the damage. We did not agree with this assessment as there was no damage to any of the bearings in the system and we usually observe bearing damage before gear damage when electric arcing is the culprit. We did, however, evaluate this possibility as part of our investigation of the pinion damage.

Reference Data

As we do with every failure analysis project, we also reviewed all historical information available with respect to this gear system, especially the operational characteristics and any unusual circumstances. Details are not pertinent here, but we found nothing unusual in these records.

Metallurgical Evaluation Procedure

We used the following procedure for evaluation of the component:

a. Review all data available related to the gearbox and the damaged pinion specifically.

b. In the course of our preliminary visual examination of the pinion, we conducted a detailed dimensional evaluation. We also conducted a magnetic particle inspection of the part to insure no additional cracking was present. No cracks were found. Based on the singular location of the tooth defect on each tooth, a preliminary determination of the best method to remove several tooth segments from the pinion for further metallurgical evaluation was discussed. It was decided to first remove the complete pinion tooth segment from the rest of the shaft, followed by cutting two 90-degree quadrant segments for our work. A significant portion of the remainder of the shaft was then nital etch inspected to determine if any evidence of arc burns was present in any location surrounding the defects. Any indications observed would be used as an indicator for additional locations to be removed for further examination.

c. Preliminary sectioning included removal of the thrust collars and the complete 47-helical tooth section from the pinion shaft. This was accomplished by water jet cutting. This technique supplies sufficient coolant to the part during the cutting operation so that the metallurgical properties of the teeth removed are not affected.

d. After removal of the pinion helical teeth from the shaft, two chord wise segments, containing approximately 10 teeth each, were removed from the pinion. This was done using a water jet machining technique. The segments were labeled ‘A’ and ‘B’. After removal, both segments were prepared for our metallurgical lab investigation. This included determination of surface pitting characteristics and metallurgical heat treat properties, including chemistry, of the pinion component. Both segments ‘A’ and ‘B’ were oiled for corrosion protection and staged for our metallurgical evaluation.

e. A portion of the remainder of the pinion was used for nital etch inspection, i.e., after removal of the two chordal segments. This was accomplished per the latest DST procedure. Approximately 10 adjacent teeth were evaluated.

f. In the lab, the two pinion tooth segments were visually and macroscopically examined for defect/pit locations/characteristics, topographical features, and any apparent contributing factors related to the damage observed on the helical teeth. Digital and macroscopic photographic documentation of the surface defects was accomplished.

g. Several metallurgical cross sections, through about 10 tooth surface defects, were obtained. The samples were removed by EDM machining and were then metallurgically mounted and polished. In addition, another cross section segment through the helical teeth was removed, mounted, and polished in order to characterize the heat treatment features of the pinion. This included determination of surface and core hardness values, nitrided case depth, etc. Another section of the pinion was carefully removed for chemical analysis.

h. Microhardness and microstructural characteristics were determined at and away from the damaged/pitted areas. Chemistry and hardness properties were determined in the area of the non-failed teeth. All microhardness data obtained was plotted, reviewed, and correlated.

i. Based on the results from the data obtained, the cause and nature of the tooth defects was determined and the failure mechanism characterized.

j. During the process, several brief update status reports were written detailing various characteristics of the tooth defects and results obtained.

k. Finally, a complete metallurgical report was prepared with numerous figures, comments, conclusions, and recommendations in order to prevent similar occurrences in the future.

Figure 2: Additional information/photographs (a) Schematic of overall compressor box details (b) View of compressor after removal

Test Results

Visual Examination of the Pinion Tooth Defects at C&K

The low-speed pinion shaft was part of a compressor drive gearbox. An overall schematic along with a photograph of the actual system during removal is shown in Figure 2a and Figure 2b. The as-received failed pinion shaft, after removal from the compressor, is shown at C&K in Figure 3a and Figure 3b. Close up view of the shaft identification is given in Figure 3c. The shaft consisted of two thrust collars, one on either side of the pinion teeth, as shown in Figure 1c. The LS pinion consists of 47 helical teeth (right hand). We removed the thrust collars and then sectioned the pinion teeth from the rest of the pinion shaft as shown in Figure 3d.

Figure 3: View of as-received pinion at C&K. (a) Overall view of pinion shaft (b) Close-Up view of pinion teeth (c) View of P/N S/N details on thrust collar (d) Cut lines to remove pinion teeth from shaft.

As noted, we conducted a dimensional evaluation of the helical teeth. After the dimensional evaluation was accomplished, we conducted a Magnetic Particle Inspection (wet, continuous technique) of all of the helical teeth. This inspection did not disclose any evidence of cracking, either extending from the defects or in any other location on the pinion teeth.

Visual examination of the pinion teeth revealed pit-like defects on every pinion tooth. Each pit was located on the coast side of the tooth. From a face width standpoint, the pits were concentrated within approximately 0.5 inch from the end of the helical tooth. Only one end of the teeth was affected. This is illustrated in Figure 4a, Figure 4b, and Figure 4c. The pitting was located at the top land of the tooth (“1,” Figure 4c), high on the addendum (“2,” Figure 4c) and in the root (“3,” Figure 4c). No pits were observed in the dedendum. The pitting observed on each tooth appeared to be deep and discrete in each location. Some of the pitted regions on the top land and the addendum appear to be “growing” together, but it is clear that they started as separate entities. Each region, however, shares similar characteristics. Since two of the pitted regions were located in non-contacting regions (i.e. top land and root) of the teeth, applied load did not appear to be a common causation factor. Because of the discrete nature of the tooth surface defects/pits, we conducted a nital etch inspection of the pinion teeth. Prior to this process, two 10 tooth chordal segments of the pinion teeth had been removed by water jet cutting (Figure 6) for detailed metallurgical evaluation. A portion of the remainder of the pinion, shown in Figure 5a, was used for the nital etch process.

Figure 4: Initial examination of pinion toothed section. (a) Pinion teeth after removal from shaft (b) Defects on tooth tips and root (c) Enlarged view of damage on top land, coast flank and root.

Nital etching was accomplished primarily to find evidence of arc burn type defects. Arc burns would be indicated by a dark band and/or white band surrounding the discrete defect. The color change is representative of metallurgical microstructural changes that are consistent with material property degradation caused by exposure to excessive heat. Arc burns would be indicative of the presence of a stray electrical current in the system during service life operation.

After setup as shown in Figure 5b, nital etching was accomplished per DST Nital Etch Specification AE 01.4, entitled “Specification for Acid Etch Inspection of Hard Finished Gear Teeth, Rev. C,” dated 6/3/2011. Approximately 8-10 teeth were nital etch inspected for arc-type defects. The area etched is illustrated in Figure 5c. We used a swab etch procedure on this segment. Specific attention was given to the area surrounding each of the defects.

Figure 5: Nital etch examination at C&K (a) Remainder of pinion after two tooth segments removed (b) Nital etch setup at C&K (c) Area of pinion teeth nital etched.

Shown in Figure 7a and Figure 7b are examples of the results of the procedure. In some areas, a dark edge at the edges of the defect can be seen. These are the result of the light being shadowed during photographic documentation of the area. No evidence of white or dark band, typical of a rehardened or retempered burn (respectively), was observed on the surrounding surface contiguous with any defect. Shown in this Figure 7c is a tooth that was not etched to one that was etched. Comparison of the surface characteristics of the area surrounding the defect on this latter tooth with the nital etched teeth shows no significant color difference. This illustrates the results of this inspection in that no nital indications were evident. Overall, the results of this examination disclosed no evidence of any arc burn damage on any of the teeth evaluated. This means that a stray electrical current was most likely NOT present during operation of the pinion component and that the defects are not related to electric arc occurrences.

Two other interesting observations on the teeth were noted during this procedure. These are pointed out in Figure 7c. First, close examination of the root area between the teeth and in line with the tooth tip pitted defects revealed additional localized damage. This damage was characterized by a sponge-like appearance. Second, similar type pitting damage was observed at the root edge of the pinion teeth. These areas are also illustrated in Figure 7c. Both of these additional defect areas were evaluated during our metallurgical evaluation detailed below.

Figure 6: Water jet cutting to remove lab evaluation sections.

Visual and Macroscopic Examination of Fracture Surfaces

a. The as-received teeth segments removed from the pinion are illustrated in Figure 8a and Figure 8b. Visual and macroscopic examination revealed a relatively large irregularly shaped defect at the same location on each tooth as illustrated in Figure 9a, Figure 9b, and Figure 9c. Each defect was located approximately 0.25 inch from the end of the tooth and extended for a distance of approximately 0.150 inch along the tooth face width. From a dimensional standpoint, each defect was located along the tip of the tooth coast side and extended down the addendum of the tooth profile also on the coast side. Each of the defects was concentrated on the coast side of the tip and tooth only. None of the defects extended completely across the tip of the tooth but did encompass up to about 90 percent of the tooth tip land, i.e., up to about 0.070 inch of the 0.080 inch wide land. Extension of the defect down along the profile of the coast tooth profile was observed to a distance of up to 0.100 inch. The tooth profile total length from tooth tip to root fillet was measured at 0.25 inch.

Figure 7. View of teeth defects after nital etch (a) View of several nital etched teeth (b) Close-up view of teeth showing no burns present (c) Comparison view of nital etched teeth (white arrows) next to non-nital etched teeth (blue arrow)

b. Macroscopic examination of the defects disclosed two primary features. One, visual evidence of some form of erosion, i.e., an eating away, of the surface was present. The surface had a spongy appearance. The erosion was characterized by a gradual surface attack at the edges of the defect, which became deeper and more concentrated as the center of each defect was approached. In some cases, the erosion pattern appeared to have an overall sun-burst configuration. Each eroded area was discrete, i.e. singular and concentrated on one side of the tooth only. Second, a relatively deep pit was present in many of the eroded areas. These pits, sometimes also showing evidence of a crevice, extended from the eroded tip of the tooth down into the center/core of the tooth. These characteristics can be seen in Figure 9a, Figure 9b, and Figure 9c. Figure 9d shows a close-up view of a pit type defect showing the sun burst pattern. The sunburst characteristic of each defect is illustrated in Figure 10a. This topographical feature suggests that high-pressure fluid was directly impinging on the coast side tip of the tooth, resulting in significant removal of material from the coast side only. This is illustrated in Figure 10b also.

c. Macroscopic view of several deep pits is shown in Figure 10b and Figure 10c. Close examination of these defects revealed a deep ‘cavernous’ non-uniform design. Multiple hollowed out areas could be seen looking down into the pits. The pits appeared clean, with no evidence of debris or corrosion present. Each pit was very rough in topography. Certain areas appeared to have what appeared to be spherical nodules present. Some of these can be seen by close examination of Figure 10. With the proper lighting, the pitted/eroded area exhibited a somewhat shiny appearance. This is shown in Figure 10a, Figure 11a, and Figure 11b. As indicated above, each pit was surrounded by an area of severe erosion.

Figure 8: As-received segments removed for metallurgical evaluation. (a) Segments labeled A and B—arrows show defects (b) Another view of segments A and B with defects pointed out (c) Every tooth has defect at similar location along tooth length.

Macroscopic evaluation of the defects at higher magnifications revealed that a significant amount of tooth material had been removed from the deep pitted areas. The roughness of the surface suggested that the material most likely had been ‘eaten out’ from the tooth nitrided/core areas forming the cavern type features. This is clearly illustrated in Figure 11a, which shows the eroded surface and the deep cavernous pit (which is out of focus) while Figure 11b shows the base of the pit (in focus). Because of their depth, the pit base could not be photographed with the adjacent surface erosion, i.e., with both areas in focus. Overall, the deep, cavernous, clean pits surrounded by the eroded surface are indicative of a ‘cavitation erosion’ mechanism of failure. It was most likely caused by a highly pressurized fluid/lubrication stream present in the compressor system, which impinged on the coast side tip corner only of each tooth.

d. Additional macroscopic examination of the defect areas showed another feature of this failure mechanism. Shown in Figure 12a are two defects on two adjacent teeth that have a semi-circular eroded area at the top of the pinion tooth. These are illustrated by the black arrows in the Figure. The tooth at the top of the photograph shows the presence of a deep cavernous pit (white arrow). What is notable here is that there is also erosion damage in the root area between the two eroded teeth (red arrows). This area is out of focus in Figure 12a but in focus in Figure 12b and magnified in Figure 12c. Again, DST noted that the eroded area exhibited sponge-like features. A sponge like surface attack is also considered typical of a cavitation erosion failure mechanism.

Figure 9: Close-up view of tooth defects (a) Photo shows area of erosion and deep pit (b) Enlarged view of erosion and pit areas (c) All deep pits were located in eroded areas (d) Magnified view of deep pit and sun burst pattern of erosion.

Furthermore, another eroded area was observed in the root area at the edge of the tooth as shown in Figure 13a, magnified in Figure 13b. These photographs were taken from tooth sections that were removed during the preparation of test samples A-1 to A-3 as detailed later. The characteristics at these locations are indicative of the amount of erosion attach that took place during the operation of the pinion component while in service. These Figures tend to indicate that the impingement of the high pressure fluid stream ‘splashed’ over the tooth tip corner and down into the coast side root fillet. Shown in Figure 13c is a metallurgical cross section A-1 through the pit type defects showing erosion damage at both the tip of the tooth and in the root on the coast side of the tooth only. It is clear that the damage in the tip of the tooth is greater than that in the root area indicating that the maximum impact of the cavitation erosion mechanism was greatest at the tip area of each tooth.

e. Study of the pit type defects disclosed no evidence of fatigue propagation. No beach marks or other fatigue characteristics were present. This data indicated that the defects were not related to normal tooth pitting and spalling. Further, the phenomena observed on each pit did NOT typify electrical arc type defects even though a significant amount of material was missing from the tooth in the pitted areas. Due to the internal depth of the pits, their cavernous appearance, and their location in an eroded area, none of the defects were considered characteristic of an electrical discharge problem.

Figure 10: Additional views of cavitation erosion damage on coast side of tooth (a) Another view of sun burst pattern showing surface erosion damage and pit (b) View of damage suggesting impingement of concentrated lubrication stream on tip corner of coast side(c) View of tip of tooth showing pits’ deep cavernous feature.

f. Macroscopic examination of the pinion tooth contact pattern revealed that the cavitation erosion type defects were located within this zone. This is illustrated in Figure 14. It was apparent the pattern extended to within 0.060 inch of the end of the tooth face width. It was interesting to note that the tooth contact pattern was clear on opposite sides and ends of each tooth rather than on both sides of each tooth. This could be indicative of some form of misalignment of the pinion while in service.

g. An evaluation of the surface finish of the teeth was accomplished using a GAR Surface Finish Comparator. This visual comparison technique yielded a surface finish value equivalent to 63 RMS which was produced by a shaping method of tooth profile manufacture. Examples of the tooth profile surface characteristics are illustrated in Figure 9e, Figure 12a, and Figure 14.

Figure 11: Magnified views of cavitation erosion pits (note shiny surface of erosion) (a) Deep pit (out of focus) in center of erosion damage (b) Bottom of deep pit (in focus) cavernous pit is clean and rough in topography.

Overall, visual and macroscopic examination of the singular defects observed on each pinion tooth at the same location indicated that each was the result of a cavitation erosion mechanism of failure. The eroded surface surrounding deep pits that extended into the center of the tooth area in a cavernous fashion is characteristic of this type of failure. The eroded surface was discrete and located at a similar location tooth to tooth. The characteristics of the pits also confirmed this finding, in that the pits were rough and cavernous in topography; clean, with no evidence of corrosion products observed; and deep in size. DST also noted that the erosion was observed in the root area on either side of the pitted surface. No clear evidence of electrical arc type damage or any other type of metallurgical defect was observed on any of the tooth defects metallurgically examined.

Figure 12: Views of erosion in root adjacent to tooth erosion/pit damage (a) View shows two adjacent teeth—red arrows indicate root erosion damage (b) Erosion damage, in focus, is concentrated on coast side root area (appears spongy) (c) Magnified view of spongy looking erosion damage in root.

Hardness, Chemistry and Microscopic Examination of Pinion Shaft and Defects

Three cross sections through the teeth defects from pinion Segment B were removed for determination of certain metallurgical properties. These samples were cut perpendicular to the pinion axis and were prepared such that the defects could be polished into. This was done in order to determine the microscopic characteristics of the cavitation erosion defects. Each sample was then evaluated in the unetched condition for microhardness traverse data. Subsequently, the samples were chemically etched for microstructural features both at the surface and below the surface. This included evaluation for any evidence of reharden and/or retemper burns. The three test samples were labeled ‘A-1’, ‘A-2’, and ‘A-3’. The areas removed for evaluation contained three teeth each, as shown in Figure 15.

Figure 13: Additional views of root erosion damage obtained from cross sections through pits. (a) Erosion damage observed on end of tooth (b) Magnified view of erosion at end of tooth—coast side (c) View of metallurgical mount shows defects in tooth tip and root located on coast side of tooth only.

Also shown in Figure 15 is a fourth test sample, labeled ‘B-1’. This test sample was removed from an area that was defect-free as illustrated in the Figure. It was prepared by removing a cross section perpendicular to the direction of the helical teeth. This sample was used to determine the actual metallurgical characteristics of the pinion component. It included case and core hardness properties, nitride case depth, and microstructural features of the pinion shaft teeth.

After removal of the all four areas by EDM machining, each was metallurgically mounted and polished. A polished cross-section through one of the ‘A’ test samples is illustrated in Figure 13c. All of the ‘A’ samples looked similar to that shown in the figure. After metallurgical preparation of all four test samples, a small additional area was removed for the purpose of determining the chemical makeup and quality level of the pinion shaft material. A brief summary of the test samples is given in Table 1.

Figure 14: View of cavitation erosion defect within tooth contact pattern.

a. Hardness Evaluation

1. Each of the four tooth segments was carefully removed from Tooth Segment B by electro discharge machining. Subsequent mounting and polishing was accomplished using standard metallurgical techniques. Samples were water cooled during cutting and polishing to assure that properties of the material would not be altered. Each sample was evaluated for microhardness characteristics at the surface as well as the core. All microhardness traverses were accomplished using a Leitz Microhardness Test Machine. Hardness traverse evaluation was accomplished at depths below the surface of 0.003”, 0.005”, 0.010”, and then every 0.010 inches up to 0.050 inches. DPHN values obtained were converted to Rockwell C values using standard conversion tables. A polynomial fit to the hardness vs. depth curves was used for all test samples to determine the nitride case depth. The nitride case depth was determined as the depth where the hardness dropped to 110% of the core hardness. This hardness value was determined as R/C 36. In addition, each sample was evaluated for surface and core hardness and decarburization if present.

2. The microhardness results for test samples A-1, A-2, and A-3 are illustrated schematically in Figure 16a, Figure 16b, and Figure 16c respectively. Data obtained from these curves are illustrated in Table 2. The results for test sample B-1 is also illustrated in Table 2 and shown schematically in Figure 16d.

Table 1: Test samples removed from pinion tooth segment B for further metallurgical evaluation

3. Based on the hardness vs. depth curves, each tooth surface had been nitrided. Hardness values obtained on each test sample were consistent with each other and typical of heat treated and hardened material.

4. All nitride case depth values (obtained at R/C 36 depth) met or exceeded the requirements given in Reference G (shown in the background section of this DST report and listed in Table 2. The nitride case depth values obtained on all of the samples are considered consistent with heat treated nitrided 4340 steel. Surface and core hardness values also met the requirements.

Table 2: Microhardness vs. depth below surface data

5. Comparison of the data shown schematically for test samples A-1 to A-3 with test sample B-1 shows that the values obtained from the latter sample was more consistent than that of the former. The reason for this is the cut angle of each test sample. Sample B-1 was cut on a perpendicular to the angle of the helical tooth resulting in consistent values on either side of the tooth profile. Samples A-1 to A-3 were cut on a perpendicular to the axis of the pinion shaft and, as such, were not cut perpendicular to the helical tooth angle. These samples had a skewed cross section angle through the nitride layer and thus showed one side of the layer being somewhat greater than the other. This was characteristic on all three samples. Again, the reason for the cut/mounting angle on these latter samples was not for nitride layer evaluation but rather to obtain a clear cross section through the pit type defects present on each tooth.

6. Core hardness values were slightly lower than the requirement. However, this most likely is a function of the carbon level or the base material and the tempering temperature used during the hardening cycle for the 4340 through hardened steel.

Table 3: Chemical analysis of LS helical pinion (all values listed in percentage)

7. Surface hardness values met the minimum hardness requirement and were considered typical for nitrided 4340 steel with low core hardness.

8. Overall, the hardness data obtained from samples was consistent with through hardened and nitrided 4340 steel.

9. Chemical analysis was obtained by Direct Reading Atom Emissions Spectroscopy. The data disclosed that the material conformed to the requirements of AISI 4340, UNS G43400. This material is a typical through hardened steel, which is acceptable for nitriding. The results are illustrated in Table 3. Brief review of the chemical data disclosed that the quality of the material was consistent with vacuum degassed steel. This was indicated by the relatively low phosphorus and sulfur contents.

Figure 15: View of four areas to be removed from tooth segment B for further metallurgical evaluation.

b. Microstructural Evaluation

Microstructural evaluation of each test sample was accomplished in the unetched and etched condition. The etchant used for this work was 2% Nital. All of the cross sections detailed above were evaluated. The following characteristics were observed:

1. A typical test sample cross section is illustrated in Figure 13c. Examination of test samples A-1 to A-3 disclosed that the tip of each tooth on the coast side only was severely eroded. In many areas, extending from the eroded surface was a relatively deep cavity as shown in Figure 17a. The depth of this cavity was estimated to be at least 1/3 of the overall tooth height dimension (~ 0.25 inch). It was apparent that the erosion to the tooth tips occurred on about 80% of the tip land on the coast side. None of the teeth showed any damage on the tip land on the load side of the tooth. In most cases, the tooth tip on the coast side of the pinion tooth had been completely eroded away, as shown in Figure 17b. The erosion damage extended from the tip of the tooth down along the coast side of the tooth, thus destroying any evidence of the addendum tooth profile or the corner of the tooth tip. This is illustrated in Figure 17b and Figure 17c. Closer examination of a cavity formed in the eroded area revealed that it extended from the tooth tip down in towards the center thickness of the tooth. At the bottom of the cavity, secondary cavities along with cracks extending from these were observed. This is shown in Figure 17d. This cracking reveals the method of the continuation of the growth of the cavity. Growth of the cavity through small, sharp cracks extending from the base of the large/primary cavity is most likely part of the cavitation erosion failure mechanism. Had service life continued the growth of the large cavity most likely would have progressed completely through the tooth.

Figure 16. Hardness vs. depth curves for all test samples (a) Schematic representation of hardness data from test sample A-1 (b) Schematic representation of hardness data from test sample A-2 (c) Schematic representation of hardness data from test sample A-3 (d) Schematic representation of hardness data from test sample B-1.

2. Evaluation of the root area between the eroded teeth disclosed evidence of further erosion and the formation of additional cavities. This is illustrated in Figure 18a. The surface topography of the eroded area was characterized by somewhat rounded features with small crack type defects in between. This is illustrated in Figure 18b, which shows a relatively large cavity which has formed at the base/center of the eroded root. These photographs confirm the presence of cavitation erosion in the tooth root areas as shown in Figure 12b and Figure 12c.

Figure 17: Metallurgical unetched cross sections through defects (a) View of typical metallurgical sample showing depth of pit (b) Magnified view of erosion damage observed on tooth tip (c) Significant erosion damage on another tooth with large, deep cavity partially shown (d) View of large cavity shown in Figure 16c showing growth characteristics—note erosion topographical profile.

3. The cavitation erosion profile in the root was similar to that observed on the tooth coast side and tip of the pinion teeth as pointed out in Figures 17b, 17c, and 17d. The pattern is also shown in Figure 20a. Microscopic examination, using higher magnification of the profile is shown in Figure 19b and Figure 19c. Clearly the profile resembles a somewhat spherical, honeycomb appearance. The honeycomb profile observed on the affected teeth of the LS pinion is further confirmation of a cavitation erosion mode of failure.

Figure 18: Magnified views of cavitation erosion in root area. (a) Erosion of present in root area shows formation of cavity (b) Large cavity present in root area—note erosion profile.

4. In addition to the above, microscopic examination of the unetched test samples disclosed the following:

i. Cleanliness evaluation (inclusion characterization) of each of the test sample appeared to be consistent with vacuum degassed melted material.

ii. No evidence of intergranular oxidation (IGO) was present along the gear tooth  profile of any test sample.

iii. No other metallurgical anomalies were observed on the test samples that could contributed to the cavitation erosion failures observed.

Figure 19: Magnified profile characteristics of erosion. (a) ypical tooth profile showing erosion profile (b) Honeycomb features of erosion profile – typical of cavitation erosion (c) Magnified view of honeycomb profile present in eroded areas.

5. Three of the four test samples were chemically etched using 2 percent Nital etchant. Microscopic examination of the eroded areas on the teeth showed no evidence of rehardening or retempering (see Figure 20a and Figure 20b). Rehardening, which would appear white in color with no evidence of structure within it, would indicate that the defects could have been a result of an electrical discharge, i.e., arc burn. No evidence of arc burns was observed in any of the eroded areas. In order to confirm the lack of rehardening in the eroded areas, magnified views of the tooth damage shown in Figure 19b were metallurgically evaluated. These are given in Figure 20c and Figure 20d. Each Figure shows the same area as in Figure 20b at higher magnifications. Clearly, no evidence of rehardening is present. The microstructural characteristics in the eroded root areas yielded similar results in that no evidence of rehardening was present. A typical root area is illustrated in Figure 20e.

6. Review of test sample B-1 was accomplished after etching in 2 percent Nital etchant. Microscopic examination of the nitrided case microstructure disclosed the presence of tempered martensite and nitride needles. These needles are characteristic of a nitrided layer and are shown in Figure 21a (white arrows). Evidence of the nitride layer can be seen in this Figure. No evidence of any surface defects, such as decarburization of the surface, was observed. The core microstructure, shown in Figure 21b, consisted primarily tempered martensite which is typical of properly heat treated 4340 steel.

Figure 20: Metallurgical etched cross sections through defects. (a) View of eroded tip with cavity formation present (no rehardening present) (b) Typical erosion characteristics on flank-tip erosion—area shown is not typical of electrical arc damage (c) Magnified view of flank erosion shown in Figures 19a and 19b – no rehardening present (d) Additional magnification view of tooth tip corner eroded area – arrows show rough edge with no rehardening present. (e) View of root area erosion pattern – similar to that in Figure 19.

Microscopic examination of the root area between the teeth on sample B-1 revealed presence of some white layer. Apparently the teeth were machined after nitriding, which removed the white layer along the flank profile. This is illustrated in Figure 21c. Also shown in the figure is the presence of white layer in the non-machined root. Another view of the white layer present in the root area is shown in Figure Figure 21d. The white layer was measured at less than 0.001 inch.

Overall, the microhardness, chemistry and microstructural data obtained appeared to be indicative of nitrided and hardened 4340 steel. The removal of the white layer along the flanks of the teeth is considered good practice and indicates that the white layer could not have contributed to the surface damage observed on any of the pinion teeth.

This two-part article presents a discussion of the causation, diagnosis, and metallurgical failure investigation of a lubrication erosion failure. Part II of the article will be published in the April issue of Gear Solutions.

**Printed with permission of the copyright holder, the American Gear Manufacturers Association, 1001 N. Fairfax Street, Suite 500, Alexandria, Virginia 22314.  Statements presented in this paper are those of the authors and may not represent the position or opinion of the AMERICAN GEAR MANUFACTURERS ASSOCIATION.

Previous articleONLINE EXCLUSIVE: Gear Molding – Where It’s Heading
Next articleONLINE EXCLUSIVE: Helical or Not Helical? How Helical Gears Impact the Performance of Planetary Systems
is the Gear Division foreman at Chalmers & Kubeck, a large machine shop/industrial/commercial equipment repair facility. He has over 25 years' experience in the power transmission repair and manufacturing field and has assisted DST in several Failure Analysis case studies.
is chief engineer of Drive Systems Technology, Inc., a Mechanical Power Transmission Consulting Organization that he founded in 1976.  He holds a Master of Structural Engineering degree from Drexel University (1980), a Bachelor of Mechanical Engineering Degree from The City University of New York (1967), a Master of Engineering degree from Pennsylvania State University (1973), and is a registered Professional Engineer in the States of Pennsylvania and Minnesota.
has worked in the failure analysis field for almost 50 years evaluating numerous failed components in the aerospace, commercial, and industrial areas both at Boeing Helicopters and Drive Systems Technology.  Prior to his retirement from Boeing he was manager of materials engineering, where he was involved in the development, testing, and heat treatment  of various types of gearing materials. He has written and presented numerous papers on  failure analyses and ways to optimize the heat treatment of gear materials. He holds both BS/MS degrees from Drexel University in Materials Engineering and prior to retirement was very active in several industrial  technical committees.