Bearing Steel Technologies : 12th Volume, Progress in Bearing Steel Metallurgical Testing and Quality Assurance 9780803176928, 0803176929

Citation preview

Beswick

Bearing Steel Technologies: 12th Volume, Progress in Bearing Steel Metallurgical Testing and Quality Assurance

ASTM INTERNATIONAL Helping our world work better ISBN: 978-0-8031-7692-8 Stock #: STP1623 www.astm.org

STP 1623

Cover photo courtesy SKF Group

ASTM INTERNATIONAL Selected Technical Papers

Bearing Steel Technologies:

12th Volume, Progress in Bearing Steel Metallurgical Testing and Quality Assurance STP 1623 Editor: John Beswick d.

SELECTED TECHNICAL PAPERS STP1623

Editor: John M. Beswick

Bearing Steel Technologies: 12th Volume, Progress in Bearing Steel Metallurgical Testing and Quality Assurance ASTM STOCK #STP1623 DOI: 10.1520/STP1623-EB

ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959 Printed in the U.S.A.

d.

Library of Congress Cataloging-in-Publication Data ISBN: 978-0-8031-7692-8 ISSN: 2160-2050 C 2020 ASTM INTERNATIONAL, West Conshohocken, PA. All rights reserved. This material Copyright V may not be reproduced or copied, in whole or in part, in any printed, mechanical, electronic, film, or other distribution and storage media, without the written consent of the publisher.

Photocopy Rights Authorization to photocopy items for internal, personal, or educational classroom use, or the internal, personal, or educational classroom use of specific clients, is granted by ASTM International provided that the appropriate fee is paid to the Copyright Clearance Center, 222 Rosewood Drive, Danvers, MA 01923, Tel: (978) 646-2600; http://www.copyright.com/ ASTM International is not responsible, as a body, for the statements and opinions expressed in this publication. ASTM International does not endorse any products represented in this publication. Peer Review Policy Each paper published in this volume was evaluated by two peer reviewers and at least one editor. The authors addressed all of the reviewers’ comments to the satisfaction of both the technical editor(s) and the ASTM International Committee on Publications. The quality of the papers in this publication reflects not only the obvious efforts of the authors and the technical editor(s), but also the work of the peer reviewers. In keeping with long-standing publication practices, ASTM International maintains the anonymity of the peer reviewers. The ASTM International Committee on Publications acknowledges with appreciation their dedication and contribution of time and effort on behalf of ASTM International. Citation of Papers When citing papers from this publication, the appropriate citation includes the paper authors, “paper title,” in STP title, book editor(s) (West Conshohocken, PA: ASTM International, year), page range, paper doi, listed in the footnote of the paper. A citation is provided on page one of each paper. Printed in Hanover, PA August, 2020

d.

Foreword THIS COMPILATION OF Selected Technical Papers, STP1623, Bearing Steel Technologies: 12th Volume, Progress in Bearing Steel Metallurgical Testing and Quality Assurance, contains peer-reviewed papers that were presented at a symposium held May 15–17, 2019, in Denver, CO, USA. The symposium was jointly sponsored by ASTM International Committee A01 on Steel, Stainless Steel and Related Alloys, and Subcommittee A01.28 on Bearing and Power Transmission Steels. Symposium Chair and STP Editor: John M. Beswick Montfoort, The Netherlands

iii d.

d.

Contents

Overview

ix Rolling Bearing Failure Modes and Advanced Analysis

Review on Crack Initiation and Premature Failures in Bearing Applications Reinder H. Vegter and Kenred Stadler Influence of Steel Cleanliness on RCF and WEC Formation Toni Blass, Xiaohong Xu, Kirsten Wunder, Werner Trojahn, Ke Geng, and Feng Li Rolling Contact Fatigue Transformations in Aero Steels: The Effect of Temperature on Microstructural Decay Xingzhong Liang, Finn Sykes, and Pedro E. J. Rivera-Dı´az-del-Castillo

1

26

50

Microcleanliness Relationships and Testing of Air-Melt Bearing Steels Quantitative Ultrasonic Characterization of Subsurface Inclusions in Tapered Roller Bearings Showmic Islam, Satyajeet P. Deshpande, Luz D. Sotelo, Musa Norouzian, Michael T. Lumpkin, Liesl F. Ammerlaan, Allen J. Fuller, and Joseph A. Turner

66

Developments in RCF Testing of Bearing Steels Design and Validation of a Modular Rolling Contact Fatigue/Rolling-Sliding Contact Fatigue Testing Machine Nicholas Novack, Robert L. Cryderman, and Trace A. Rimroth Investigation of Fatigue Behavior around Nonmetallic Inclusion Using a Newly Developed Rolling Contact Fatigue Test Method Takeshi Fujimatsu

82

103

Rolling Bearing Surface Damage and Effect on RCF Life Lubricant-Induced White Etching Cracks: Mechanism and Effects of Surface Finishing Mohanchand Paladugu

v d.

131

Surface Damage in Rolling Bearings and the Impact on Rolling Contact Fatigue Markus Dinkel, Xiaohui Zheng, Michael Warmuth, and Martin Correns

147

Effect of MnS on the Micropitting Behavior of Through Hardened Bearing Steel during Rolling Contact Fatigue Vikram Bedekar, Carl Hager, and R. Scott Hyde

169

Influence of Material, Heat Treatment, and Microstructure in Resisting White Etching Crack Damage Mohanchand Paladugu and R. Scott Hyde

182

Dimensional Stability Influence of Heat Treatment Conditions on the Dimensional Stability of SAE 52100 Christine Sidoroff-Coicaud, Christophe Le Bourlot, Carole Dessolin, Michel Perez, Victor Lejay, Pierre-Emmanuel Dubois, and Pierre Dierickx

202

Complexity of Dimensional Stability of Case-Hardened Bearing Components Pei Yan and Mohamed Y. Sherif

246

Developments in Bearing Component Manufacturing Selective Laser Melting (SLM) of M50NiL—Enabling Increased Degrees of Freedom in New Design Concepts Patrick Mirring, Andreas Rottmann, and Carsten Merklein Manufacturing of Large-Diameter Rolling Element Bearings by Steel-Steel Multimaterial Systems Timm Coors, Maximilian Mildebrath, Florian Pape, Thomas Hassel, ¨rgen Maier, and Gerhard Poll Hans Ju Ultrasonic Evaluation of Tailored Forming Components Florian Pape, Timm Coors, Tim Matthias, Bernd-Arno Behrens, and Gerhard Poll

261

277

300

Developments in Air-Melt Clean Bearing Steel Steelmaking Improved Processing Techniques for Inclusion-Free Steel for Bearing and Mechanical Component Applications Christopher DellaCorte Performance and Reliability of Powder Metallurgy Steels for Aerospace Bearings Jacques Bellus, Christine Sidoroff-Coicaud, Viktor Sehlstedt, Atman Benbahmed, Johanna Andre´, and Olivier Blanchin The Use of SEM-EDS and PDA-OES Techniques to Help the Development of the Production of Bearing Steel Audrey Col, Andrea Spadaccini, Daniel Acevedo, and Christophe Stocky Advances in Billet Cast Carbon Steel Quality for High-Performance Rolling Bearings Eduardo Scheid and Denise Correa de Oliveira

vi d.

313

332

352

366

Improvements in GCr15 (52100) High Carbon Bearing Steel Steelmaking and Their Effect on Inclusions, Segregation, and Fatigue Properties Xiaohong Xu, Jigang Liu, Guoqing Xu, Qing Yin, Xudong Zhang, and Hans-Åke Munther

381

New and Novel Steel Compositions for Advanced Rolling Bearing Usage Temperature-Resistant, Corrosion-Tolerant Carburizing Bearing Steel for Aero-Engine Applications Aidan Kerrigan, Alexandre Mondelin, Jean-Baptiste Coudert, ´o Mohamed Y. Sherif, and Yves Mahe

403

High Performance Ferrium Steels for Aerospace Gearing and Bearing Applications Kerem Taskin

421

Hybrid Steel and Its Potential for Bearing Applications Jan-Erik Andersson, Fredrik Lindberg, and Steve Ooi

436

Application of Fracture Mechanics to Bearing Steel Property Characterization Effect of Carbide Segregation on Mode I Fatigue Resistance Properties of the Bearing Tool Steel Roll Blade Die Aleksej Molokanov, Martin Rawson, Tim Moreton, and Geoff West

455

Relevance of Fracture Mechanics in Rolling Bearing: Functional Property Determination and Steel Quality Assurance ´o Jean-Baptiste Coudert, Aidan Kerrigan, Alexandre Mondelin, and Yves Mahe

474

VIM-VAR Steel Know-How—Aero Steels Metallurgy and Functional Properties VIM-VAR Steelmaking for Bearing Steel Grades Stephen Carey Melt Methods and Their Effects on Cleanliness for Bearing Performance Colleen Tomasello and George Shannon Spall Propagation Characteristics of Life-Tested VIM-VAR M50 and Pyrowear 675 Bearing Steels Hitesh K. Trivedi, DaMari A. Haywood, Lewis Rosado, and Mathew S. Kirsch

499

515

528

Spall Propagation Characteristics of As-Manufactured Aerospace Bearing Steels Hitesh K. Trivedi, DaMari A. Haywood, Mathew S. Kirsch, and Lewis Rosado

551

Testing to Reveal Tribology Mechanisms for Advancing Bearing Steels Lavern D. Wedeven, William F. Black, Graham G. Wedeven, Robert J. Homan, Nelson H. Forster, Herbert A. Chin, and Rainer Fluch

574

Adhesive Wear Performance of Pyrowear 675 in All-Metal and Hybrid Configuration Daulton D. Isaac, Mathew S. Kirsch, Patrick T. Hellman, Andrew S. Foye, and Hitesh K. Trivedi

605

Accelerated Life Testing of Pyrowear 675 Material at 218°C (425°F) Using a Ball-on-Rod Rolling Contact Fatigue Tester Hitesh K. Trivedi and Lewis Rosado

vii d.

628

Author Index

649

Subject Index

651

d.

Overview There are occasions when jewels evolve from quite mundane beginnings, and the ASTM International bearing steel symposium series, starting in 1946, is an example of something special. The symposium “Rolling Bearing Steel: Progress in Bearing Steel Metallurgical Testing and Quality Assurance” was held May 15–17, 2019, in Denver, CO, USA. Symposia dedicated to bearing steel technologies follow an established tradition of ASTM support for the topic. Information on the past ASTM bearing steel symposia and related STP publications are given as follows:

Topic

Year

Location

Chairman

STP

70

Symposium on Testing of Bearings

1946

Buffalo

—

Rating of Non-Metallic Inclusions

1974

Boston

Joe Hoo

575

Roller Contact Fatigue Testing

1981

Phoenix

Joe Hoo

771

Effect of Steel Manufacturing Pro-

1986

Phoenix

Joe Hoo

987

San Diego

Joe Hoo

1195

New Orleans

Joe Hoo and

1327

cesses on the Quality of Bearing Steels Creative Use of Bearing Steels

1991

Bearing Steels: Into the 21st Century 1996

Bill Green Sixth International Symposium on

2001

Phoenix

John Beswick

1419

2005

Reno

John Beswick

1465

2009

Vancouver

John Beswick

1524

2011

Tampa

John Beswick

1548

2014

Toronto

John Beswick

1580

2016

Orlando

John Beswick

1600

2019

Denver

John Beswick

1623

Bearing Steels Advances and State of the Art in Bearing Steel Quality Assurance Developments on Rolling Bearing Steels and Testing Advances in Rolling Contact Fatigue Strength Testing and Related Substitute Technologies Advances in Steel Technologies for Rolling Bearings Progress in Steel Technologies and Bearing Steel Quality Assurance Progress in Bearing Steel Metallurgical Testing and Quality Assurance

ix d.

The aim of the ASTM bearing steels symposia has always been to facilitate an exchange of relevant technical information on rolling bearing steel technologies. Global participation of experts has always been a key feature, and without presenters and participants from outside North America, the symposia would not have been a success over the years. Bearing steel technologies plan and look forward to the ASTM events, and the attendance numbers are relatively stable but the global coverage continues to expand. The majority of the presentations from the symposium have been compiled as peer-reviewed papers for publication as ASTM selected technical papers (STP1623). A rigorous peer-review process has been applied as befits a reputable technical publication. The STP editor is beholden to the peer reviews for finding time and motivation to perform this critical task. Experience has shown that the STPs are an excellent bearing steel technology reference, and the authors, their respective companies, and the anonymous peer reviewers are congratulated for their commitment to publication. The symposium program contained 37 presentations with 110 persons registering for the 10 sessions during a two-and-half-day event. The symposium comprised the following sections: Rolling Bearing Failure Modes and Advanced Analysis Microcleanliness Relationships and Testing of Air-melt Bearing Steels Developments in RCF Testing of Bearing Steels Rolling Bearing Surface Damage and Effect on RCF Life Dimensional Stability Developments in Bearing Component Manufacturing Developments in Air-melt Clean Bearing Steel Steelmaking New and Novel Steel Compositions for Advanced Rolling Bearing Usage Application of Fracture Mechanics to Bearing Steel Property Characterization VIM-VAR Steel Know-How—Aero Steels Metallurgy and Functional Properties Bearing metallurgists will recognize some familiar trends and some new directions in the STP1623 papers. The world is changing and the alloy steel technology trends have synergies, and the remit of the ASTM Subcommittee A01.28 has been revised since the 12th Bearing Steel Symposium. The subcommittee now covers bearing and transmission steels and as such is responsible for bearing and transmission steel specifications. The next ASTM bearing steel symposium event is destined to be the 1st ASTM Bearing and Transmission Steels Symposium, with new opportunities to apply alloy steel know-how to support the manufacture of better and added value transmission steel products. The opportunities arising from collaborations within the bearing and transmission steel technologies cannot be overestimated, and the merging of the relevant steel purchasing specifications can result in

x d.

improvements from both the technical and commercial perspectives. Merging of the specifications could improve the commercial margins to support purposeful R&D, related publications and the future ASTM symposia. The STP editor is indebted to Jeff Fuller, the ASTM A01.28 Subcommittee Chairman, for supporting the symposium arrangements, and to the ASTM International organization for fostering the continuation of the symposia series. The following organizations provided financial support for the symposium; very many thanks for your sponsorship:

Jiangyin Xingchen Special Steel Amsted Rail Co., Inc

China USA

Ascometal

France

Bo ¨hler Edelstahl GmbH Carpenter Technology Corporation

Austria USA

Charter Steel

USA

FNsteel B.V. Georgsmarienhu ¨tte (GMH) GmbH

The Netherlands Germany

Gerdau Special Steel North America

USA

Ovako AB Saarstahl AG

Sweden Germany

Sanyo Special Steel

Japan

SKF B.V. Timken

The Netherlands USA

TimkenSteel

USA

It has been the symposium chairman’s privilege to have, on behalf of ASTM International, prepared the program, chaired the symposium, and edited the STP1623. John M. Beswick Montfoort, The Netherlands

xi d.

d.

BEARING STEEL TECHNOLOGIES: 12TH VOLUME

STP 1623, 2020 / available online at www.astm.org / doi: 10.1520/STP162320190054

Reinder H. Vegter1 and Kenred Stadler2

Review on Crack Initiation and Premature Failures in Bearing Applications Citation R. H. Vegter and K. Stadler, “Review on Crack Initiation and Premature Failures in Bearing Applications,” in Bearing Steel Technologies: 12th Volume, Progress in Bearing Steel Metallurgical Testing and Quality Assurance, ed. J. M. Beswick (West Conshohocken, PA: ASTM International, 2020), 1–25. http://doi.org/10.1520/STP1623201900543

ABSTRACT

In this paper, a review of various root causes for premature bearing macrospalling failure will be presented. Previously, premature failures were often associated with the occurrence of white etching areas, which were seen as the cause of these failures. It has been determined, however, that white etching areas are not limited only to premature failures; they occur in all types of failures during bearing life. A wide range of bearing operation conditions that can lead to rolling contact fatigue, including cracking with the development of white etching areas, has been shown. This is not limited to one specific type of failure or operating condition such as the presence of hydrogen, electrical current passage, overload, or others. The current investigations indicate that premature bearing failure mainly can be linked to two operation conditions, namely stresses higher than anticipated or material strength lower than expected (or a combination of both). A scheme to unravel the complex interactions of bearing steel properties and operating conditions that lead to premature failure will be presented. This scheme is very useful for application analysis and for finding solutions for the cases where premature failures occur. Based on the observations made on premature failures and white etching cracks, a proposal for a crack initiation mechanism at non-metallic inclusions is presented.

Manuscript received April 29, 2019; accepted for publication September 18, 2019. 1 SKF RTD, Meidoornkade 14, 3993 AE Houten, The Netherlands https://orcid.org/0000-0003-3523-9516 2 SKF RTD, Gunnar-Wester-Straße 12, 97421 Schweinfurt Germany 3 ASTM 12th International Symposium on Rolling Bearing Steel: Progress in Bearing Steel Metallurgical Testing and Quality Assurance on May 15–17, 2019 in Denver, CO, USA. C 2020 by ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959. Copyright V

d.

1

2

STP 1623 On Bearing Steel Technologies: 12th Volume

Keywords bearing steel, white etching cracks, inclusions, crack initiation

Introduction Bearings are critical machine components for carrying loads and transmitting motions. The challenge of ever-increasing power density from modern equipment manufacturing imposes higher demands for the load-carrying capacity on one hand and the reliability of bearings on the other hand. For this reason, bearings are made of highly engineered steel with excellent properties.1 Despite this, bearing steel occasionally can fail in an unexpectedly short life even if manufactured and used in accordance with the bearing industry recommendations. Under normal rolling contact fatigue (RCF) conditions, life is calculated using standardized methods; but in certain bearing applications and under specific conditions, bearing steel can fail prematurely. These failures are often caused by cracks that are decorated with so-called white etching areas (WEAs). Therefore, they are often called irregular white etching crack (WEC) networks and sometimes are also called irregular, or white structured flaking (WSF).2–6 The root cause for this failure mode has been elusive and a topic of considerable investigations within the bearing industry as well as the academic environment. Various hypotheses on the root cause of such failures have been put forward, such as the following: • Hydrogen embrittlement7 due to hydrogen uptake resulting from decomposition of lubricant,3 mixed friction and slip,8 stray current,9 or corrosion10 • Tribological reactions leading to an electromagnetic initiation of WEC by electrothermal carbide transformation11 • Adiabatic shear banding resulting from impact of rollers onto the ring that leads to WEC networks12 • A process of surface cleavage cracking caused by frictional traction, followed by corrosion fatigue crack growth13 • Initiation of butterflies due to high loads and followed by propagation of butterflies into large WEC networks14–17 In general, these phenomena can be classified under two headings: (a) the situations where stresses are higher than expected in the application and (b) the situations where, due to several possible factors, the material is weaker than expected and therefore cannot cope with the requirement anymore. Schematically, this is shown in figure 1. Many hypotheses for the origin of WEC failure neglect the fact that an initial crack is needed to form a system of white etching cracks. However, the definition of a crack onset can differ from hypothesis to hypothesis. In this work, new observations of the WEC system, including hardness measurements, will be presented. d.

VEGTER AND STADLER, DOI: 10.1520/STP162320190054

FIG. 1 Simplified summary of several root cause hypotheses for WEC generation.

Following from that, an inclusion crack initiation model is proposed and a new classification scheme based on figure 1 will be described. In addition, some examples of different bearing investigations from test and field cases will be presented to help in clarifying the role of WEC in rolling contact fatigue as well as the sequence of events until failure. From this work, it will be reasoned that WEC in failed bearing steels is a symptom rather than a root cause for failure.

Occurrence of White Etching Areas in Bearing Steel White etching refers to the white appearance of the altered microstructure of a polished and etched steel sample. The affected areas, consisting of ultrafine, nanorecrystallized, carbide-free ferrite, appear white under an optical microscope because of their low etching response to the etchant. The opposite of white etching is the socalled dark etching. As the name implies, a dark etching area appears darker than the bulk material after etching. d.

3

4

STP 1623 On Bearing Steel Technologies: 12th Volume

White etching phase transformation is known to occur in bearing steels for different reasons. These include: • Steel reaustenization (e.g., caused by grinding burns, scuffing/smearing conditions, and electric erosion remelting) • Martensite decay in bearing steels leading to the formation of white etching ferritic bands at low and high angles to the rolling surface, which is associated with a large volume of predominantly cyclic plastic deformation • Preexisting material defects in hardened steels under rolling contact—the phenomena known as butterflies • Rubbing between two surfaces leading to tribological transformed structures • Adiabatic shear bands formed in ballistic impact tests More details on the differences in the formation of the aforementioned white etching phase transformations can be found in several published studies.1,6,14,17–22 In this paper, some examples of white etching crack networks will be given from field cases (fig. 2) and from internal early failure (or WEC) testing, as well as from standard fully lubricated endurance testing (rolling contact fatigue). Currently, it is postulated, on the one side, that the occurrence of irregular WECs is generated by special situations (such as “bad” lubricants, vibrations, stray currents, water contamination, additional loads, and so on). Several works support this argument.7–13,16,23

FIG. 2 Typical irregular WEC found in a premature failed bearing after sectioning and Nital etching.

d.

VEGTER AND STADLER, DOI: 10.1520/STP162320190054

FIG. 3 Dark etching region development in the inner ring of a deep groove ball bearing (Type 6309). Contact pressure is 3.7 GPa; the outer ring temperature of the tested bearing is 53 C: (A) new bearing, (B) 10
106 revolutions, (C) 63
106 revolutions, and (D) 660
106 revolutions. All micrographs are taken in a plane parallel to rolling direction.30

However, on the other side, irregular WEC and WEA generation is also found to be a part of bearing rolling contact fatigue.17,19,24–29 This important observation makes it necessary to take a closer look into the overall role of WEC as a common symptom of fatigue. Under normal laboratory conditions, WEAs may not be observed. Therefore, many research activities have focused on the occurrence and development of the so-called dark etching region (DER). This is the region where global material degradation occurs due to the mechanical and thermal loading in rolling contact fatigue. One of the most common illustrations of this development is shown in figure 3. Here, the growth of the DER is shown as a function of the number of revolutions in tested deep groove ball bearings. The location of the DER, away from the surface, is explained by the profile of the shear stresses coming from the Hertzian stress calculation: The absence of shear stresses at the surface explains the unaltered material very close to the surface, whereas the highest level of material degradation is shown at the position where the maximum shear stress is located. Figure 3 shows that the microstructure alterations are homogeneously distributed, which means that no concentration of damage is taking place.30 This can be attributed to the fact that the material is very clean and no non-metallic inclusions play a major role in local initiation of fatigue damage and cracks. On top of this, the global development of DERs is also accompanied by the development of compressive residual stresses, which in turn also suppress the formation of eventual damage or cracks. The nature of DER formation is rather complex in terms of microstructure changes. It is a transformation induced by the shear stresses that influence the carbide distribution in the steel. It is assumed that small e-carbides and g-carbides go into solution and let larger carbides grow. As a side effect, the chemical reactivity of the material changes, which gives the possibility of visualizing the location of the microstructure changes by etching, as visible in figure 3. In the very advanced stages d.

5

6

STP 1623 On Bearing Steel Technologies: 12th Volume

of microstructure degradation (fig. 3D), white bands occur in the microstructure. These white bands are a consequence of recrystallization. The recrystallization process is not visible in the transition from figure 3C to D, but it is clear from earlier work that the material in figure 3 shows texture and grain refinement.31 Until the stage in which recrystallization occurs, the microstructure changes are mainly dominated by the change in carbide morphology and the dislocation structure. Both phenomena develop simultaneously, where the presence of an increasing dislocation density and networking of dislocations supports the dissolution of carbides and carbon diffusion. The mechanism of recrystallization is not fully understood in this case. Analogous to the occurrence of white etching areas, the continuously increasing dislocation density could lead to very small grain size, which is then followed by grain growth in the preferred orientation. However, in addition to the pictures shown in figure 3, limited experimental evidence exists for this process at the end of the life of a bearing with only DER development. Very often, cracks occur before the stage shown in figure 3D. These cracks are then the dominating failure cause and prevent the recrystallization from becoming visible. This is also the case for the occurrence of white etching cracks; in many bearing applications, these cracks do appear earlier than the advanced development of circumferential macro DER bands, meaning that, in bearing applications, failure due to macro DER is relatively seldom.

Metallurgical Characterization The metallurgical nature of white etching areas has been described by various authors. In this section, a few examples of typical microstructures showing white etching cracks will be presented. GENERAL PICTURE OF WEA MICROSTRUCTURE

A common picture of a white etching area is that it is a part of the microstructure that strongly deviates from the original bearing steel microstructure. A typical example is given in figure 4. The WEA is a changed microstructure adjacent to cracks in the material. The WEA is surrounded by the original microstructure, and the transition from the altered microstructure to the normal bearing steel microstructure is very sharp. In some cases, this transition is made by a crack; but in many cases, the transition from WEA to the steel microstructure shows bonding at the interface. The pictures in figure 5 show the microstructure of the white etching area in a crack in the same bearing (as shown in fig. 4) with increasing magnification using scanning electron microscopy. In figure 5B, a banded structure can be seen. Small grains are present inside and outside this band (fig. 5C). A more detailed picture of the grain structure is shown in figure 5D. Here, it can be seen that the grain size is smaller than 50 nm. Because this picture reached the end of the resolution of the scanning electron microscope used, further work was performed using transmission d.

VEGTER AND STADLER, DOI: 10.1520/STP162320190054

FIG. 4 Crack in a bearing component showing white etching area development.

FIG. 5 White etching area structure in a used bearing. The pictures show increasing magnification of the same area: (A) Magnification 2609x, (B) magnification 18,051x, (C) magnification 72,204x and (D) 288,815x.

d.

7

8

STP 1623 On Bearing Steel Technologies: 12th Volume

FIG. 6 TEM photo of WEA.

electron microscopy (TEM). The microstructure as seen by TEM is given in figure 6, and the accompanying electron diffraction pattern is shown in figure 7. The circles in this picture indicate that the microstructure of figure 6 has a random orientation, which means that the grains have been growing without a texture in this structure. A more detailed photo is given in figure 8, which shows that the typical grain size in the WEA is 30 to 50 nm. To investigate the transition layer between the WEA and the original steel microstructure, a picture was taken from the interface (fig. 9). It confirms that the interface is uninterrupted, which means that, at this location, there is no crack or other phase between the WEA and the steel matrix. HARDNESS OF WEA

The material properties of a white etching area are not easy to characterize. One of the few methods possible is the use of nanoindentation. A measurement on a bearing taken from a field application that shows WEA in the subsurface regions is presented here. To have an accurate hardness measurement, the indents were made on an aspolished sample with the cracks as marks for orientation (fig. 10A). After the indents were made, the sample was Nital-etched, and an analysis was made as to whether the indents fell in the white etching area or in the steel matrix (fig. 10B). In some cases, it could not be determined exactly whether the indent was in the WEA or in the steel matrix; in that case, the indent was marked half. d.

VEGTER AND STADLER, DOI: 10.1520/STP162320190054

FIG. 7 Electron diffraction pattern of the area shown in figure 6.

The results of the measurements are shown in figure 11. A scatter is observed, but the hardness values of the WEA are up to 50% higher than the steel matrix values. CRACKS ORIGINATING AT INCLUSIONS (BUTTERFLIES)

A systematic study on the role of inclusions has been performed by Grabulov, Petrov, and Zandbergen28 and by Grabulov.29 It was observed that the onset of a white etching area can be caused by a debonding of the inclusion. The debonding acts as the initiation of a crack, which can be seen in figure 12, from Grabulov.29 The pictures were taken from a tested roller with artificially included Al2O3 inclusions, which was tested for 1.3
107 load cycles under a contact pressure of 2.6 GPa at 100 C. Figure 12A shows the inclusion and the surrounding matrix. This inclusion was located in a zone where the RCF loading was relatively mild (i.e., no high shear stress was applied on the interface between the inclusion and the steel matrix). It can be seen that the transition zone is a sharply defined crack, where the microstructure of both the inclusion (Al2O3) and the steel are unaffected. Figure 12B shows that the interface looks different when the inclusion is located in the highest loaded zone around the z0 depth, where the shear forces are maximal due to the Hertzian pressure. The interface is less sharp, and the steel matrix has been damaged due to the crack face rubbing. d.

9

10

STP 1623 On Bearing Steel Technologies: 12th Volume

FIG. 8 WEA microstructure showing grains smaller than 50 nm.

Based on this “rubbing assumption,” the work describes how this damaged zone adheres to the inclusion and how a crack can redirect itself from the interface into the steel matrix. The starting point in this reasoning is the fact that non-metallic inclusions such as Al2O3 are not fully bonded to the surrounding matrix. In many cases, there is an area of debonding, which can be treated as a small crack. The first question that arises is why this crack does not propagate along the inclusion/steel interface but usually starts entering the steel matrix. During this propagation, it is then usually accompanied by a WEA, making it a butterfly. The microstructure of the very thin layer consists of a WEA. This is formed because the stresses concentrate at the interface and could cause microplastic deformation32 under the influence of alternating stresses. This very small zone consists of a material that contains an increased amount of dislocations and other crystallographic defects, such as vacancies and vacancy-carbon complexes. This initial situation is sketched in figure 13; a zone with highly damaged material is adjacent to the crack. However, this zone is not only put under high shear stress during the load cycle but, due to the hydrodynamic stresses, is also pressed onto the inclusion during each load d.

VEGTER AND STADLER, DOI: 10.1520/STP162320190054

FIG. 9 TEM photo showing a sharp boundary between the WEA and the matrix.

FIG. 10 Nanoindentation in WEA: (A) polished surface and (B) detail of etched surface showing locations of some indents in the WEA.

cycle. During this pressing, a chemical or physical bonding can occur. This type of bonding is likely because the inclusion/steel interface is under vacuum conditions. Next, a crucial observation is made in figure 14, which shows that the inclusion surface, after the adjacent steel matrix has been damaged, does adhere to some of the highly damaged steel. This can be seen because the Al2O3 is debonded at the line indicated by Position 1 but is bonded at the line in Position 2. The Al2O3-steel d.

11

12

STP 1623 On Bearing Steel Technologies: 12th Volume

FIG. 11 Nanoindentation results in WEA, at the edge and in the matrix. Indentations are shown in figure 10.

FIG. 12 SEM image of alumina-steel matrix interface of (A) an inclusion far away from the loaded zone at a depth of 2100 μm below the raceway and (B) an inclusion in a higher loaded zone at 650 μm below the raceway. The shear load is higher in (B) but still very moderate with a value of approximately 50 MPa when the contact pressure in the used geometry is 2.6 GPa.29

d.

VEGTER AND STADLER, DOI: 10.1520/STP162320190054

FIG. 13 Situation at the debonded inclusion/steel interface in the early stage of fatigue.

FIG. 14 TEM picture of Al2O3-steel matrix interface with two EDS element area maps. (A) shows the element distribution in region A and (B) shows the element distribution in region B.29

matrix interface is not very distinct between the positions of both lines but shows an area between the steel and the inclusion that is a mixture of material and defects (pores). This is confirmed by the energy dispersive spectroscopy (EDS) mapping that is also visible in the figure (made in Field A and Field B). The iron (Fe) area d.

13

14

STP 1623 On Bearing Steel Technologies: 12th Volume

FIG. 15 EDS line spectrum of Position 1 in figure 14.29

does not stop abruptly, but a rather mixed zone is present before the aluminum (Al) signal on the left side increases sharply. This is also illustrated in figure 15 and figure 16, which show an EDS line spectrum for Positions 1 and 2, respectively. In figure 15, the presence of a real gap between the inclusion and steel can be seen in the micrograph, but this is also confirmed by the line spectrum, which shows the absence of elements between 0.2 and 0.4 lm. Figure 16 shows a different picture.

FIG. 16 EDS line spectrum of Position 2 in figure 14. The arrow corresponds to the arrow in the figure.29

d.

VEGTER AND STADLER, DOI: 10.1520/STP162320190054

Here, the gap is showing Fe on both sides. This leads to the conclusion that the gap has been moving from the interface into the steel matrix. In the aforementioned example, the crack movement into the steel matrix is very small. It can be imagined that when this process continues, the crack gets further away from the inclusion, leaving a highly damaged area of material behind that is attached to the inclusion. In this material, the level of defects is high; therefore, some of this can be seen as being in a semi-amorphous state. The driving force for recrystallization is high and a fine crystalline structure will be formed, which is the known white etching area structure. The hypothesis here is that the strength of the WEA exceeds the strength of the normal steel matrix. An important argument for this is the fine grain size, accompanied by the high hardness. It is well-known in materials science that the strength of a material increases with decreasing grain size (the Hall-Petch relation). A consequence of this is that the WEA is less prone to cracking than the steel/WEA interface or the steel itself. If follows from this that the crack, once away from the inclusion and separated by an area of WEA from the inclusion, will propagate away from the inclusion/steel interface. The proposed mechanism is shown in figure 17. The first step (Situation A) is a partially debonded non-metallic inclusion. This inclusion and the surrounding

FIG. 17 Proposed mechanism for crack initiation at the inclusion/steel interface. (A) A debonding is present, (B) crack face rubbing causes first WEA, (C) WEA attaches to the inclusion and a crack occurs between matrix and WEA, and (D) this process continues, leading to a growing butterfly wing.

d.

15

16

STP 1623 On Bearing Steel Technologies: 12th Volume

matrix are subjected to the Hertzian stresses during every load cycle. It is important to mention here that the Hertzian shear stresses reach their maximum when the normal stresses are relatively low, and the normal stresses are high when the shear stress is zero. Because the shear stresses are causing microplastic deformation, their detrimental effect happens when the normal stresses are not suppressing this. In the stage when the shear stress is zero, the normal hydrostatic stresses press the damaged microstructure onto the inclusion. As shown in figure 16, this can lead to a transfer of material to the other side of the crack (Situation B in fig. 17). This process may continue and repeat itself as more and more microplastically deformed material adheres to the WEA and recrystallizes, forming a very strong phase in the steel. In this way, the crack is initially moving sideways; but later in the process, it may also get longer by shearing in the classical way (Situation C and Situation D). From that moment on, the WEA may be formed on both sides due to crack-face rubbing, but eventually the transfer of the material mechanism may still occur, leading to asymmetric WEAs along cracks. This mechanism initially leads to butterflies, which have a typical shape as shown in figure 18. Later, these may develop in larger crack systems where WEAs occur along the crack faces during RCF.

FIG. 18 Typical butterfly type of WEA originating at a non-metallic inclusion. It was observed in the inner ring of a tapered roller bearing 32064, which was tested at 1.8 GPa for 137
106 revolutions (test temperature, 50 C).

d.

VEGTER AND STADLER, DOI: 10.1520/STP162320190054

CRACKS OBSERVED IN CLASSICAL ROLLING CONTACT FATIGUE LIFE

In the previous section, the origin of cracks and the formation of normal butterflies were discussed. The butterflies are usually relatively small features in comparison with larger white etching cracks that are observed. The step between the butterfly types of cracks and large systems of white etching cracks has not been properly described so far. In this section, the similarities between WEAs at butterflies and large-scale WECs are discussed in more detail on the basis of a number of examples. It was observed that in some cases, larger crack systems can originate by the occasionally occurring situation where there are multiple inclusions forming a butterfly simultaneously. One example is given in figure 19. Visible in this figure are three inclusions that have initiated a butterfly. Because the butterflies are located close to each other, the WEA of each butterfly has grown so far that the ends are (almost) in contact with each other. It can be expected that the moment these areas are in full contact, the system would be judged to be a white etching crack with a length of 250 lm instead of three individual butterflies. In this way, it can be proven that white etching cracks can be caused by nonmetallic inclusions in the bearing steel. The shown microstructure comes from a standard tapered roller bearing (32064) that was endurance tested under a normal contact pressure of 1.8 GPa. No special conditions were applied during the bearing test; the bearing was operating under fully flooded lubrication conditions with a lubrication parameter of j greater than four, which indicates that there is no metal-to-metal

FIG. 19 Three butterflies originating at three inclusions in a tested tapered roller bearing. Picture was taken 0.56 mm below the raceway surface. Bearing was tested for 137
106 revolutions with 1.8 GPa contact pressure.

d.

17

18

STP 1623 On Bearing Steel Technologies: 12th Volume

FIG. 20 Butterfly-to-WEC transition in the case of a single inclusion.

contact. Observations of the surface confirm the lack of surface fatigue in this bearing. The bearing was stopped due to some spalling damage, which was far away from the shown WEA formation. Another example of the transition of butterflies to WECs is shown in figure 20. This picture shows a WEC in the same bearing studied in figure 5. The origin of the crack is a non-metallic inclusion that developed a butterfly with two wings. One wing was “arrested” and did not grow much, but the crack and the adjacent white etching area on the left side did continue their growth and would be recognized as a WEC if the inclusion were not visible. Systematic studies performed by Evans et al.14 have confirmed the hypothesis that an inclusion can be found in every WEA. In the studies by Evans et al., some tests were performed with hydrogen-charged steel. In the current work, the initial shape of a WEC as it is formed out of three crack-initiating inclusions is studied. This work shows that under normal RCF conditions, as applied in a SKF large size bearing test, WEAs can be formed and that white etching cracks are not uniquely related to premature failure. CRACK INITIATION PRECEDING WEA

Another method of accelerating the occurrence of a WEA in bearing steel is to initiate cracks artificially. One way of doing this is to start with already precracked microstructures; in this case, it is very easy to generate WEAs, as shown by Solano-Alvarez and Bhadeshia.33 Another method is to charge the bearing steel with hydrogen and d.

VEGTER AND STADLER, DOI: 10.1520/STP162320190054

FIG. 21 Crack in hydrogen-charged inner ring after 260,000 revolutions.

then test the material. Vegter and Slycke7 showed that a WEA occurs after a relatively short testing time. The question is whether this WEA occurs together with the cracks or after the cracks have been formed. To study this in more detail, a bearing test was performed with a deep groove ball bearing, where the inner ring was charged with hydrogen. The bearing was tested with a contact pressure of 3.2 GPa and had been running with a hydrogen content of 4 to 5 ppm wt. The test was stopped after 260,000 revolutions. Several cracks were found in the subsurface of the inner ring; one example is shown in figure 21. Looking at this figure, it is not immediately apparent whether there are cracks with a white etching area around them. However, a more detailed study shows that this is not the case (fig. 22). These observations show that the assumption that cracks are preceding WEAs is strongly supported. A WEA is a consequence of the presence of cracks in the microstructure and can occur if cracks are subjected to RCF loading for a certain amount of load cycles. In the example given, the amount of load cycles that the crack has experienced is too limited (or the load is too mild) to generate WEAs along the crack faces. DAMAGE DEVELOPMENT DURING ACCELERATED FATIGUE VERSUS RCF

The conditions under which cracks occur can be various. In this work, two groups of situations are distinguished, namely the conditions under which the load in the bearing is higher than anticipated and the condition where the material is weakened, which means that it cannot cope with the load as it would do so under normal conditions. d.

19

20

STP 1623 On Bearing Steel Technologies: 12th Volume

FIG. 22 Detail of crack as visible in figure 11. Adjacent to the crack, no WEA is visible.

In addition to these situations, in the normal state of a bearing, RCF occurs after a very long bearing life. This manifests itself through dark etching regions, white etching bands, and also eventually through cracks. As shown in the previous section, in normal situations, these cracks may develop into butterflies and white etching cracks. To clarify the relationships among the various possible situations, a scheme was developed and is shown in figure 23).34 The examples given in the prior sections show that WEAs can occur under a variety of conditions. In various situations, the root causes of the cracks preceding the white etching areas are different; however, all lead to the situation where WEAs can occur due to accumulation of microstructural damage along the crack faces.35

Discussion The aforementioned observations about WECs show that in the absence of elevated hydrogen, it is very likely in many cases that cracks start at inclusions. More specifically, the inclusion/steel matrix interface can act as a crack that exists prior to RCF loading. The cracks presented in figure 11 and figure 12 do not have white etching matter around them. This supports the assumption that cracks start as normal cracks and white etching matter forms in a later stage, which agrees with the findings of Solano-Alvarez and Bhadeshia.33 From this observation, a generic diagram listing the failure routes has been developed (fig. 24). The core of the diagram is the d.

VEGTER AND STADLER, DOI: 10.1520/STP162320190054

FIG. 23 Scheme to relate accelerated fatigue to normal RCF.34

FIG. 24 White etching crack occurrence (simplified) in rolling element bearings.34

d.

21

22

STP 1623 On Bearing Steel Technologies: 12th Volume

observation that all subsurface failures are related to rolling contact fatigue. Under safe operating conditions, the classical fatigue mechanism dominates, which leads to macro DERs under high-contact pressures. At lower contact pressures, this does not necessarily occur. But at high numbers of revolutions, the bearing will eventually develop cracks from non-metallic inclusions that then can turn into white etching cracks and subsequent failure (depending on the crack propagation rate, if sufficient time for crack face rubbing is available). In the case of premature failure, the same mechanism appears, however, in a different time/load domain. If the application conditions have been calculated to give a certain L10 life, the actual life may be shorter due to either (a) a deviation in actual load or (b) a mechanism that weakens the bearing steel (or both). Deviation in actual load can occur if the application is poorly understood and the assumptions on the loading of the bearing are false. The effect is then a shorted life because the classical fatigue phenomena are accelerated, as mentioned previously. The other option is material weakening. This may occur due to hydrogen that enters the steel and then also accelerates the fatigue phenomena. The end effect is, however, a manifestation of fatigue in the same way as normal fatigue, namely DER development or crack initiation around inclusions (or both) with white etching matter development. In recent years, some industries confronted with the WEC phenomenon have considered FE8-based 81212 (cylindrical roller thrust bearing) tests to qualify lubricants with respect to early crack formation and WECs. In addition, many researchers refer to similar types of tests.36 It can be shown that some lubricants in combination with the FE8-based 81212 test conditions give indications of a hydrogen uptake.36,37 Unfortunately, these tests do not discriminate sufficiently among different oils. As a matter of fact, most WEC failures in FE8 tests have been produced with only a few specific nonrepresentative oil formulations. However, to the authors’ knowledge, nearly all oils used in the field (e.g., in wind gearboxes) where premature failures with WEC have occurred would survive the FE8 81212 test conditions. In short, many WEC-generation mechanism hypotheses are based on FE8 81212 testing that does not reflect the reality in the field. In other words, other WEC-generation mechanisms as observed in FE8 testing may be relevant as well (e.g., see the discussion related to WEA development before crack versus crack before WEA development) depending on the experimental test setup upon which the researchers base their observations. The conclusions given here are based on an analysis of bearings in the field38,39 as well as those bearings used in experimental testing.

Conclusions The work presented in this paper leads to the following conclusions: • Premature bearing failures can have a range of root causes, which can be classified as situations where the load is higher than expected or the material is weakened due to external factors.

d.

VEGTER AND STADLER, DOI: 10.1520/STP162320190054

•

•

•

•

Cracks can initiate at non-metallic inclusions by a nonclassical crack propagation mechanism, which is characterized by material transfer and recrystallization. Small cracks can form butterflies, which then can grow into larger cracks, often indicated as white etching cracks. Despite the mechanisms shown here, some premature failures result from other mechanisms, such as tribochemistry or hydrogen ingress. A classification scheme for root causes of white etching cracks and premature failures was developed. This scheme contributes to finding solutions for premature failure problems.

ACKNOWLEDGMENTS

The authors thank Dr. J. Th. M. de Hosson of Groningen University for help with the high-resolution microscopy pictures shown in this paper. We thank also J. Lai of SKF Research & Technology Development for fruitful discussions.

References 1. 2.

3. 4.

5. 6.

7.

8.

9. 10.

H. K. D. H. Bhadeshia, “Steels for Bearings,” Progress in Materials Science 57, no. 2 (2012): 268–435. M. H. Evans, “White Structure Flaking (WSF) in Wind Turbine Gearbox Bearings: Effects of ‘Butterflies’ and White Etching Cracks (WECs),” Material Science and Technology 28, no. 1 (2012): 3–22. H. Uyama and H. Yamada, White Structure Flaking in Rolling Bearings for Wind Turbine Gearboxes (Alexandria, VA: American Gear Manufacturers Association, 2013). R. Errichello, R. Budney, and R. Eckert, “Investigations of Bearing Failures Associated with White Etching Areas (WEAs) in Wind Turbine Gearboxes,” Tribology Transactions 56 (2013): 1069–1076. A. Greco, S. Sheng, J. Keller, and A. Eridemir, “Material Wear and Fatigue in Wind Turbine Systems,” Wear 302 (2013): 1583–1591. K. Stadler, J. Lai, and R. H. Vegter, “A Review: The Dilemma with Premature White Etching Crack (WEC) Bearing Failure,” in Bearing Steel Technologies: 10th Volume, Advances in Steel Technologies for Rolling Bearings, ed. J. Beswick (West Conshohocken, PA: ASTM International, 2015), 487–508, https://doi.org/10.1520/STP158020140046 R. H. Vegter and J. T. Slycke, “The Role of Hydrogen on Rolling Contact Fatigue Response of Rolling Element Bearings,” Journal of ASTM International 7, no. 2 (2010): 1– 12, https://doi.org/10.1520/JAI102543 K. Stadler and A. Stubenrauch, “Premature Bearing Failures in Industrial Gearboxes,” in Proceedings of Antriebstechnisches Kolloquium ATK 2013 (Aachen, Germany: RWTH, 2013): 113–133. K. Tamada and H. Tanaka, “Occurrence of Brittle Flaking on Bearings Used for Automotive Electrical Instruments and Auxiliary Devices,” Wear 199 (1996): 245–252. I. Strandell, C. Faijers, and T. Lund, “Corrosion—One Root Cause for Premature Failure” (paper presentation, 37th Leeds-Lyon Symposium on Tribology, Leeds Trinity University, Leeds, UK, September 7–10, 2010).

d.

23

24

STP 1623 On Bearing Steel Technologies: 12th Volume

11.

12. 13.

14.

15.

16.

17.

18. 19.

20.

21.

22.

23.

24.

25.

26. 27.

M. Scepanskis, A. Jakovics, I. Kaldre, W. Holweger, B. Nacke, and A. M. Diederichs, “The Numerical Model of Electrothermal Deformations of Carbides in Bearing Steel as the Possible Cause of White Etching Cracks Initiation,” Tribology Letters 59, no. 37 (2015), https://doi.org/10.1007/s11249-015-0564-8 J. Luyckx, “WEC Failure Mode on Roller Bearings” (paper presentation, Wind Turbine Tribology Seminar, NREL Workshop, Broomfield, CO, November 15–17, 2011). J. Gegner, “Frictional Surface Crack Initiation and Corrosion Fatigue Driven Crack Growth” (paper presentation, Wind Turbine Tribology Seminar, NREL Workshop, Broomfield, CO, November 15–17, 2011). M. H. Evans, A. Richardson, L. Wang, and R. Wood, “Serial Sectioning Investigation of Butterfly and White Etching Crack (WEC) Formation in Wind Turbine Gearbox Bearings, Wear 302 (2013): 1573–1582. T. Bruce, E. Rounding, H. Long, and R. S. Dwyer-Joyce, “Characterization of White Etching Crack Damage in Wind Turbine Gearbox Bearings,” Wear 338–339 (2015): 164–177, https://doi.org/10.1016/j.wear.2015.06.008 T. Bruce, H. Long, T. Slatter, and R. S. Dwyer-Joyce, “Formation of White Etching Cracks at Manganese Sulfide (MnS) Inclusions in Bearing Steel Due to Hammering Impact Loading,” Wind Energy (2016), https://doi.org/10.1002/we.1958 T. B. Lund, “Sub-Surface Initiated Rolling Contact Fatigue—Influence of Non-Metallic Inclusions, Processing History, and Operating Conditions,” in Bearing Steel Technology, 8th Volume; Developments in Rolling Bearing Steels and Testing, ed. J. Beswick (West Conshohocken, PA: ASTM International, 2010), 81–96, https://doi.org/10.1520/STP49124S T. W. Wright, The Physics and Mathematics of Adiabatic Shear Bands (Cambridge, UK: Cambridge University Press, 2002). A. Grabulov, U. Ziese, and H. W. Zandbergen, “TEM/SEM Investigation of Microstructural Changes within the White Etching Area under Rolling Contact Fatigue and 3-D Crack Reconstruction by Focused Ion Beam,” Scripta Materialia 57 (2007): 635–638. K. Stadler, D. Vaes, and M. Ersson, “Premature Bearing Failures and White Etching Cracks,” in Proceedings of Antriebstechnisches Kolloquium ATK 2015, ed. G. Jacobs and P. Drichel (Aachen, Germany: RWTH Publications, 2015): 105–120. J. Loos, I. Bergmann, and M. Goss, “Influence of Currents from Electrostatic Charges on WEC Formation in Rolling Bearings,” Tribology Transactions 59, no. 5 (2016): 865–875, https://doi.org/10.1080/10402004.2015.1118582 B. Podgornik, M. Kalin, J. Vizintin, and F. Vodopivec, “Microstructural Changes and Contact Temperatures during Fretting in Steel–Steel Contact,” Journal of Tribology 123, no. 4 (2001): 670–675. J. Lai and K. Stadler, “Investigation on the Mechanisms of White Etching Crack (WEC) Formation in Rolling Contact Fatigue and Identification of a Root Cause for Bearing Premature Failure,” Wear 364–365 (2016): 244–256. D. Scott, B. Loy, and G. H. Mills, “Paper 10: Metallurgical Aspects of Rolling Contact Fatigue,” Proceedings of the Institution of Mechanical Engineers, Conference Proceedings 181, no. 15 (1966), https://doi.org/10.1243/PIME_CONF_1966_181_303_02 H. Harada, T. Mikami, M. Shibata, D. Sokai, A. Yamamoto, and H. Tsubakino, “Microstructural Changes and Crack Initiation with White Etching Area Formation under Rolling/Sliding Contact in Bearing Steel,” ISIJ International 45, no. 12 (2005): 1897–1902. K. Sugino, K. Miyamoto, M. Nagumo, and A. Aoki, “Structural Alterations of Bearing Steels under Rolling Contact Fatigue,” ISIJ International 10 (1970): 98–111. ¨ sterlund, O. Vingsbo, L. Vincent, and P. Guiraldenq, “Butterflies in Fatigued Ball R. O Bearings—Formation Mechanism and Structure,” Scandinavian Journal of Metallurgy 11 (1982): 23–32.

d.

VEGTER AND STADLER, DOI: 10.1520/STP162320190054

28.

29. 30. 31.

32.

33.

34. 35.

36.

37.

38.

39.

A. Grabulov, R. Petrov, and H. W. Zandbergen, “EBSD Investigation of the Crack Initiation and TEM/FIB Analyses of the Microstructural Changes around the Cracks Formed under Rolling Contact Fatigue (RCF),” International Journal of Fatigue 32, no. 3 (2010): 576–583. A. Grabulov, “Fundamentals of Rolling Contact Fatigue” (PhD thesis, Delft University of Technology, Materials Innovation Institute, 2010). A. Voskamp, “Microstructural Changes during Rolling Contact Fatigue” (PhD thesis, Delft University of Technology, 1996). A. P. Voskamp and E. J. Mittemeijer, “Crystallographic Preferred Orientation Induced by Cyclic Rolling Contact Loading,” Metallurgical and Materials Transactions A 27 (1996): 3445–3465, https://doi.org/10.1007/BF02595437 R. H. Vegter, H. Krock, Y. Kadin, and V. Ocelik, “Nonmetallic Inclusion Bonding in Bearing Steel and the Initiation of White-Etching Cracks,” in Bearing Steel Technologies: 11th Volume, Progress in Steel Technologies and Bearing Steel Quality Assurance, ed. J. M. Beswick (West Conshohocken, PA: ASTM International, 2017), 519–532, http://dx.doi.org/ 10.1520/STP160020160145 W. Solano-Alvarez and H. K. D. H. Bhadeshia, “White-Etching Matter in Bearing Steel. Part II: Distinguishing Cause and Effect in Bearing Steel Failure,” Metallurgical and Materials Transactions A 45A (2014): 4916–4931. K. Stadler, R. H. Vegter, and D. Vaes, “White-Etching Cracks,” SKF Evolution 1 (2018): 21–29. S. W. Ooi, A. Gola, R. H. Vegter, P. Yan, and K. Stadler, “Evolution of White-Etching Cracks and Associated Microstructural Alterations during Bearing Tests,” Materials Science and Technology 33, no. 14 (2017), 1657–1666, https://doi.org/10.1080/ 02670836.2017.1310431 A. D. Richardson, M. H. Evans, L. Wang, M. Ingram, Z. Rowland, G. Llanos, and R. J. K. Wood, “The Effect of Over-Based Calcium Sulfonate Detergent Additives on White Etching Crack (WEC) Formation in Rolling Contact Fatigue Tested 100Cr6 Steel,” Tribology International 133 (2019): 246–262, https://doi.org/10.1016/j.triboint.2019.01.005 F. G. Guzman, M. Ozel, G. Jacobs, G. Burghardt, C. Broeckmann, and T. Janitzky, “Reproduction of WECs under Rolling Contact Loading on Thrust Bearing and Two-Disc Test Rigs,” Wear 390–391 (2017): 23–32, http://dx.doi.org/10.1016/j.wear.2017.06.020 B. Gould, A. Greco, K. Stadler, and X. Xiao, “An Analysis of Premature Cracking Associated with Microstructural Alterations in an AISI 52100 Failed Wind Turbine Bearing Using X-Ray Tomography,” Materials and Design 117 (2017): 417–429, http://dx.doi.org/ 10.1016/j.matdes.2016.12.089 B. Gould, A. Greco, K. Stadler, R. H. Vegter, and X. Xiao, “Using Advanced Tomography Techniques to Investigate the Development of White Etching Cracks in a Prematurely Failed Field Bearing,” Tribology International 116 (2017): 362–370.

d.

25

26

BEARING STEEL TECHNOLOGIES: 12TH VOLUME

STP 1623, 2020 / available online at www.astm.org / doi: 10.1520/STP162320190084

Toni Blass,1 Xiaohong Xu,2 Kirsten Wunder,1 Werner Trojahn,1 Ke Geng,2 and Feng Li2

Influence of Steel Cleanliness on RCF and WEC Formation Citation T. Blass, X. Xu, K. Wunder, W. Trojahn, K. Geng, and F. Li, “Influence of Steel Cleanliness on RCF and WEC Formation,” in Bearing Steel Technologies: 12th Volume, Progress in Bearing Steel Metallurgical Testing and Quality Assurance, ed. J. M. Beswick (West Conshohocken, PA: ASTM International, 2020), 26–49. http://doi.org/10.1520/STP1623201900843

ABSTRACT

It is well known that the amount and size of nonmetallic inclusions (NMIs) in bearing steel determines its performance in terms of classical rolling contact fatigue (RCF), especially if the bearing is subjected to high loads. Therefore, decreasing the NMI content in steel is seen as a key factor in prolonging the lifetime of bearing parts. Furthermore, it is presumed that the white etching cracks (WECs) formed in bearings as a consequence of so-called “additional loads” acting during time in service (e.g., electrical current, strong dynamics, critical additives) originate at very small size NMIs. Therefore, it is speculated in the literature that a cleaner steel should have a positive impact on this failure mechanism as well. In this study, bearing component tests with steels exhibiting different levels of cleanliness have been conducted. To start, classical RCF tests were performed, which are known to be highly sensitive to NMI content in the material. In addition, WEC tests with defined WEC-provoking additional loads were conducted to identify the role of NMIs for this nonclassical fatigue mechanism. It is observed that the microscopic cleanliness strongly affects the lifetime if classical material fatigue is the dominating failure mode, and that

Manuscript received June 28, 2019; accepted for publication October 21, 2019. 1 Schaeffler Technologies AG & Co. KG, Georg-Schaefer Str. 30, Schweinfurt, 97421, Germany T. B. https://orcid.org/0000-0002-1286-8144 2 Jiangyin Xingcheng Special Steel Works. Co., Ltd., No. 297, Binjiang East Road, Jiangyin City, 214400, http://orcid.org/0000-0003-3233-3072, K. G. https://orcid.org/0000-0002-9701P.R. China X. X. 2694, F. L. https://orcid.org/0000-0002-0057-6267 3 ASTM 12th International Symposium on Rolling Bearing Steel: Progress in Bearing Steel Metallurgical Testing and Quality Assurance on May 15–17, 2019 in Denver, CO, USA. C 2020 by ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959. Copyright V

d.

BLASS ET AL., DOI: 10.1520/STP162320190084

so-called superclean steels such as XCS3 have superior performance compared to standard qualities. In contrast, the formation of WECs is unaffected from the level of microscopic cleanliness. Keywords rolling contact fatigue, white etching cracks, nonmetallic inclusions, steel cleanliness

Introduction STEEL CLEANLINESS AND IMPROVEMENT IN MANUFACTURING

Nonmetallic inclusions (NMIs) belong to commercial steel as a result of several manufacturing steps in which molten steel comes in contact with ceramic parts of production facilities (e.g., refractories). In addition, internal gas reactions may form some inclusions. Therefore, several parameters determine the cleanliness of the material in terms of size, amount, and type of the NMIs. It is a well-known fact that one of the dominant criteria for the quality of bearing steel is the cleanliness of the steel, which is influenced by its oxygen (O) content and directly linked to oxidic inclusions. Significant efforts have been undertaken to lower the content of oxide inclusions by reducing the oxygen content of the steel. A lowered oxygen content usually results in a higher fatigue duration.1,2 Similar relationships can also be obtained from bearing parts; therefore, low content and size of inclusions are beneficial for the endurance of those components.3,4 Low oxygen content not only extends the lifetime of the parts but also increases the dynamic load rating (dynamic load capacity) of bearings, as shown in figure 1. As an example, cylindrical roller bearing NU1010 data were collected from various official bearing catalogues from company FAG / Schaeffler from past years. The decrease of the total oxygen level in the steel from 20 to 30 ppm in the 1950s to the current state of 4 to 15 ppm has drastically increased the dynamic load rating for rolling bearings. As it is not only the oxides that have an impact on the lifetime of bearings, similar efforts have been taken to decrease the sulfur (S) content and subsequently sulfidic inclusions.5 INFLUENCE OF MATERIAL CLEANLINESS ON ROLLING CONTACT FATIGUE

Inclusions are weak points within the microstructure of bearing parts. Macroscopic inclusions can cause premature component failures if located in the overrolled volume. Minimizing these macro inclusions should increase the reliability of bearings. Additionally, and pronounced at higher Hertzian loads, smaller microscopic inclusions may act as the origin of cracks, while their size and type steer the criticality. The dependence of the number of overrollings was systematically investigated by Bo¨hm et al.6 Several steel batches with different contents of nonmetallic inclusions using angular contact ball bearings type 7205 were tested. Bo¨hm et al.’s study resulted in a correlation of B10 life with cleanliness according to SEP 1570 (predecessor of the German standard DIN 50602); whereas a modified K4 value as a sum d.

27

28

STP 1623 On Bearing Steel Technologies: 12th Volume

FIG. 1 Increase in dynamic load rating and rating life of NU1010, taken from catalogue values and calculated rating life.

of K4 (oxides) and 1/3 K4 (sulfides) gave the smallest scattering area (fig. 2). The factor 1/3 for the sulfidic inclusions refers to the difference in the harmfulness observed for the different inclusion types. INFLUENCE OF MATERIAL CLEANLINESS ON WHITE ETCHING CRACK FORMATION

White etching crack (WEC) formation is a unique failure mode because it leads to premature failures of bearings way before they reach their calculated lifetimes. Several studies have been published in order to identify the root cause and influencing factors for WEC formation. The papers of Evans contain comprehensive overviews.7,8 Loos et al.9 postulated that so-called additional loadings, such as the influence of electrical currents, strong dynamics in the system, and so on can lead to WEC formation and published a “chain of events” as a hypothesis explaining how the additional loadings can finally result in the formation of WECs in the subsurface of bearing elements. The authors also discussed the strong influence of lubricants on the formation of WECs. Despite the great effort taken to understand this kind of failure mode, there are still a lot of open questions. For example, is the white matter or the crack the first feature of this failure mechanism? Since the WEC failure mode has become a big issue in the bearing industry, several researchers have also tried to identify the role of NMIs in WEC formation. Evans et al.10 and Richardson et al.11 analyzed bearing parts affected with WECs to identify the possible origin of WEC structures. They applied serial sectioning on d.

BLASS ET AL., DOI: 10.1520/STP162320190084

FIG. 2 Plot of load cycles (B10 life) obtained on different batches of 100Cr6 over the K4mod-value as the sum of K4 (oxides plus 1/3 sulfides).6

washers and rollers out of cylindrical roller thrust bearings (type 81212). Both studies concluded that butterflies and WECs can originate or interact with small inclusions (fig. 3, extracted from Richardson et al.11). Missing in this work is an inclusion rating according to common cleanliness measurement standards and a benchmark with tolerance limits given for bearing steels. For instance, the globular oxide inclusions in figure 3 are rated according to DIN 5060212 and would be counted as size class K0 or even less (K0 rating for inclusions, which are within an area range of 71 lm2 to 141 lm2). Gould et al.13 recently published results on computer-tomography (CT) measurements of field complaints affected by a large number of WEC structures. Exemplarily, Gould et al. presented four different inclusions, with their threedimensional tomography, reconstructed out of the CT scans. Each of them was supposed to be an initiation site for cracks and white etching structures. Inclusion classification was also missing in this study. Unfortunately, the rolling direction of the ring was not available, so a standard conformance assessment of inclusion rating was not possible. But to estimate the rating of the four inclusions, the largest dimension of the inclusions and the thickness were chosen to calculate the area of d.

29

30

STP 1623 On Bearing Steel Technologies: 12th Volume

FIG. 3 Etched microcuts of white etching cracks, originated at or interacting with small globular NMIs, found in the rollers after testing for 18 h on WEC test rig FE8.11 The numbers in the brackets indicate the size in the two visible directions or the equivalent diameter if only one number is given. Therefore, the rating, according to DIN 50602, can be derived.12 (The figure is reproduced without modification from Richardson et al.11)

the inclusions, respectively. With such a methodology, all inclusions shown in 2 2 figure 4 were rated as K2, having areas within the limits of 283 lm to 566 lm . Basically, the inclusions for air-melted and vacuum-degassed bearing steels are limited in terms of their oxidic K4 level (for example, as required in ISO68317:199914). So only inclusions having an area larger than 1,131 lm2 are counted

FIG. 4 Three-dimensional tomography reconstructions of the inclusions that initiated cracks and WEC structures.13 Taking the largest dimension and the thickness Tmax as the base for size calculation, all the inclusions would be rated as K2.

d.

BLASS ET AL., DOI: 10.1520/STP162320190084

and taken into assessment. The finding that inclusions can act as starting points for WECs is obvious in the former mentioned studies, but all inclusions found are much smaller than the tolerance limits. Loos et al.9 studied the influence of different remelted steel on the WEC lifetime and presented results obtained with axial cylindrical thrust bearings tested on an FE8 test rig and those of cylindrical roller bearings NU222 tested on an R4G test rig. Both the tests were conducted under WEC critical environments. The authors could not find any difference in the lifetimes due to WEC failure between bearings made from remelted steel in comparison to bearings made from standard bearing steel.

Experimental CHARACTERIZATION OF STEEL 100CR6 (EQUIVALENT TO SAE 52100) BATCHES INCLUDING THE CLEANLINESS ASSESSMENT

In total, six different batches from 100Cr6, produced by conventional air-melting and vacuum-degassing, were used for this study. For the rolling contact fatigue (RCF) test, five batches of 100Cr6 were chosen, with different degrees of microscopic cleanliness. These test samples included two batches of normal quality (Batch A, used as a baseline, and Batch C) and three different batches of pure steels (Batch B and two batches of “super-clean steels”: XCS3-I and XCS3-II). The diameters of the steel bars varied between 50 mm and 60 mm. Another super-clean steel, XCS3-III, was chosen to be tested for WEC performance. This steel batch had a diameter of 120 mm in order to manufacture the washers for cylindrical roller thrust bearing type 81212. All of the steels investigated fulfill the requirements of the ISO 683-17:201415 specification in terms of their chemical composition. Table 1 focuses on the content of S, O, and aluminum (Al) because they are steering factors for the cleanliness of the materials. Additionally, the sum of nickel (Ni) and copper (Cu) are listed to get information about the content of scrap used for steel production. The most common methods to determine the cleanliness and to rate the batches according to raw-material specifications are DIN 5060212 (mainly used in

TABLE 1 Overview of 100Cr6 batches, used for the RCF and WEC tests

Test

Batch

Manufacturing

RCF

A (Baseline)

Continuous casting

B

Ingot casting

C

Continuous casting

WEC

Cu þ Ni [  102%]

Al [  103%]

O [  104%]

S [  103%]

10

3

12.0

4

23

22

3.8

1

8

5

8.0

6 1

XCS3-I

Continuous casting

8

28

4.2

XCS3-II

Continuous casting

8

25

4.3

1

XCS3-III

Continuous casting

3

24

3.7

3

d.

31

32

STP 1623 On Bearing Steel Technologies: 12th Volume

Germany) and ASTM E45:2018, Standard Test Methods for Determining the Inclusion Content of Steel.16 DIN and ASTM mainly differentiate between sulfidic NMI types (SS or A) and oxidic NMI types (OA, OS, OG, or B, C, D). For this study, microscopic cleanliness was assessed on the raw-material bars according to the German standard DIN 50602, Method K, counting sulfidic as well as oxidic inclusions starting with size K0. Among other advantages, this evaluation method is based on the assessment of inclusion area, rather than the field-based ASTM E45. To have a suitable statistic, a total area of 4,800 mm2 was divided into 24 cross sections for investigation. The inclusions found were further statistically analyzed using extreme value analysis according to ASTM E2283, Standard Practice for Extreme Value Analysis of Nonmetallic Inclusions in Steel and Other Microstructural Features,17 which says that 24 single values for each type of inclusion are needed for good statistical certainty. Results of the cleanliness assessment can be seen in table 2, where the K0 to K4 sum values are listed for the sum of oxide and sulfide inclusions and for the oxide inclusions alone. Within the test batches, a significant spread in the microscopic cleanliness can be seen. Very good results were obtained with the XCS3 Batches I, II, and III, and with Batch B. For comparison, the ISO 683-17:199914 standard announced a K1 (oxides) value of less than or equal to six for electro-slag remelted (ESR) qualities and of a K1 (oxides) less than or equal to one for vacuum-induction melted (VIM) and vacuum-arc remelted (VAR) steel qualities. Within these samples, Batches A and C have the worst degree of cleanliness, whereas the latter ones also showed K4-rated inclusions. As indicated by the higher levels of sulfur, the degree of cleanliness is steered by the sulfide in Batches A and C, whereas the B and XCS3 batches have almost no sulfidic inclusions. If all the inclusions found in the investigated area of 4,800 mm2 are normalized to this area, one can obtain a better image of the inclusions in the steel batches, as seen in figure 5 and figure 6. Batch XCS3 has the best cleanliness level, followed by Batches B, A, and C.

TABLE 2 Results of cleanliness assessment according to DIN 50602, Method K

KX (oxides þ sulfides) Test

Batch

K0

K1

K2

RCF

A (Baseline)

WEC

KX (oxides)

K3

K4

K0

K1

K2

K3

K4

39.1

23.3

11.2

3.4

0.0

1.5

0.4

0.2

0.0

0.0

B

4.2

1.3

0.3

0.1

0.0

4.2

1.3

0.3

0.1

0.0

C

89.8

50.7

20.7

5.5

0.8

5.0

2.3

0.9

0.4

0.2

XCS3-I

0.9

0.7

0.4

0.2

0.0

0.9

0.7

0.4

0.2

0.0

XCS3-II

0.6

0.3

0.1

0.0

0.0

0.6

0.3

0.1

0.0

0.0

XCS3-III

0.6

0.3

0.0

0.0

0.0

0.3

0.2

0.0

0.0

0.0

d.

BLASS ET AL., DOI: 10.1520/STP162320190084

FIG. 5 Plot of the sum of oxidic and sulfidic inclusions from size classes K0 to K4 per mm2 for the different steel batches.

FIG. 6 Plot of the oxidic inclusions from size classes K0 to K4 per mm2 for the different steel batches.

d.

33

34

STP 1623 On Bearing Steel Technologies: 12th Volume

The analysis of the inclusions according to ASTM E2283 revealed the picture shown in figure 7. Here, the calculated value (99% probability) of the longest inclusion is shown for the different types of inclusions. The calculation was also done for the largest area (99% probability) of inclusion, as seen in figure 8. RCF TESTING

Several methods for determining material reliability have been devised over the years. Starting from the simple ball-on-rod or ball-on-washer test to realistic bearing tests, a plethora of different test rigs are available. One of the test stages able to balance between short, reproducible tests and low, mounting and dismounting effort is the L17 test rig. This test rig was introduced in the 1970s by Bo¨hm et al.6 and since then has been further developed to optimize the reproducibility of results by slight modifications on load, lubrication, and test bearings. Figure 9 presents a schematic sketch of the test rig. Two angular contact ball bearings (type 7205B) are tested in parallel, whereas the inner rings (IRs) were made of the different steel batches (turning out of bars, no intermediate forging step) and used as the primary tested part. By varying the lubricant viscosity and therefore the relative lubricating film thickness (kappa value), different damage mechanisms can be triggered. In particular, the surface separation (elasto-hydrodynamic lubrication, or EHL) leading to subsurface fatigue and the surface contact (mixed friction) resulting in surface-initiated fatigue. The parameters for testing can be found in table 3. With the aim of generating

FIG. 7 Bar chart of maximum length (99% probability) of the four inclusion types, rated.

d.

BLASS ET AL., DOI: 10.1520/STP162320190084

FIG. 8 Bar chart of the maximium area (99% probability) of the four inclusion types, rated.

FIG. 9 Schematic sketch of test rig L17.

d.

35

36

STP 1623 On Bearing Steel Technologies: 12th Volume

TABLE 3 Parameters chosen for the RCF tests on the L17 test rig

Surface Separation (Medium Load)

Test target

Surface Separation (High Load)

Subsurface fatigue

Surface Contact

Surface fatigue

7205B (martensite, about 60 HRC, retained austenite  5%)

Bearing type Rotational speed

12,000 rpm

Hertzian pressure

2,900 MPa

3,500 MPa

2,630 MPa

 80 C

Temperature Kappa ratio

2.2

0.26

subsurface failures, one medium load test was set up with 2,900 MPa and another with a higher load of 3,500 MPa. The test to trigger surface-initiated damages was performed at a load of 2,630 MPa. All tests were conducted in sudden-death mode, which means that the damaged bearing was counted as a failure, whereas the other bearing (tested in parallel) was rated as a suspended part (failure probability of 0% in the statistics). WEC TESTING

The FE8 test rig type, operating with special settings, is one of the most frequent test rigs used to produce WECs in bearings within a reasonable time. This test rig was originally invented to test lubricants for their lubricity and is standardized in DIN 51819-1:2016.18 Theoretically, a broad range of different rolling bearings can be used. In our study, the cylindrical roller thrust bearing type 81212 was chosen. The test rig setup itself is quite simple (fig. 10 and fig. 11): The force is applied by a spring, and the rotational speed is generated by an electric motor. Lubricant is supplied by a closed loop system with temperature control and regulation. During the test, different signals are monitored for torque and temperature control, as well as noise and acceleration for failure detection. It has been discovered that WEC formation can be triggered within 40 h by application of specific conditions, as seen in table 4. Comparable to previous studies,19 a low-viscosity automotive gearbox oil, ISO VG 68, was used. This oil consists of polyalphaolefin base oil and a combination of 1.4% calcium sulfonate and 2.6% overbased zinc dithiophosphate. This combination was identified previously to provoke WEC formation on FE8 test rigs.20 The kappa value of 0.46 indicates mixed friction operation conditions. The modified reference rating life Lhmr of 163 h is usually exceeded in conventional bearing lifetime tests without WEC formation. At the end of the test, which may be either due to excessive noise or acceleration, or completion of the predefined maximum test duration time, the washers are investigated for WECs using ultrasound technology (UT) with a test frequency of 25 MHz. This method allows precise localization of subsurface reflectors and thus helps to identify subsurface WEC structures by microscopic sectioning. d.

BLASS ET AL., DOI: 10.1520/STP162320190084

FIG. 10 Schematic sketch of FE8 test rig.

FIG. 11 Photo of FE8 test rig.

d.

37

38

STP 1623 On Bearing Steel Technologies: 12th Volume

TABLE 4 FE8 test conditions, adapted from Blass, Trojahn, and Merk.19

81212 (martensite, about 60 HRC, retained austenite  5%)

Bearing type

Rotational speed

750 rpm

Applied load

60 kN

Hertzian pressure

2,075 MPa

Lubrication viscosity class

ISO VG 68

Kappa ratio

0.46

Lhmr acc. ISO/TS 16281

163 h

Results ROLLING CONTACT FATIGUE TESTS ON L17 WITH SURFACE SEPARATION AND 2,900 MPA HERTZIAN PRESSURE

The results of the RCF tests, conducted under surface separation conditions and 2,900 MPa load, can be seen in the Weibull plot in figure 12. Only IRs that exhibited subsurface failures were counted as failures in the Weibull evaluation. Bearing tests that were stopped due to other reasons (e.g., failure of balls or outer rings) were

FIG. 12 Weibull plot of the results of RCF tests under surface separation with a pressure of 2,900 MPa.

d.

BLASS ET AL., DOI: 10.1520/STP162320190084

rated as suspended. The calculated modified reference rating life Lhmr is noted in the chart as well. All variants tested show a Bh10 lifetime (running time in hours, where the Weibull lines cross the 10% unreliability horizontal line) higher than this calculated reference lifetime. This indicates that the calculation method itself is reliable. Looking at the results, in particular, a spread on the results of more than half an order of magnitude can be seen. Although no one-to-one correlation exists between the Bh10 lifetimes and the cleanliness ratings, the general trend is that the Bh10 lifetimes increase with better cleanliness. Without doubt, the cleanest steels, XCS3-I and XCS3-II, exhibited the best performance in the test. ROLLING CONTACT FATIGUE TESTS ON L17 WITH SURFACE SEPARATION AND 3,500 MPA HERTZIAN PRESSURE

For these tests, only the Batch A steel and the XCS3-II steel batch were tested. Again, only the IRs that exhibited subsurface damages were counted as failures. The calculated modified reference rating life Lhmr is still exceeded by both variants (fig. 13). In comparison to the test at 2,900 MPa, the difference in the running times Bh10 of Batches A and XCS3-II is somewhat decreased. In the microcuts through the IRs displayed in figure 14, pronounced material alterations could be found, even in those IRs from XCS3-II having the shortest running time. These flat white bands

FIG. 13 Weibull plot of the results of the RCF tests under surface separation with a Hertzian pressure of 3,500 MPa.

d.

39

40

STP 1623 On Bearing Steel Technologies: 12th Volume

FIG. 14 Microcut through the earliest failure of XCS3-II, showing pronounced flat white bands.

could also act as weak points in the material (hardness of about 600 HV, compared to the matrix having 700 HV) and therefore could act as possible origins of failures. This means that besides the influence of NMIs, the effect of material alterations can also contribute to the failures. ROLLING CONTACT FATIGUE TESTS ON L17 WITH MIXED FRICTION AND 2,630 MPA

Similar to the tests with surface separation and 3,500 MPa Hertzian pressure, these tests were conducted with steel Batches A and XCS3-II, only. The failed IRs in these tests showed pittings that originated at the surface but also subsurfaceinduced ones. Again, the calculated modified rating life is exceeded with both sample batches (fig. 15). Applying this test setup, the difference in the Bh10 lifetimes between both tested batches is much more pronounced compared to the test with surface separation. A ratio of 11 was achieved within these tests. This is somewhat contrary to consensus but can be explained by the findings of Dinkel and Trojahn.21 Dinkel and Trojahn showed that under surface contact (i.e., mixed friction conditions), the sulfidic cleanliness plays a significant role. Comparing the sulfidic cleanliness of the two tested batches, Batch A has a sum value K1 (sulfides) of 19.9, whereas Batch XCS3-II has a sum value K1 (sulfides) of 0. Also, in the former batch, sulfidic inclusions having a K3 size were present. In some of the failed IRs from Batch A, residues of sulfide stringers were identified in the origin of the pittings, as seen in figure 16. d.

BLASS ET AL., DOI: 10.1520/STP162320190084

FIG. 15 Weibull plot of the results of the RCF tests performed under surface contact with a pressure of 2,630 MPa.

FIG. 16 Pitting in a sample from Batch A, the residue of a manganese sulfide (MnS) stringer was identified in the origin (indicated by the arrows).

d.

41

42

STP 1623 On Bearing Steel Technologies: 12th Volume

WHITE ETCHING CRACK FORMATION ON FE8

On the FE8 tests with the WEC creating additional load (mixed friction in conjunction with special lubricant) the steel Batch XCS3-III was tested. For comparison, several tests conducted with commercially available axial cylindrical thrust bearings (Type 81212) made from standard 100Cr6 (cleanliness according to ISO 68317:1999) were taken. Some of them were already published in Blass, Trojahn, and Merk.19 From the Weibull plot in figure 17, it can be observed that in contrast to experiments on L17, the calculated reference rating life Lhmr cannot be achieved by either of the tested variants. This mirrors the WEC failure mode as a premature failure. The Weibull curves of the XCS3-III samples and those of the standard material are almost parallel. This indicates that no influence of material cleanliness could be found within this test. Since all the samples tested underwent a detailed damage analysis as described previously, no differences in the density and appearance of the WEC structures could be found (figs. 18 and 19). By applying serial sectioning, as performed in Richardson et al.,11 in rare cases, small NMIs could be identified as an origin of the WEC structures. Mostly, they were identified as globular oxides and had a size of less than K0 (fig. 20).

FIG. 17 Weibull plot of the WEC tests on FE8.

d.

BLASS ET AL., DOI: 10.1520/STP162320190084

FIG. 18 WEC structure out of a 100Cr6 standard sample.

FIG. 19 WEC structure out of a sample from XCS3-III.

d.

43

44

STP 1623 On Bearing Steel Technologies: 12th Volume

FIG. 20 Microcut of a washer from Batch XCS3-III with WEC starting from a small oxide inclusion with a diameter of 8 μm; it is, therefore, rated to have a size less than K0.

Discussion RCF TESTING

The results of classical RCF testing shows a clear tendency of better cleanliness leading to higher lifetime to failure. To rate the results of the present work with those obtained from Bo¨hm et al.,6 it is necessary to convert the Bh10 lifetime in hours to the B10 life in load cycles for each variant tested at 2,900 MPa by equation (1): B10 ¼

Bh10  ½nðIRÞ  nðcageÞ  z  60 106

(1)

where: B10 ¼ B10 in load cycles, Bh10 ¼ B10 in hours, n(IR) ¼ rotational speed of the inner ring in rpm (12,000 rpm), n(cage) ¼ rotational speed of the cage in rpm, and z ¼ number of rolling elements in the bearing (13 balls). A conversion of load cycles obtained with the current tests at 2,900 MPa and those conducted at 2,600 MPa by Bo¨hm et al.6 can be conservatively estimated (equations [2] and [3]) with a stress-life exponent of 9 as proposed by Zaretsky.22 d.

BLASS ET AL., DOI: 10.1520/STP162320190084

  B10ð2600MPaÞ 2900MPa 9 ¼ B10ð2900MPaÞ 2600MPa   2900MPa 9  B10ð2900MPaÞ B10ð2600MPaÞ ¼ 2600MPa

(2)

(3)

Thus, a factor of 2.67 is used to convert the theoretical B10 life of a test at 2,900 MPa to results obtained at 2,600 MPa load as in equation (4). B10ð2600MPaÞ ¼ 2:67  B10ð2900MPaÞ

(4)

By means of this calculation, the results obtained in this study can be compared with the previous results from Bo¨hm et al.6 In figure 21, the data are plotted together. In the 1970s, a good correlation of the B10 life with the K4 (oxides plus 1/3 sulfides) was obtained. Such a correlation was not applicable for the present results because only Batch C exhibited K4-rated inclusions. Futhermore, it can also be observed from the results that all of the latest tested steel batches exhibited a longer B10 life than any of the batches tested in the 1970s. This shows the improvement in steel quality over the past decades and is a further justification for the increased power density of current bearings in comparison to those previously used.

FIG. 21 Plot of load cycles (B10 life) over the K4mod-value as the sum of K4 (oxides plus 1/3 sulfides). Original data from Bo ¨ hm et al.6 and current results are matched.

d.

45

46

STP 1623 On Bearing Steel Technologies: 12th Volume

So, to account for the increased cleanliness, smaller inclusion sizes must be considered. Therefore, it is proposed to use the K1 values to rate the different tested batches. Subsequently, the B10 life of the current tests, conducted at a Hertzian pressure of 2,900 MPa, is correlated to the K1 mod values of the steel batches as the sum of K1 (oxides) plus 1/3 K1 (sulfides). The resulting chart can be seen in figure 22, where an inverse square root dependency gives a good trend. WEC TESTING

To explain the results of WEC testing on the FE8 test rig, which in general do not show an influence of the steel batches tested, one has to provide details about the amount of potential initiation points for the white etching structures present in the stressed volume. As identified by several researchers10,11 as well as is seen in the present study (fig. 20), very small inclusions can act as starting points for WECs. Holappa and Helle5 calculated the number of globular oxide inclusions out of the total oxygen content in the steel, presuming that all oxygen is bound in the inclusions and that those inclusions are of a perfectly spherical shape. So, for example, steels containing 5 ppm of oxygen theoretically have a number of 3  104 oxide inclusions with a size of 10 lm in a volume of 1 cm3. Even if the total oxygen in the steel can be lowered to 1 ppm, which is far below the current technical limits, the content of inclusions would not decrease below 103 inclusions per cm3 (fig. 23).

FIG. 22 Plot of load cycles (B10 life) of the five steel batches tested in this study at surface separation and 2,900 MPa pressure over the K1mod-value as the sum of K1 (oxides) plus 1/3 K1 (sulfides).

d.

BLASS ET AL., DOI: 10.1520/STP162320190084

FIG. 23 Calculated density of globular inclusions with different sizes with respect to the total oxygen in the steel. Dotted lines indicate the density of 10-μm sized inclusions in dependency on total oxygen content. Data taken from Holappa and Helle.5

Of course, this is a theoretical and simplified calculation; nevertheless, it can explain that if bearings are used in a WEC critical environment, there are always enough possible initiation sites for the WEC development. And, depending on the bearing size, the stressed volume can exceed several tens or hundreds of cm3.

Conclusion •

• •

•

•

Cleanliness of XCS3 steel produced according to conventional production (air-melted and vacuum-degassed) can meet the requirements of remelted steel grades. Microscopic cleanliness has a strong influence on bearing life. The steels with better cleanliness exhibit longer classical fatigue lifetimes during surface separation testing and also during surface contact testing. K4 values are not a suitable base for characterization of modern steels. Rather, attention should be drawn to smaller-sized inclusions. Therefore, the rating of K1 is proposed for clean steels. Under surface contact (mixed friction) conditions, the difference in fatigue lifetime observed between standard clean steel (Batch A) and especially clean steel (Batch XCS3-II) is greater than under surface separation (EHL).

d.

47

48

STP 1623 On Bearing Steel Technologies: 12th Volume

This can be explained mainly by the differences in the sulfidic cleanliness, as published previously.21 • The cleanest steels available did not exhibit an increased WEC resistance in bearing tests on FE8. • Even the cleanest steels have a huge number of (very) small-sized inclusions below or near the detection limit that can act as initiation sites for WECs. • In a WEC-critical environment (i.e., when so-called “additional loads” are applied on the bearings), the content and the size of NMIs did not influence their lifetimes. This validates previous findings.9 In order to increase the performance of bearing parts further, enabling for higher dynamic load rating and potential for downsizing, it is beneficial to improve the steel cleanliness. However, as proven under varying tests conditions, it will not be a countermeasure to prevent WEC formation.

References 1.

2. 3.

4.

5. 6.

7.

8.

9.

10.

T. Masuda and Y. Kato, “Production and Evaluation of High Cleanliness Steel,” in 14th Process Technology Conference Proceedings (Orlando, FL: Iron & Steel Society, Process Technology Division, 1995), 101–108. Y. Murakami, Metal Fatigue: Effects of Small Defects and Nonmetallic Inclusions (Oxford, UK: Elsevier, 2002). M. Kinoshi and A. Koyanagi, “Effect of Nonmetallic Inclusions on Rolling-Contact Fatigue Life in Bearing Steels,” in Bearing Steels: The Rating of Nonmetallic Inclusion, ed. J. Hoo, P. Kilhefner, and J. Donze (West Conshohocken, PA: ASTM International, 1975), 138–149, https://doi.org/10.1520/STP32290S N. Tsunekage, K. Hashimoto, T. Fujimatsu, and K. Hiraoka, “Initiation Behavior of Crack Originated from Non-Metallic Inclusion in Rolling Contact Fatigue,” Journal of ASTM International 7, no. 2 (2010): 1–9, https://doi.org/10.1520/JAI102612 L. Holappa and A. Helle, “Inclusion Control in High-Performance Steels,” Journal of Materials Processing Technology 53, nos. 1–2 (1995): 177–186. K. Bo ¨hm, H. Schlicht, O. Zwirlein, and R. Eberhard, “Nonmetallic Inclusions and Rolling Contact Fatigue,” in Bearing Steels: The Rating of Nonmetallic Inclusion, ed. J. Hoo, P. Kilhefner, and J. Donze (West Conshohocken, PA: ASTM International, 1975), 96–113, https://doi.org/10.1520/STP32288S M.-H. Evans, “White Structure Flaking (WSF) in Wind Turbine Gearbox Bearings: Effects of ‘Butterflies’ and White Etching Cracks (WECs),” Material Science and Technology 28, no. 1 (2002), https://doi.org/10.1179/026708311X13135950699254 M.-H. Evans, “An Updated Review: White Etching Cracks (WECs) and Axial Cracks in Wind Turbine Gearbox Bearings,” Materials Science and Technology 32 (2016), https:// doi.org/10.1080/02670836.2015.1133022 J. Loos, T. Blass, J. Franke, W. Kruhoeffer, and I. Bergmann, “Influences on Generation of White Etching Crack Networks in Rolling Bearings,” Journal of Mechanical Engineering and Automation 6, no. 2 (2016): 85–94. M.-H. Evans, A. D. Richardson, L. Wang, R. J. K. Wood, and W. B. Anderson, “Confirming Subsurface Initiation at Non-Metallic Inclusions as One Mechanism for White Etching Crack (WEC) Formation,” Tribology International 75 (2014): 87–97.

d.

BLASS ET AL., DOI: 10.1520/STP162320190084

11.

12.

13.

14. 15. 16.

17.

18.

19.

20.

21.

22.

A. D. Richardson, M.-H. Evans, L. Wang, R. J. K. Wood, M. Ingram, and B. Meuth, “The Evolution of White Etching Cracks (WECs) in Rolling Contact Fatigue-Tested 100Cr6 Steel,” Tribology Letters 66, no. 6 (2018), https://doi.org/10.1007/s11249-017-0946-1 Microscopic Examination of Special Steels Using Standard Diagrams to Assess the Content of Non-Metallic Inclusions, DIN 50602, withdrawn standard (Berlin, Germany: Beuth Verlag, 1985). B. Gould, A. Greco, K. Stadler, E. Vegter, and X. Xiao, “Using Advanced Tomography Techniques to Investigate the Development of White Etching Cracks in a Prematurely Failed Field Bearing,” Tribology International 116 (2017): 362–370. Heat-Treated Steels, Alloy Steels and Free-Cutting Steels—Part 17: Ball and Roller Bearing Steels, ISO 683-17 (Geneva: International Standards Organization, 1999). Heat-Treated Steels, Alloy Steels and Free Cutting Steels—Part 17: Ball and Roller Bearing Steels, ISO 683-17 (Geneva: International Standards Organization, 2014). Standard Test Methods for Determining the Inclusion Content of Steel, ASTM E45-18a (West Conshohocken, PA: ASTM International, approved June 1, 2018), https://doi.org/ 10.1520/E0045-18A Standard Practice for Extreme Value Analysis of Nonmetallic Inclusions in Steel and Other Microstructural Features, ASTM E2283-08(2019) (West Conshohocken, PA: ASTM International, approved November 1, 2019), https://doi.org/10.1520/E2283-08R19 Testing of Lubricants—Mechanical-Dynamic Testing in the Roller Bearing Test Apparatus FE8–Part 1: General Working Principles, DIN 51819-1 (Berlin, Germany: German Institute for Standardization, 2016). T. Blass, W. Trojahn, and D. Merk, “Influence of Material and Heat Treatment on the Formation of WECs on Test Rig FE8,” in Bearing Steel Technologies: 11th Volume, Advances in Steel Technologies for Rolling Bearings, ed. J. M. Beswick (West Conshohocken, PA: ASTM International, 2017), 129–150, https://doi.org/10.1520/STP160020160139 J. Franke, W. Holweger, H. Surborg, T. Blass, J. Fahl, T. Elfrath, and D. Merk, “Influence of Tribolayer on Rolling Bearing Fatigue Performed on a FE8 Test Rig,” (paper presentation, 19th International Colloquium Tribology: Industrial and Automotive Lubrication, Ostfildern, Germany, January 21–24, 2014). M. Dinkel and W. Trojahn, “Influence of Sulfur Inclusion Content on Rolling Contact Fatigue Life,” in Bearing Steel Technologies: 10th Volume, Advances in Steel Technologies for Rolling Bearings, ed. J. M. Beswick (West Conshohocken, PA: ASTM International, 2014), 1–17, https://doi.org/10.1520/STP158020140059 E. Zaretsky, R. Parker, and W. Anderson, W., “NASA Five-Ball Fatigue Tester—Over 20 Years of Research,” in Rolling Contact Fatigue Testing of Bearing Steels, ed. J. Hoo (West Conshohocken, PA: ASTM International, 1982), 5–45, https://doi.org/10.1520/ STP36131S

d.

49

50

BEARING STEEL TECHNOLOGIES: 12TH VOLUME

STP 1623, 2020 / available online at www.astm.org / doi: 10.1520/STP162320190073

Xingzhong Liang,1 Finn Sykes,1 and Pedro E. J. Rivera-Dı´az-del-Castillo1

Rolling Contact Fatigue Transformations in Aero Steels: The Effect of Temperature on Microstructural Decay Citation X. Liang, F. Sykes, and P. E. J. Rivera-Dı´az-del-Castillo, “Rolling Contact Fatigue Transformations in Aero Steels: The Effect of Temperature on Microstructural Decay,” in Bearing Steel Technologies: 12th Volume, Progress in Bearing Steel Metallurgical Testing and Quality Assurance, ed. J. M. Beswick (West Conshohocken, PA: ASTM International, 2020), 50–65. http://doi.org/ 10.1520/STP1623201900732

ABSTRACT

Bearing components undergo failure as a result of rolling contact fatigue (RCF), a prevalent issue in a range of applications including aero engines. RCF can be triggered by microstructural changes at the subsurface including the formation of white etching areas (WEAs), dark etching regions (DERs), and white etching bands (WEBs). For room temperature RCF, such microstructural alterations have been modelled with our recently proposed dislocationassisted carbon migration theory, which is able to describe the occurrence of microstructural transitions reported in the literature over the last 70 years. This approach naturally incorporates temperature and rotational speed to describe microstructural decay and failure. The model is validated with the literature data available for aerospace bearings. It is shown that temperatureaccelerated microstructural transitions and failure can be described with the proposed dislocation-assisted carbon migration theory. The shortcomings of

Manuscript received June 7, 2019; accepted for publication August 20, 2019. 1 Dept. of Engineering, Lancaster University, Lancaster LA1 4YW, United Kingdom X. L. http://orcid.org/ 0000-0003-2912-2696, P. E. J. R http://orcid.org/0000-0002-0419-8347 2 ASTM 12th International Symposium on Rolling Bearing Steel: Progress in Bearing Steel Metallurgical Testing and Quality Assurance on May 15–17, 2019 in Denver, CO, USA. C 2020 by ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959. Copyright V

d.

LIANG ET AL., DOI: 10.1520/STP162320190073

this approach are outlined, and the need for new theory and experimental data is discussed. Keywords bearing steels, microstructural alteration, rolling contact fatigue

Introduction Bearings are components used to facilitate rotational motion with large loads. In use, they are subject to very large Hertzian contact stresses, often in extreme operating conditions. Because of this, bearings are manufactured from high-strength martensitic steels but still fail due to rolling contact fatigue (RCF). This makes RCF life a major concern for the design of bearings; RCF prediction is a challenge due to the many factors affecting it, such as temperature, contact pressure, and rotational speed.1 Fatigue at the surface of bearings can essentially be avoided by the use of adequate lubrication, leaving subsurface fatigue as the critical contributor to RCF.2 Hertzian contact stresses in bearings induce damage in the material subsurface where the shear stress component culminates.3 Such damage manifests as alterations in the microstructure of the material to form white etching areas (WEAs), dark etching regions (DERs), and white etching bands (WEBs).4–6 The changes to the microstructure influence properties such as toughness, leading to failure.7 Written descriptions for the alterations follow, and there are schematic diagrams in figure 1 explaining their formations. Under a wide range of rolling conditions, WEAs are the first form of microstructural change to occur; they are typically observed around 0.001 L10.8 L10 is the

FIG. 1 Schematic representation of WEA, DER, and WEB formation; Jd is the carbon flux driven by dislocation gliding; hc stands for the thickness of carbon-enriched cell wall; rp denotes the half width of carbon-enriched precipitates; and lLC is the thickness of carbon-enriched lenticular carbides. Cell wall

WEAs

Jd

Jd

DERs

Carbides

WEBs

Carbides

Jd Lencular carbides

d.

51

52

STP 1623 On Bearing Steel Technologies: 12th Volume

number of rotations at which 10% of bearings fail. They are ferritic areas made of dislocation cells 10 to 100 nm in size. WEAs are also known as butterflies because their wing shape resembles that of a butterfly. The wing formations tend to occur roughly at 45 to the over-rolling direction, relating them to the maximum subsurface shear stress. WEAs occur adjacent to nonmetallic inclusions and primary carbides.9–12 They are formed largely due to carbon (C) migration, which occurs as a result of stress concentrations found adjacent to the rubbing of cracks during RCF.13 This causes the carbides in the area to dissolve, and extreme grain refinement subsequently occurs.9,14–16 DERs tend to occur later in bearing life, at around 0.1 L10.17 They are regions less hard than the original material’s martensitic microstructure18 that grow as the number of stress cycles increases. DERs form as a result of a type of martensite decay thought to be caused by the normal shear stress component.18,19 This results in the redistribution of carbon, which is believed to be incorporated into existing carbides.18 WEBs form at roughly 0.1L5017 in 100Cr6 steels, with L50 being the number of rotations at which 50% of bearings fail. In aerospace bearings, WEBs are present after a very high number of cycles (approximately 108).6,20 WEBs are ferrite bands where the carbon has been exhausted; they are decorated by lenticular carbides (LCs) occurring at 30 and 80 to the over-rolling direction when viewed in the circumferential section. They form because the derivative stress components cause the carbon from ferrite bands to redistribute to the boundaries.21–23 This then results in the precipitation of LCs. WEBs form in DERs or in tempered martensite where the hardness decreases by about 50 HV.21 The microstructural alterations are observed by first grinding and polishing the surface of the test material and then etching with 2% nital. Observing the material using optical microscopy, it is seen that WEAs and WEBs both exhibit a white contrast in the shapes and angles detailed previously. Nital “attacks” carbide-ferrite boundaries, so when using optical microscopy to observe DERs, a dark contrast is seen, explaining their name.24 Additionally, atom probe tomography (APT) and electron microscopy are sometimes used to examine material microstructures on a more detailed level.13,25 In the aerospace industry, bearings undergo extreme operating conditions. They must tolerate unusually high-applied contact stresses, bending moments, and rotational speeds at high temperatures. Over the last half century, a series of bearing grades have been designed for such applications; for example, bearing grade M50 has superior RCF performance at elevated temperatures up to about 300 C.8 M50 was then modified by adding nickel (Ni) and reducing the carbon content, which produced M50 NiL. Later, M50 NiL (N) was developed by applying nitriding, which increased hardness and resulted in better RCF performance. In recent years, new bearing grades have been produced, such as Pyrowear 675, AMS 6558, and ASP 2060.26 In these grades, microstructural alterations have also been observed under RCF, including WEAs, DERs, WEBs, and the newly found light etching regions (LERs).20 d.

LIANG ET AL., DOI: 10.1520/STP162320190073

Recent studies have proposed a dislocation-assisted carbon migration theory to model the microstructural alterations in bearings.18,22,27 The theory assumes that gliding dislocations accelerate the diffusion of carbon and thus the formation of WEAs, DERs, and WEBs. According to the theory, the microstructural alterations can be well described in 100Cr6 bearing steels. In this study, we attempt to extend this theory to evaluate the effect of temperature on microstructural decay and to review what is required to describe microstructural decay in aerospace bearings.

Properties and Microstructural Alterations Hard bearings possess good rolling contact properties, and it is desirable to have a higher hardness in aerospace bearings. Figure 2 shows the hardness values of some typical aerospace bearings. The hardness of these grades ranges from 700 to 1200 HV. Some grades have a hardness gradient from surface toward material center, for example, M50 NiL (N) is hardened by nitriding where an approximately 50 HV increment can be achieved. Aerospace bearings contain a variety of alloying elements, generally including carbon, molybdenum (Mo), chromium (Cr), and others. Table 1 shows chemical compositions of some aerospace bearing grades. The carbon content varies from 0.05 up to 2.35 wt.%. In some newly developed bearings, more alloying elements

FIG. 2 Hardness values of some typical aerospace bearings; error bar obtained according to the scatter from Kirsch and Trivedi20 and from Allison et al.26

ASP 2060 ASP 2055 ASP 2062 Pyrowear 675 HTT M50 M50 NiL (N) M50 NiL

HV, depth