Investigations on Additively Manufactured Stainless Bearings

Round 1

Reviewer 1 Report

In this paper, the authors applied Additive manufacturing by means of plasma-transferred arc welding to develop a new material system consisting of Rockit® (0.52 wt.% C, 0.9% Si, 14% Cr, 0.4% Mo, 1.8% Ni, 1.2% V) as cladding material on ASTM A572 as substrate. This material possesses High fatigue-resistant rolling bearing raceways. It was achieved that, the high hardness and strength of the cladding material is attributed to the high quantity of carbides, which are distributed throughout the cladding microstructure. Compared to the previously studied hybrid AISI 52100 - AISI 1022M bearing washers, the hybrid Rockit® - ASTM A572 bearing washers are characterized with significantly higher compressive residual stresses over a large depth range. 

 

This is a clear, concise, and well-written manuscript. The introduction is relevant and theory based. Sufficient information about the previous study findings is presented for readers to follow the present study rationale and procedures. The text is clear and easy to read, and the results are sufficiently discussed. Overall, the manuscript is well thought out and written, the objectives clearly stated, experimental methods are advanced, data statistically analyzed, the conclusions well supported by the data presented. In my opinion, the manuscript is suitable for publication as it is.

Author Response

Dear Reviewer,

We thank you for reviewing the initial version of the manuscript and for providing positive evaluation of our manuscript.

Reviewer 2 Report

The article for review, entitled “Investigations on additively manufactured stainless bearings” concerns the evaluation of an alternative method to increase the service life of rolling bearings. The authors used the powder plasma surfacing process for this purpose. A powder with the trade name Rockit was selected as the coating material. It is a material with a composition corresponding to martensitic stainless steel. The substrate was ASTM A572 steel. The coating and the base material were heat treated by martensitic hardening and low tempering. And then mechanical shaping (turning) and finishing, grinding and polishing. The article assesses the impact of the proposed technology on the microstructure, the distribution of chemical elements, the state of internal stresses in the coating and the substrate, hardness and corrosion resistance obtained as a result of processing a new material system. The literature review in the introduction was based on 43 source materials. The cited literature is sufficiently up-to-date. The introduction itself introduces the reader to the issue raised in the article very well. The research methodology proposed by the authors is adequate to the purpose of the research. I have no major comments on the way of presenting the obtained research results and their discussion.  The conclusions are correctly formulated and follow from the content of the article. My only suggestion is that the conclusions should be technology related and not sample related.

Below are some of my detailed comments:

 1) In my opinion, the paragraph in the lines between 135-157 should be included in the Introduction chapter.

2) Table 3. Please explain the concept of "Load equivalent" in the article

3) Figure.5.  From the hardness change characteristics, you can determine the thickness of the Rockit material layer. Were the samples used for these tests different thicknesses of the cladded layer?

4)  Figure 11. Which of them is 11a or 11b has not been indicated in the figures. You have to figure it out.

5) Verse 4 and 504 The sum of the stated components of the Rockit alloy is not 100%. I'm not sure all readers will guess that the rest is iron.

Author Response

Dear Reviewer,

The authors wish to thank you very much for your constructive comments. We would like to respond to the comments point by point as follows:

1. Thanks for the advice to focus the conclusion more on the technology. We have implemented this where it was possible.
2. Thank you for this comment as well. However, we do not want to implement it because we would like to focus on the concept of additively manufactured rolling bearings (or raceways) in the Introduction. In the chapter in question, however, we deal with the basic principles of material selection for our approach. A brief description of the state of the art for the selected material therefore seems more appropriate for us at this point.
3. The load equivalent, load ratio, or dynamic load safety factor C/P, is the ratio between the basic dynamic load rating of the specific bearing (here: C= 137 kN) and the actual equivalent dynamic load P of the bearing during operation. This information has been added at the corresponding passage in the manuscript on page 7.
4. This is correct. Cladding layer thickness, however, was specified by microscopy on different cross sections. The samples were taken from the actual batch and the data was verified on multiple parts. This information has been added on page 9.
5. The label in the picture was added.
6. Thank you, due to routine-blindness we would not have noticed. We have added iron as balance when specifying the element composition.

Reviewer 3 Report

1.       A few typos to be correct (e.g., missing close parenthesis line 125, missing “ly” on approximate line 517, “dont” -> “do not” line 518). Just needs a final proofread I think.

2.       It also seems like the authors use a mix of “.” and “,” to denote thousands. They also sometimes use “.” to denote decimal. Consider using the SI and US-NIST standard notation of a thin space between the third and fourth digit to avoid potential confusion. For instance, in the same paragraph you have “temperatures between 1.000 C and 1.150 C” and “0.01 s^-1” – read semantically, these could easily be misinterpreted.

3.       How was composition measured? Is this of the powder/precursor or built material? This is not described sufficiently in the manuscript.

4.       Relatedly, were trace amounts of “contaminant” elements measured? E.g., N might exist in small quantities and act as an austenite stabilizer, but was unquantified. Same for O and others.

5.       The authors mention that a 1:5 ratio of two different composition powders were mixed, what is the justification for this? Furthermore, how was the mixing accomplished and can the authors guarantee that local concentrations do not vary from the bulk mixing ratio? For instance, on the macroscale, location-to-location variability may drive differences in behavior from that observed by the very small scope of the data captured by nanoindentation maps.

6.       The resolution of the SEM images seems rather poor. Could we see the “high magnification” images that the author mention in the discussion section to better show the martensitic and carbide structures?

7.       The specifics of tool motion strategy are likely important to the achieved material (at least pre-heat treatment, but potentially after heat treatment as well). Please expand the description and if possible, provide a diagram and/or appendix with full control plans.

8.       Corrosion testing was reported, but it is unclear how this relates to the rest of the work – were fatigue tests conducted under non-ambient environment? If so, this was not explained well enough. Or consider removing the tests and results about corrosion as off-topic.

 

Author Response

Dear Reviewer,

The authors wish to thank you very much for your constructive comments. We would like to respond to the comments point by point as follows:

1. Thank you for correcting the typing errors. Our proofreading by a native speaker was probably not perfect yet, so we had the text proofread again by another professional.
2. Again, thank you for pointing this out. Because of the different authors with their own habits, we have now consistently followed your recommendations. The corresponding values in the manuscript and in the figures have been changed.
3. The spark spectrometer ”SPECTROMAXx” (SPECTRO ANALYTICAL INSTRUMENTS GMBH) was used to determine the alloy content of the cladded material after finishing. The values coincide with the precalculated mixing ratio of the powders Rockit 706 and Rockit 401. This description has been added on page 5.
4. Table 1 already considered the contaminant of the elements P and S. The containment of N is less than 0.05 wt.%. In general the values of the accompanying elements correspond to the DIN EN 10088-3 (Technical delivery conditions of corrosion resisting steels for general purposes). This also has been added on page 5.
5. The ratio of 1:5 was chosen in order to obtain a corrosion resistant alloy. Therefore, the content of chromium must be at least 12%. Due to the risk of intercirstalline corrosion, the chromium content was set higher to 15.7%. Of course there are different local concentrations, due to the fact that at the grain boundaries the carbide Cr₂₃C₆ was indicated. As a direct result of these carbide precipitations, the chromium content nearby the grain boundaries decreases below 15.7%. In an other paper, which is not yet published, we focus on these local concentrations. The mixing was accomplished before the welding process started, whereby the powders were weigthed and put in a mixer for powders.
6. Due to the compression of the PDF file for reviewing, the images seem to have been rendered at a lower resolution. Please compare the source file "Gefüge Rockit ohne Umformung.pdf" with >600 ppi in the Latex directory. We will check for a high resolution of all figures during the final proofreading of the file to be published (hopefully).
7. This has been specified in section 2.4 on page 5.
8. The purpose of the corrosion test was only to show that the material is resistant to corrosion, as it was required during material selection. Fatigue testing was carried out under “normal” conditions, i.e. non-corroded specimens after finishing. This has been clarified on page 11.

Reviewer 4 Report

It is hard to judge this paper for a special edition of protective coatings for tribologically loaded surfaces, because there is almost no  work done on tribological properties of the surfaces.  The paper remains shows a lot of material characterisation, but when it comes to a single tribological property, i.e. fatigue resistance in a modified FE8 setup, it is not clear what the results for these coatings are.  As the conclusions state, the quantification of fatigue life improvement was not possible.  So there is no tribological information in this paper.

Suggest to publish elsewhere.

 

Author Response

Dear Reviewer,

While we very much appreciate your feedback, we respectfully disagree with the statement.

The description of the Special Issue explicitly addresses deposition-welded coatings under tribological load, which was in focus of our research. Our essential finding, that the coating material we have researched has a longer fatigue life under cyclic loading of a tribological contact than the current standard for bearing steel, is due to the properties of the tribological system. Simply, if other components of the tribological system fail, then the coating presented here is more advantageous for the tribological application. If the tribological load would’ve been directly applied to the base materials surface, failure would be expected after very low number of load cycles (<<10^{^4}) due to comparatively low fatigue strength. In fact, it is more likely that plastic deformation would have already occurred during the application of the load due to the low hardness. All of these characteristics are clearly properties of the tribosystem – and therefore related to the coating material. The determination of explicit tribological parameters of the coating material, such as coefficient of friction or specific wear parameters, is currently being carried out in other studies and would also be dependend on the tribosystem as a whole. Thus, this is not the subject of the research here and is not possible with the test methodology presented. For example, the total friction torque of the FE8 test head did not differ between AISI 52100 and Rockit. Since both are steel materials, this would not be expected either. The weighing of the specimens before and after the test for determination of possible weight loss, which is mandatory for FE8 testing, did not provide any information during the evaluation and was therefore not included.

But we have included the above considerations in the conclusion. It is absolutely correct that, in case of doubt, a more in-depth analysis with further tests is necessary for quantification, and we also presented this in the conclusion. However, since this could not be foreseen in advance, we here present the step that was necessary beforehand.

The authors thank you again for accepting to review this manuscript and we appreciate the time and effort you have put into it. In the present manuscript, however, it is not possible for us to further implement your comment. If you have constructive suggestions on how we can address your criticisms with accompanying research, please let us know.

Round 2

Reviewer 4 Report

Thank you for your explanation.  However, I remain unclear with the conclusions in this paper.  I refer to your discussion text :

5. "When analyzing the worn surfaces, it was noticed that the Rockit® washers already showed signs of surface distress after a running time of approximately 500 h. This was not observed even with the conventional bearing washers that failed due to fatigue after twice the running time. However, from the results presented here, a higher fatigue life can be expected. Due to the small number of specimens and very long experimental run times without failures, this cannot be quantified."

If I understood the discussion about figure 10 correctly, this figure shows that Rockit washers of the second batch have not failed after 2500 hours.  But also some AISI52100 washers would not fail above 2500 hours. So can you safely conclude from these measurements, that Rockit washers have a significantly better fatigue life than AISI52100 without coatings ?  From my understanding (which is limited !), I would say that figure 10 only shows that the Rockit second batch is 'as good as AISI52100' but I don't know if it indicates that it is a lot better ?

I am no specialist in the fatigue testing, so if this is the case, it should be made more clear in the text, especially for non-specialist readers.

I understand the limitations in testing, 2500 hours of testing is a very long time.  But perhaps this could be made more clear : are the Rockit second batch coatings certainly much better than no coating at all ?

The way the text is currently built up, with long but unclear discussions in chapter 3.2 and an apparently contradiction in the discussion chapter, the whole flow of the article is confusing.

In the conclusions you have added :

"The newly developed hybrid material and processing strategy showed promising results in tribological fatigue life tests on a rolling bearing test bench, exceeding the fatigue life of conventional bearing washers made of monolithic AISI 52 100." 

Are you stating with certainty that the coatings exceed the fatigue life of the conventional bearing washers ?  Is this a solid conclusion from figure 10 ?  If so, then I have no problem with this conclusion, but if it is a hypothesis (as you said 'promising results...') then it should also be made clear.

I hope you can comment on that.

Author Response

Dear Reviewer,

Thank you again for your comments. I have tried to break them down to answer them point by point:

1. It is not quite correct, that Figure 10 “shows that Rockit washers of the second batch have not failed after 2500 hours”. It shows the calculated Weibull failure probability of the failed discs of all other materials over the duration to failure in revolutions. The longest test of monolithic AISI 52100 ran more than 50 million revolutions due to failure of an AISI 52100 disc (black X at about 50% probability of failure). The Rockit disc in the same test (per test 1x Rockit, 3x AISI 52100; see section 2.7) did not fail up to this point. Another set of counter-disks made of AISI 52100 was used to eventually provoke damage to the Rockit disk. This did not occur even after approx. 75 million revolutions ( = 2500 h), which is why the test was canceled. However, this extended test was not included in the evaluation, since only up to the failure of the AISI 52100 disc at approx. 50 million revolutions can it be ensured that the Rockit disc did not suffer secondary damage. This was further specified in the manuscript on page 12.
2. It is correct, that “some AISI 52100 washers would not fail above 2500 hours”. This is taken into account in the analysis using the maximum likelihood method for censored Weibull distributions, see page 13. However, this is due to the general scatter in fatigue testing, why repeated tests and a statistical evaluation have been carried out.
3. If the weakest link theory is followed, the weakest component was the AISI 52100 washer, which always failed before the Rockit washer. Based on this, it is safe to say that “Rockit washers have a […] better fatigue life than AISI 52100 without coatings”. Due to the sample size, statistical significance cannot be derived from this. It is also not possible to quantify the improved fatigue life of the Rockit washers due to the absence of fatigue damage, see page 17.
4. Since in Figure 10 the Rockit samples have already been marked as sample passed and all the above points have been taken up in Section 3.2, we do not know how it could be further specified. If a non-specialist reader is interested in rolling contact fatigue of a custom material for additive manufacturing, we would like to refer to the basic literature [1-3], more specific publications [5-21, 26-28, 30] and our previous work [22-25].
5. We cannot say if the Rockit coatings are "much better" because we cannot quantify life improvement at this time, see answer 3 and sections 3.2 and 6 of our manuscript. However, as also answered above, they are proven to be better than the current standard.
6. If you have specific suggestions on how we can improve section 3.2, we are happy to implement this. We do not notice an "apparent contradiction". Can you please specify this as well?
7. Yes, as explained in all the previous points, we can safely say that the Rockit coatings exceed the fatigue life of conventional bearing washers. This does not follow explicitly from Figure 10 (see answer 1), but from the underlying statistical evaluation of the fatigue life tests explained in section 3.2.
8. "Promising" refers to the potential variety of applications for tribologically stressed machine elements and components that can develop through additive manufacturing with Rockit. We have replaced it with "good performance".

Round 3

Reviewer 4 Report

The author's explanations clarify a lot, but they were not so clear in the original paper.  Suggestion to incorporate the author's replies to revision 2 into the final paper, then I can agree with publication.

Author Response

Dear Reviewer,

We thank you for the positive feedback concerning our response. From our standpoint however, all of your comments from Review #2 are addressed in the text. Unfortunately, we are having trouble revising the text due to the inspecific nature of your initial review #2. To address this, we have summarized the answers to your comments in review #2 in our text point by point:

1. The terminated endurance run is described in lines 400-408.
2. The test specifications and the use of the maximum likelihood method are described in lines 231-234 and 421-426 respectively.
3. This is the main statement of our research, which can be found in lines 495-498, lines 548-550 and lines 563-570.
4. Do you imply that we should add an additional reference to the provided literature? If so, at which point in the text would you consider this as necessary?
5. The better performance of the Rockit washers is addressed in lines 408-409, 432-437, 496-498, 528-531, and especially lines 569-571.
6. We hope that we were able to clear up any contradiction observed in the text through our other responses. If this is not the case would you care to specify what contradiction you still see in the article?
7. Please see answers 1, 3, and 5.
8. This has been renamed in lines 548/549.

We hope this clears up any possible misunderstanding. If not, we urgently ask for specific suggestions for improvement, preferably including a line number. Of course, we would like to prepare our manuscript to its best form for all interested readers. However, it is not possible for us to do so apart from the revisions that have already been incorporated.