OLYMPIC MATERIALS: Faster, Higher, Stronger

FEBRUARY 2010 JA publicationO of The Minerals, Metals & Materials Society www.tms.org/jom.html

MECHANICAL PROPERTIES Q Aluminum Q Nanocomposite Materials Q Bulk Metallic Glasses

MECHANICAL PROPERTIES JC O O N T E N T S Volume 62, Number 2 February 2010

58: The Current Activated 93: Metallic Glasses: Gaining ALUMINUM Pressure Assisted Plasticity for Microsystems: 23: Aluminum Recycling—An Densifi cation Technique for Yong Yang, Jianchao Ye, Integrated, Industrywide Producing Nanocrystalline Jian Lu, Yanfei Gao, and Approach: Subodh K. Das, Materials: J.E. Alaniz, Peter K. Liaw John A.S. Green, J.R. Morales, J. Gilbert Kaufman, and J.E. Garay FEATURES Daryoush Emadi, 13: TMS: An Essential 63: Nanostructured Metal and M. Mahfoud Connection to an Composites Reinforced Evolving Profession: 27: Aluminum Industry with Fullerenes: George T. “Rusty” Gray and Climate Change— Francisco C. Robles- Assessment and Responses: Hernández and H.A. Calderon 14: The Journal Talks with Subodh K. Das and 2009 TMS President John A.S. Green BULK METALLIC GLASSES Ray Peterson 69: Bulk Metallic Glasses: 32: Adaptive Control of Feed in 17: Advanced Materials Are a Overcoming the Challenges the Hall–Héroult Cell using a Game Changer in the Winter to Widespread Applications: Neural Network: Olympics: Peter K. Liaw, Gongyao Wang, K.D. Boadu and F.K. Omani Lynne Robinson and Judy Schneider 37: The Hot Formability of an 70: Understanding the ALSO IN THIS ISSUE Al–Cu–Mg–Fe–Ni Forging Properties and Structure insert: Member News Disk: M. Aghaie-Khafri of Metallic Glasses at the and A. Zargaran Atomic Level: T. Egami insert: The Young Leader 42: The Use of Friction Stir 76: Mechanical Response of insert: Professional Preface Welding for Manufacturing Metallic Glasses: Insights Small-scale Structures: from In-situ High Energy 2: In the Final Analysis T. Hirata and K. Higashi X-ray Diffraction: 3: News & Update Mihai Stoica, Jayanta Das, NANOCOMPOSITE Jozef Bednarčik, Gang Wang, MATERIALS 99: Meetings Calendar Gavin Vaughan, 49: Nanocomposite Materials— Wei Hua Wang, and 101: Materials Resource Leading the Way in Novel Jürgen Eckert Center: Jobs, Consultants, Materials Design: Marketplace 83: Amorphous Metals for Jonathan E. Spowart Hard-tissue Prosthesis: 104: End Notes: “Global 50: Fatigue and Fracture Marios D. Demetriou, Innovations Symposium Toughness of Epoxy Aaron Wiest, Douglas C. Flies High at TMS 2010” Nanocomposites: Hofmann, William L. Johnson, I. Srivastava and N. Koratkar Bo Han, Nikolaj Wolfson, Gongyao Wang, and Peter K. Liaw

About the Cover: The 2010 Vancouver Winter Olympics will be occurring a little less than 200 kilometers from the 2010 TMS Annual Meeting in Seattle, Washington, February 14–18. In recognition of this, the JOM feature on page 17 looks at the role that advanced materials technology will play in several Olympic events. Skis are among the most complex pieces of winter sports equipment as demonstrated by Tim Kelley, U.S. Ski Team, as he trains slalom in Saas Fee, Switzerland. (Photo courtesy of Adam Perreault/U.S. Ski Team.)

Vol. 62 No. 2 • JOM www.tms.org/jom.html 1 In the “And Now, a Word from Our Sponsor.” Final Analysis —Classic Segue from Programming to Advertising Oftentimes, we look at commercials as unwanted breaks in wanted activities. */ Me, if I’m watching television, I want to see what Jack Bauer is about to do next, not a string of commercials for automobiles and big-box stores; if I’m listening to 184 Thorn Hill Road, the radio, I want to hear Muse or U2, not a pitch for some lite beer; if I’m browsing Warrendale, PA 15086, USA the web, I don’t want to be assaulted with Flash highjacks and pop-up windows for Telephone: (724) 776-9000 mortgage companies. Fax: (724) 776-3770 Web: www.tms.org/jom.html But, if I’m at a conference, I want to see the program and the sponsoring orga- E-Mail: [email protected] nizations. And, in many cases, I want to see the exhibition most of all. A confer- ence exhibition is nothing like TV ads . . . it’s like shopping at a specialty mall that exists at one place on earth for three days only and that is built just for you. For PUBLISHER those of you reading this editorial while ensconced at the TMS 2010 Annual Meeting James J. Robinson in Seattle, that mall is here, and those three days are now. FEBRUARY ADVISORS Exhibitions are a trove of potential delights: Will someone who I want to visit Aluminum be hosting a booth? Will I see some neat stuff demonstrated? Will there be irresist- Aluminum Committee ible new widgets? Will there be food? Will there be gotta-have souvenirs? Most Pierre Homsi, Consultant, importantly, will I fi nd something that is going to make my workplace more effi cient, retired from Rio Tinto Alcan competitive, and (dare I hope?) fun? Nanocomposite Materials Composite Materials Committee I’m ancient enough to recall the fi rst TMS Annual Meeting Exhibition. It was held Jonathan E. Spowart, U.S. Air Force during 1987 in Denver and debuted in complement to festivities celebrating the cen- Research Laboratory tennial of the Hall-Héroult process. The exhibition was heavily light metals oriented Bulk Metallic Glasses Mechanical Behavior of Materials Committee back then and has tracked that way for two decades. In recent years, however, the Peter K. Liaw, University of Tennessee profi le has started to evolve, with aluminum still playing a prominent role, but with all points of the materials tetrahedron now taking on greater profi le on the exhibit fl oor. EDITORIAL STAFF Maureen Byko, Editor TMS’s director of partner relations, Trudi Dunlap, leads our effort to make sure Cheryl M. Geier, Senior Graphic Designer that there is something for everyone in the exhibition mall, err, hall. The effort includes Shirley A. Litzinger, Production Editor input from the recently formed TMS Exhibitors Council—exhibitors who offer sug- Elizabeth Rossi, Web Developer gestions on how to improve and grow the event and provide a better experience for exhibitor and attendee alike. Trudi believes this effort important for many reasons, PARTNER RELATIONS AND not the least of which is her astute observation that the exhibit hall is home to the ADVERTISING STAFF Trudi Dunlap, Director suppliers that make the technologies presented in the session rooms and in journals Colleen Leary, Sales Representative like JOM possible. Exhibitors represent the infrastructural backbone of the community. She told me, for example, that . . . JOM (ISSN 1047-4838) is published monthly by Springer Sciences & Business Media, LLC • Software vendors Thermo-Calc (booth 100), CompuTherm (booth 428), and (Springer) 233 Spring St., New York, NY 10013 in Sente (booth 414) produce and distribute tools used by computation materials cooperation with The Minerals, Metals & Materials science and engineering professionals as described in the October 2009 issue. Society (TMS). • DISCLAIMER: The opinions and statements ex- Cutting-edge testing equipment manufacturers like Hysitron (booth 400), Agi- pressed in JOM are those of the authors only and lent (booth 404), and CSM Instruments (booth 239) help scientists know how are not necessarily those of TMS or the editorial small volumes of materials deform and help everyone more easily understand staff. No confi rmations or endorsements are intended the physical behavior of materials, which are common topics in any issue of the or implied. journal. SUBSCRIPTIONS, ORDERS, AND OTHER FULFILLMENT INQUIRIES: In the Americas, • Two of the top aluminum producers in the world, Rio Tinto Alcan (booth 309) contact Journals Customer Service, P.O. Box 2485, and Hydro Aluminium Hycast (booth 608), participated in a special roundtable Secaucus, NJ 07096, USA; telephone (800) 777- on Challenges for Sustainable Growth in the Aluminum Industry, as described 4643 (in North America) or (212) 460-1500 (outside North America); e-mail [email protected]. in the November 2009 issue. Outside the Americas, contact Journals Customer Good stuff and all true. . . . Just a minute. . . . I think that Trudi shared her thoughts Service, Springer Distribution Center, Haberstr. 7, 69126 Heidelberg, Germany; telephone 49-6221- in hopes that I might mention the exhibition in an upcoming editorial. 345-4303; e-mail [email protected]. Looks like I just did. It was easy because I love the exhibition! TMS MEMBERS: Access this and back issues of JOM on-line at no charge via members.tms.org POSTMASTER: Send address changes to: JOM, Springer, 233 Spring Street, New York, NY 10013, USA. Periodicals postage paid at New York, NY, James J. Robinson and additional mailing offi ces. Publisher

2 www.tms.org/jom.html JOM • February 2010 News & Update Items of Note from the Field, Profession, and Society

12 Teams Step Up to the 2010 Materials Bowl Challenge 2010 TMS HONORS AND AWARDS RECIPIENTS In a test of materials science knowl- TMS and its divisions honor outstanding members annually for various achievements in their fi eld and on behalf of the Society. TMS congratulates the following individuals who edge and competitive resolve, 12 student have received awards and honors this year and will be recognized at the TMS 2010 Annual teams are slated to participate in the Meeting in Seattle, Washington. To learn more about each award or how to nominate a col- TMS 2010 Materials Bowl, sponsored league for an award, visit http://www.tms.org/society/tmsawards.aspx. by Alcoa, on Sunday, February 14 at Fellow—Class of 2010: The highest honor bestowed by TMS, the honorary class of Fellow was the TMS 2010 Annual Meeting. Three established in 1962. To be inducted, a candidate must be recognized as an eminent authority intense rounds of questions will focus and contributor within the broad fi eld of metallurgy, with a strong consideration of outstanding on such broad topics as Energy Materi- service to the Society. The maximum number of living Fellows cannot exceed 100. als, Materials and Society, TMS and Jeff DeHosson: For his outstanding contributions to structure-property relationships JOM, Seattle’s Best (materials ques- in materials. Says DeHosson, “I consider the Fellow Award as a major honor in the tions related to the TMS 2010 Annual international community of materials science and an important milestone in my career. Meeting city), Recycle Bin (questions TMS played a crucial role in my scientifi c life since the Society provided me a forum from previous Materials Bowls), and for the dissemination of my scientifi c accomplishments. Most importantly, the award is based on the recognition by my peers, i.e. my outstanding colleagues whose critical Potpourri. In addition to bragging rights, opinion I value most. Thanks are due to my research group, to the Zernike Institute for the winning team will receive a cash prize Advanced Materials-Groningen, and the Netherlands Materials Innovation Institute for and take home the traveling Materials their continuous support.” Bowl Trophy. The fi nal championship James W. Evans: For his pioneering and seminal work in extractive/process metallurgy round, which will take place at 8:30 and his reputation as a leader in this fi eld worldwide. Says Evans, “My years of member- p.m. on February 14, will also feature ship in TMS, since the late 70s, have taught me much that has been valuable in my career the unveiling of the winner of the TMS, and given me the opportunity to interact with my colleagues in the Society to my great MSE & Me video contest. benefi t. The Fellows of TMS are a highly distinguished group of people and I am deeply TMS members are welcomed and honored by this award.” encouraged to take a few moments to Easo P. George: For his outstanding contributions to understanding deformation and cheer on this year’s Materials Bowl fracture in intermetallic and metallic alloys and to designing new materials and functional contestants, representing the follow- use. “It is indeed a great pleasure to be elected TMS Fellow,” says George. “I joined TMS ing schools: Boise State University; as a student, shortly after which I went to my fi rst technical meeting, giving a talk and making my fi rst outside scientifi c contacts, and was hooked. The impact of TMS on my California Polytechnic State University; professional development—through meetings, publications, and friends made along the Florida International University; Georgia way—is immeasurable, and I am profoundly grateful for it.” Institute of Technology; Indian Institute Richard G. Hoagland: For outstanding contributions in fracture mechanics and atom- of Technology, Kanpur; Iowa State istic modeling of dislocation mechanisms of deformation and fracture in a broad range University; University of British Colum- of materials. Says Hoagland, “Well, this is really something! If I have done anything to bia; University of Idaho; University of deserve this recognition, it is because I have had the mighty broad shoulders of colleagues Minnesota, Twin Cities; University of and mentors to stand on, and the help of the best students and postdocs anyone could Tennessee; University of Washington; hope for. Some are listed among the TMS Fellows alumni. But many others are, or have Washington State University. been, at places like Battelle, Ohio State University, McMaster University, Washington State University, and Los Alamos National Lab. TMS meetings have been important to me because they have provided a forum where I can meet these friends again, for exchanging new results with them and others and for learning new things. And, learning is what it is all about, isn’t it?” Philip J. Mackey: 139th Annual Meeting & Exhibition For the development of new metallurgical processes for the production of non-ferrous metals, long service to the metallurgical community and, in particular, for bridging copper metallurgists of the two Americas. “TMS has always been an exciting and important organization for me,” says Mackey. “Throughout my career, it has played an important role in my professional life. Through TMS, I have met and worked with leaders in the industry. I fi rst joined TMS in 1996 and it remains as relevant today as back then. It is a great honor and privilege to be recognized as a recipient of the 2010 TMS Fellow Award.” American Institute of Mining, Metallurgical, and Petroleum Engineers (AIME) Honor- ary Member: AIME Honorary Membership is one of the highest honors that the Institute can bestow on an individual. It is awarded in appreciation of outstanding service to the Institute or in recognition of distinguished scientifi c or engineering achievements in the fi elds embracing the activities of AIME and its member societies.

Vol. 62 No. 2 • JOM www.tms.org/jom.html 3 Y. Austin Chang Special Lectures Examine AIME Robert Earll McConnell Award: This award recognizes benefi cial service to society Key Industry Issues and through signifi cant contributions which tend to enhance a nation’s standard of living or replenish its natural resources. Developments Diran Apelian Distinguished experts will challenge TMS 2010 Annual Meeting attendees to AIME James Douglas Gold Medal: This award recognizes distinguished achievement consider, discuss, and explore a broad in non-ferrous metallurgy, including both the benefi cation of ores and the alloying and utilization of non-ferrous metals. range of topics during the special lec- tures scheduled throughout the event. James C. Williams These include: William Hume-Rothery Award: The William Hume-Rothery Award was established in • Extraction & Processing/Materi- 1972 by the Institute of Metals Division of TMS and recognizes outstanding scholarly als Processing & Manufacturing contributions to the science of alloys. The award includes an invitation to the recipient to Joint Division Luncheon Lec- be the honored lecturer at the William Hume-Rothery Memorial Symposium during the TMS Annual Meeting. ture: “Titanium: Its Attributes, Characteristics, and Applications,” Didier DeFontaine: Lecture: “How Hume-Rothery’s Work Led to Computational Rod Boyer, Boeing Commercial Thermodynamics.” “When I was a metallurgical engineering student in my native Belgium at the Catholic University of Louvain, the instructor taught the physical metallurgy course Planes (Tuesday, February 16). out of a textbook by William Hume-Rothery,” says DeFontaine. “As a result, I learned This presentation will focus on the about the great Oxford physicist/metallurgist Hume-Rothery very early on and his excellent desirable and unique attributes of textbook played no small role in the orientation of my career. Hume-Rothery was a hero this alloy system, as well as provide in undergraduate school, but of course, I could not even dream that some day my name a general overview of some of the would be associated with that of the great man himself.” approaches being taken to reduce Institute of Metals Lecturer & Robert Franklin Mehl Award: The Institute of Metals the cost—some of which could Lecture, established in 1921, recognizes an outstanding scientifi c leader who is selected impact non-aerospace industries. to present a lecture at the TMS Annual Meeting. The Robert Franklin Mehl Award was • Extraction & Processing Division established in 1972. Distinguished Lecture: “Alloy Robert Ritchie: Lecture: “Nature-Inspired Structural Materials.” Says Ritchie, “I have been Formation during Electrochemical a member of TMS for well over 30 years, and can remember with clarity listening to the Cementation Reactions,” J. Brent great Mehl Award lectures of many years ago by such luminaries as Jack Christina, Mike Ashby, Peter Haasen, and Bill Nix. I am honored, yet humbled, to be the 2010 recipient Hiskey, associate dean, University of the TMS Robert Franklin Mehl Award and only hope that I am worthy enough to be of Arizona (Tuesday, February 16). included with this elite group of individuals.” Metal displacement (cementation) Application to Practice Award: The Application to Practice Award, established in 1986, reactions have been important recognizes an individual who has demonstrated outstanding achievement in transferring to many hydrometallurgical pro- research results or fi ndings into commercial production and practical use. cesses for centuries. For the most Carol Handwerker: “For decades, TMS has played a major role in nurturing the discussion part, these reactions involve rather of the materials science and engineering of electronic interconnects,” says Handwerker. “In straightforward electrochemical the transition from Sn-Pb solders to Pb-free electronics, TMS has been the forum for the steps and the deposition of unique community to disseminate its newest results, discuss open issues, and debate the tradeoffs alloys by this technique has been between different alloys with a serious focus on applications. I am very grateful to TMS reported for several systems. This for conferring on me the TMS Application to Practice Award for my work on Pb-free presentation describes these sys- solder and for continuing to provide an open, exciting scientifi c forum where materials scientists and engineers can address problems of national concern.” tems and provides an explanation for this phenomenon. John Bardeen Award: The John Bardeen Award, established in 1994, recognizes an • Light Metals Division Luncheon individual who has made outstanding contributions and is a leader in the fi eld of electronic materials. Lecture: “Aluminum—Are We There Yet?” Wayne Hale, executive Eugene Haller: Says Haller, “Winning the 2010 John Bardeen Award of TMS is a vice president and CEO, Century great honor and a much appreciated recognition of the research contributions that my collaborators, students, postdocs, and I have made. The award carries the name of one Aluminum (Wednesday, February of my greatest heroes in the sciences, John Bardeen. I was fortunate to speak with John 17). With the global recession and Bardeen on a few occasions and was always impressed by his friendly and low-key style the associated impact on com- of discussing science. As a young postdoctoral fellow at the Rad Lab in Berkeley, now the modities, industry consultants and Lawrence Berkeley National Laboratory, I discussed our ultra-pure germanium research pundits continually prognosticate and development with John Bardeen, who at the time was reviewing our work for the the future of the aluminum industry. Atomic Energy Commission, now the U.S. Department of Energy. He listened carefully with an approving smile!” This presentation offers a review of the industry’s current position, as Bruce Chalmers Award: The Bruce Chalmers Award was established in 1989 by the well as the necessary milestones Materials Processing & Manufacturing Division of TMS and is named for Bruce Chalmers, widely acknowledged as a father of modern solidifi cation science. The award recognizes and potential bumps along the way, outstanding contributions in the fi eld of solidifi cation science. in addressing the vexing question, (Continued on page 6.) “Are we there yet?” Also highlighting the Annual Meeting

4 www.tms.org/jom.html JOM • February 2010 is an array of symposia keynotes: • Institute of Metals/Robert Frank- lin Mehl Award Lecture: “Nature- Inspired Structural Materials,” Robert Ritchie, Chua Distinguished Professor of Engineering, Univer- sity of California (Tuesday, Febru- ary 16). • William Hume-Rothery Award Lecture: “How Hume-Rothery’s Work Led to Computational Ther- modynamics,” Didier DeFontaine, Professor Emeritus, University of California (Monday, February 15). • JIM International Scholar Award Lecture: “Development of Coher- ent X-Ray Diffraction Microscopy and Its Application in Materials Science,” Yukio Takahashi, Osaka University (Wednesday, February 17). • Vittorio de Nora Award Lecture: “Designing Crushing and Grinding Circuits for Improved Energy Effi - ciency,” Zeljka Pokrajcic, Worley- Parsons Services Pty Ltd—Miner- als and Metals, Australia (Wednes- day, February 17).

Ever wonder Out of the blue. From its beginning, Permatech has been innovating. where Indeed, it was founded to create an entirely new product in the casting industry—asbestos-free, pre-cast refrac- new ideas tory components. Created as it was to pioneer a new idea, the company didn’t stop there. In the nearly four decades since, Permatech has introduced one new idea come from? after another. Even now, we’re working on what’s next.

Graham, NC • permatech.net • 336.578.7728 Honors and Awards (Continued from page 4.) Cristoph Beckermann: “The Bruce Chalmers award is a great honor and recognition by Sponsors Ensure Convenience, my peers that my research has had some impact in the fi eld of solidifi cation.It is mainly Comfort at TMS Annual through TMS meetings that I learned about the many fascinating aspects of solidifi cation, Meeting from both scientifi c and industrial perspectives. TMS also allowed me to make the con- Making sure that the TMS 2010 nections that are necessary to work in this interdisciplinary fi eld. I would like to share Annual Meeting is a positive, productive this award with all of my former graduate students and collaborators.” experience down to the last detail are the Educator Award: The Educator Award, established in 1986, recognizes an individual sponsors who have underwritten an array who has made outstanding contributions to education in metallurgical engineering and/or of events and amenities. For instance, materials science and engineering. Fives Solios has once more been named John Moore: Says Moore, “I am very honored to be nominated and awarded the 2010 the meeting’s Premier Green Sponsor, Educator Award. I have been a member of TMS since 1979 and consider TMS to be actively supporting TMS in its efforts the ‘professional home’ of metallurgical and materials engineers worldwide. TMS has to make the Annual Meeting environ- provided me with the communication, networking, and professional interaction that I have needed for the past 30 years and also provides excellent support and interaction mentally sound by helping to eliminate for materials in higher education.” waste, encourage recycling, and conserve energy. A supporter of the TMS Annual Robert Lansing Hardy Award: Established in 1955, the Robert Lansing Hardy Award recognizes outstanding promise for a successful career in the broad fi eld of metallurgy Meeting for more than 20 years, Fives by a metallurgist under the age of 35. In addition, the winner receives a $500 stipend Solios will provide canvas messenger from the Ford Motor Company. bags to all full conference registrants and Diana Lados: “It is a great honor to receive this recognition and join a most distinguished is fl agging the recycling containers in group of materials scientists and engineers,” says Lados. “It is my intent to continue the Washington State Convention Center the tradition of research excellence established by this award and engage the younger with its name and logo. Headquartered generation in materials science and TMS. TMS is an excellent career development in St. Germain En Laye, France, Fives platform that facilitates valuable personal interactions. It is rewarding to be part of the Solios designs and supplies processing TMS family and contribute to the exciting future of our fi eld.” equipment and turnkey plants for alumi- Leadership Award: The Leadership Award, established in 1986, recognizes an individual num producers worldwide, particularly who has demonstrated outstanding leadership in the fi eld of metallurgy and materials. in the reduction areas of gas treatment Donald Gubser: “To be chosen as the recipient of the 2010 TMS Leadership award centers and bath processing plants. is a special honor to me. It provides recognition from a very prestigious professional Keeping Annual Meeting participants society that my science leadership has been judged by my peers to be of high quality,” connected to home and work has been says Gubser. “It is especially gratifying to know that this award basically emanates made possible by Wahl Refractory Solu- from those with whom I have worked closely, both from within my own organization, tions, Inc., which has underwritten the as well as within the larger scientifi c community through professional activities and cost of the free Wi-Fi made available in duties. To be appreciated and valued by those I have worked with for many years, and by the members of TMS, is indeed a very high and humbling honor.” the Annual Meeting Exhibit Hall. Stellar Materials is also sponsoring the Exhibit Champion H. Mathewson Award: The Champion H. Mathewson Award, established Hall’s Cyber Center—computer pods in 1933, is awarded to an author(s) of a paper or series of closely related papers which represents the most notable contribution to metallurgical science during the period under which can be accessed by all attend- review. ees and exhibitors. Helping meeting attendees stay on track and informed Guillaume Reinhart: Paper Title: “In-Situ and Real-Time Analysis of the Formation of Strains and Microstructure Defects during Solidifi cation of Al-3.5 Wt Pct Ni Alloys.” throughout the event is the Informa- Says Reinhart, “This award represents the achievement of several years of work between tion Booth sponsor, Harbison-Walker the IM2NP (formerly L2MP) laboratory in Marseille where I did my Ph.D. and the Refractories, as well as B&P Processing, European Synchrotron Radiation Facility where the experiments were performed. I’m which is sponsoring the Annual Meeting very grateful for this nomination, which also motivates my research group even more to keep on contributing to metallurgical and materials science.” Early Career Faculty Fellow Award: Established in 2006, the Early Career Faculty Fellow Award honors an assistant professor for accomplishments that have advanced the academic institution where employed, as well as recognizes his or her abilities to broaden the technological profi le of TMS. Xingbo Liu: Lecture: “Energy Materials—Past, Present, and Future.” Says Liu, “Since I became a member in 2002, TMS has played a critical role in my career development. Through attending and organizing symposia and participating in different committees, I have had the valuable opportunity to interact with scientists and engineers in the fi eld of high-temperature materials for energy conversion from all over the world. I am thrilled to be the recipient of the TMS 2010 Early Career Faculty Fellow Award and will exert more of my efforts in contributing to the TMS community.”

Shri Ram Arora Award: The Shri Ram Arora Award for Materials Science and Engineer- The annual meeting exhibit hall, with its ing Education was established in 2001 to recognize, encourage, and support the quest for displays of materials-related products and knowledge within the materials science and engineering community. processes, is a popular gathering spot.

6 www.tms.org/jom.html JOM • February 2010 and Exhibit Hall entrance, and Pamas Apu Sarkar:“It is a great honor to receive the prestigious Shri Ram Arora award. This & Company, the 2010 Annual Meeting award will help me gain international recognition, while encouraging me and boosting lanyard sponsor. my confi dence to work on new and challenging research problems,” says Sarkar. A complete list of Annual Meeting Vittorio de Nora Prize for Environmental Improvements in Metallurgical Industries: sponsors can be found on the TMS 2010 New for 2010, the de Nora Prize recognizes outstanding materials science research and Annual Meeting Web site. development contributions to the reduction of environmental impacts as applied in global metallurgical industries, with a particular focus on extractive processing. Women in Science Breakfast Zeljka Pokrajcic: Lecture: “Designing Crushing and Grinding Circuits for Improved Lecture Highlights ADVANCE Energy Effi ciency.” Says Pokrajcic, “I am both honored and excited to be the fi rst Initiative recipient of the Vittorio de Nora Prize. It is a prestigious and coveted award from an An overview of the successes and chal- equally highly regarded organization. The award will signifi cantly lift the profi le of lenges experienced by the ADVANCE environmental improvements and increased energy effi ciency in the energy-intensive minerals industry. I hope that my contribution through the research already undertaken Center for Institutional Change at the and the work planned as a result of the prize will further contribute to environmental University of Washington will be the improvements in metallurgical industries.” focus of the annual Women in Science Alexander Scott Distinguished Service Award: Since 1994, TMS has recognized a member Breakfast Lecture, to be held February 15, for outstanding contributions to the Society with the Distinguished Service Award. from 7 to 8 a.m. at the Sheraton Seattle. Eve A. Riskin, Washington’s associate Brajendra Mishra: “The membership of TMS is a privilege to me,” says Mishra. “It helped me shape my career not only as a student, but also as a practicing materials engi- dean of academic affairs, professor of neer. The leadership opportunities provided to me by TMS have been most benefi cial, electrical engineering, and center direc- along with the chance to serve the profession. Recognition of my service to TMS through tor, will discuss the program’s outcomes this Distinguished Service Award is an honor that I shall forever proudly cherish.” since its inception in 2001 and strate- Electronic, Magnetic & Photonic Materials Distinguished Service Award: This award gies it has used to enhance recruitment, recognizes an individual whose continuous service to the TMS Electronic, Magnetic & retention, and leadership development Photonic Materials Division (EMPMD) has clearly facilitated the Society’s capability of its female faculty. She will also share to serve its electronic, magnetic, and photonic materials-oriented members and their interventions that can support success of supporting organizations. female faculty in the science, technology, Elizabeth Holm: “I am honored to accept this award for service to TMS and the EMPMD. engineering, and mathematics (STEM) However, I must admit that I have received much more from my TMS membership disciplines. An initiative of the National than I have given back,” says Holm. “TMS has supported my scientifi c and professional Science Foundation, ADVANCE seeks advancement since I was a student. I hope to reciprocate by continuing to serve the to improve the climate for success of Society throughout my career.” women in STEM at universities in the Electronic, Magnetic & Photonic Materials Distinguished Scientist/Engineer Award: United States. The Women in Science This award recognizes an individual for research excellence in one or more areas related Breakfast Lecture is free to all Annual to electronic, magnetic, and photonic materials science. Meeting attendees. Brent Fultz: “TMS meetings have been a home for me since my days as a graduate student, consistently offering a stimulating environment for the exchange of new ideas. Lamborghini Revs Up Over the years, my TMS colleagues and friends were some of the best for discussing TMS 2010 Exhibition ideas about the vibrational entropy of materials, and studies of entropy by inelastic scattering methods. These discussions helped me optimize the research that is recog- The TMS 2010 Exhibition will be the nized by this award.” epicenter of networking activity when Extraction & Processing Distinguished Service Award: This award recognizes an it opens for business at the Washington individual who has demonstrated outstanding long-term service to industries served by State Convention Center at noon on the Extraction & Processing Division. Monday, February 15. New this year is the TMS Lunch and Learn Session, Patrick Taylor: “It is an honor to receive the 2010 TMS EPD Distinguished Service Award. I attended my fi rst TMS meeting in 1975 and gave my fi rst professional pre- “Carbon Fiber Composites Research sentation. Over these past 34 years I have truly benefi ted, professionally and personally, and Development at Automobili Lam- from my participation in TMS EPD. The many hours that I have devoted to help TMS borghini,” scheduled to begin at 12:30 and EPD has been paid back, many times, through the positive impact of this society p.m. in the exhibit area on February 15. In on my professional career and the friends that I have made. addition to hearing fi rst-hand how Lam- Extraction & Processing Distinguished Lecture Award: This award recognizes an borghini, the University of Washington, eminent individual in the fi eld of extraction and processing of nonferrous metals with an and Boeing have partnered to advance invitation to present a comprehensive lecture at the TMS Annual Meeting. excellence in carbon fi ber composite J. Brent Hiskey: Lecture: “Alloy Formation during Electrochemical Cementation technology, program attendees will get Reactions.” a chance to see a Lamborghini Murcié- Extraction & Processing Science Award: This award, established in 1955, recognizes a lago LP 670-4 SuperVeloce up close. paper that represents notable contributions to the scientifi c understanding of the extraction (The Murciélago will be on display in and processing of nonferrous metals. the exhibit hall for the duration of the Stanko Nikolic, Peter C. Hayes, Hector Henao, and Eugene Jak: Paper Title: “Advan- exhibition.) tages of Continuous Copper Fire Refi ning in a Packed Bed.” TMS 2010 Annual Meeting attendees

Vol. 62 No. 2 • JOM www.tms.org/jom.html 7 Extraction & Processing Technology Award: This award recognizes a paper contain- will also have ample opportunity to learn ing notable contributions to the advancement of technology related to the extraction and about the latest technologies, products, processing of nonferrous metals. and services from approximately 100 Gabriel Riveros, Andrze Warczok, Daniel Smith, and Ariel Balocchi: Paper Title: exhibitors covering the spectrum of “Phase Equilibria in Ferrous Calcium Silicates, Parts IÐIV.” materials science and engineering. These include long-time exhibition support- Light Metals Distinguished Service Award: This award, established in 1995, recognizes an individual whose continuous service to the TMS Light Metals Division has clearly ers Rio Tinto Alcan, Outotec, Dubai facilitated the Society’s capability to serve its light metals-oriented members and their Aluminium, Hydro Aluminum Hycast, supporting organizations. Fives Solios, and Gillespie+Powers, Inc. In addition, TMS is welcoming a number Neale Neelameggham: “It is an honor and pleasure to receive the Distinguished Service of new exhibitors to this year’s event, Award from the Light Metals Division of TMS. I have been a member of TMS for over 35 years and actively engaged in the different committees during the past ten years. I such as Big C, Agilent Technologies, appreciate the confi dence my co-volunteers and peers from the Magnesium Commit- CIMM Group, Farra Engineering Ltd., Magnesium Elektron, Major, Momen- tum Press, Oli Systems, and SunStone. NIST Technology Innovation Program Selects New Projects The U.S. National Institute of Stan- dards and Technology (NIST) announced on December 15 that it will provide up to $71 million through its Technol- ogy Innovation Program (TIP) for 20 new projects that support innovative, high-risk research in new technologies. The new projects include developing unmanned, hovering aircraft for bridge inspections, a high-speed sorting system for recycling aerospace metals, and nano- The door to the future of material testing materials for advanced batteries. The has been opened with the AG-X. From awards will be matched by other funding concept to reality, count on the AG-X— sources to achieve nearly $150 million the innovative, user-friendly solution— in new research over the next two to fi ve for your testing requirements. years. The projects were selected from World-class Specifications a TIP competition announced on March ■ 0.5% load cell measurement precision 26, 2009. A complete list of these initia- (1 to 1000) tives can be accessed at http://www.nist ■ 5kHZ real sampling rate .gov/public_affairs/releases/20091215 ■ 1M pulse/second resolution _tip_awards.htm. Powerful, Multifunctional Trapezium-X TIP is a merit-based, competitive pro- Software gram that provides cost-shared funding ■ “My Memory-My Tester”—Insert a for research projects by single small- or USB card to build your specialized medium-sized businesses or by joint machine ventures that may include universities, ■ Innovative Web-based re-analysis function non-profi t research organizations, and national laboratories. The intent of the program is to support and accelerate innovation in the United States through high-risk, high-reward research in areas You demand performance and of critical national need. productivity—Shimadzu delivers. TIP also announced that it is currently seeking comments on white papers pre- pared by its staff and posted on the TIP Web site. These white papers articulate Learn more about Shimadzu’s universal testing machines. the goals and scope of technical chal- www.ssi.shimadzu.com/AGX Call (800) 477-1227 or visit us online at lenges that could be addressed in future TIP competitions. The white papers can be accessed at http://www.nist.gov/tip Shimadzu Scientifi c Instruments Inc., 7102 Riverwood Dr., Columbia, MD 21046, U.S.A. Order consumables and accessories on-line at http://store.shimadzu.com /wp_cmts/index.html. One Source One Partner for Leading Brand Solutions & Expertise in Bauxite Mining & Refining, Calcination, and Smelting

High Performance for Alumina Handling Systems Every day, worldwide, our equipment crushes, conveys, grinds, digests, clarifies, precipitates, stores, and calcinates bauxite to produce alumina. Combining the respected brand names of MÖLLER, KOCH-MVT, FULLER-TRAYLOR, EIMCO, DORR-OLIVER, KREBS, and ABON, FLSmidth offers a broad range of equipment and processes while increasing recoveries, lowering energy consumption, and providing proven reliability. www.FLSmidthMinerals.com Please visit us at TMS 2010 booth 227. JOM tee, the Energy Committee, LMD Council, and other committees have had with my contribution in conferring this award. My humble thanks to them.” READER POLL Light Metals Technology Award: This award recognizes an individual who has dem- To vote, go to onstrated outstanding long-term service to the light metals industry by consistently www.tms.org/jomsurvey.html. providing technical and/or operating knowledge that has enhanced the competitiveness of the industry. Last month’s question: Douglas Granger: “To be chosen by ones peers to receive the prestigious Light Metals Technology Award is a great honor that I truly appreciate,” says Granger. “At the same time, it is a humbling experience to be joining a roster that includes so many distinguished In the 2009 salary survey by the scientists and technologists. Throughout my more than 30-year career in light metals American Association of Engi- research and development, TMS has been of immense importance to me. Much of what I have accomplished would not have occurred without the stimulating discussions and neering Societies, the median pay thoughtful guidance provided by my TMS colleagues.” for materials engineers with 35+ Light Metals Award: This award, established in 1983, is conferred upon the author(s) of years since earning their baccalau- a paper, presented in the preceding year in a Light Metals Division sponsored session at reate degrees was $119,224. This the Annual Meeting, which notably exemplifi es the solution of a practical problem. month, JOM asks, how would you Anthimos Xenidis, Charalabos Zografi dis, Ioannis Kotsis, and Dmitrios Boufounos: characterize materials engineers’ Paper Title: "Reductive Roasting and Magnetic Separation of Greek Bauxite Residue for its Utilization in the Iron Ore Industry.” Says Xenidis, “My collaborators and I feel salaries? deeply honored to be the recipients of the 2010 Light Metals Best Paper Award of TMS. This is recognition of the quality of the research work we carry out at the National Last month’s Technical University of Athens. Rightly so, it fi lls us with pride. However, this award is not just a valued distinction, but also a high level standard for which we will have to answers: strive in our future endeavors.” Inadequate 59.46% Light Metals Division JOM Best Paper Award: This award, established in 1994, is Fair 24.32% presented to the author(s) of a paper published in an issue of the preceding volume year of JOM under a light-metals-related topic. Generous 16.22% Barry Welch, Martin Iffert, and Maria Skyllas-Kazacos: Paper Title: “Applying Fun- This month’s question: damental Data to Reduce the Carbon Dioxide Footprint of Aluminum Smelters.” Materials Processing & Manufacturing Division Distinguished Scientist/Engineer Award: This award recognizes an individual who has made a long-lasting contribution This issue contains a cover story to the design, syntheses, processing, and performance of engineering materials with signifi cant industrial applications. about the use of advanced materials Yuntian Zhu: “It is a great honor to be recognized with this award by colleagues,” in the Olympic games (which will says Zhu. “For my entire academic career, TMS has provided me with essential and be held in Vancouver, just hours great opportunities for academic growth, service, and networking with colleagues. The away from the TMS 2010 Annual importance of membership in TMS cannot be overstated for professionals in materials Meeting in Seattle, Washington), science and engineering.” and notes the controversy that can Materials Processing & Manufacturing Division Distinguished Service Award: This award recognizes an individual whose dedication and commitment to the Materials Pro- result when athletes are too reliant cessing & Manufacturing Division has made a demonstrable difference to the objectives on technology to improve their and capabilities of the division and the Society. performances. This month, JOM John Smugeresky asks, what effect has materials sci- Structural Materials Distinguished Materials Scientist/Engineer Award: This award, ence had on the competitive spirit established in 1996, recognizes an individual who has made a long-lasting contribution to of the Olympics? the fundamental understanding of microstructure, properties, and performance of structural materials for industrial applications. Reinhold Dauskardt: “I am delighted that I am receiving this award and recognition This month’s from TMS,” says Dauskardt. “I have been a long-standing member of TMS and have answers: been very impressed at how the Society has embraced new and non-traditional areas of Little impact: Superior athletes will materials and biomaterials research. This has kept the society scientifi cally and techno- logically relevant and the reason that I enjoy attending the annual meetings.” prevail regardless of the equipment Structural Materials Division Distinguished Service Award: This award recognizes they used. an individual whose dedication and commitment to the Structural Materials Division has made a demonstrable difference to the objectives and capabilities of the division and the Moderate impact: Technology can Society. provide the push needed for a good Dallis Hardwick athlete to become a winner. Structural Materials Division JOM Best Paper Award: This award, established in 2007, Signifi cant impact: Advanced mate- is presented to the author(s) of a paper published in an issue of the preceding volume year of JOM under a structural-materials-related topic. rials have given an unfair edge to Marc A. Meyers, Sirirat Traiviratana, V.A. Lubarda, David J. Benson, and Eduardo the teams that can afford the best M. Bringa: Paper Title: “The Role of Dislocations in the Growth of Nanosized Voids technology. in Ductile Failure of Metals.”

10 www.tms.org/jom.html JOM • February 2010 Science and Engineering Indicators WASHINGTON NEWS 2010: Declaring that the state of the science and engineering enterprise in the United States is strong but slipping, the National Science Board released the biennial Science and Engineering Indicators report at a From Betsy Houston, briefi ng at the White House on Janu- Federation of Materials Societies ary 15. The report, mandated by law, provides information on the scope, quality and vitality of America’s sci- ence and engineering enterprise. This year, it stressed the nation’s position in the global economy. Over the past decade, R&D intensity—how much of a country’s economic activity or gross domestic product is expended on R&D—has grown considerably in Asia, while remaining steady in the United States. Annual growth of U.S. R&D expenditures has averaged fi ve to six percent, while Asian countries have skyrocketed to double, triple, or even four times that. Other key indicators include intellectual research outputs; patents; the globalization of capability; funding; performance and portfolio of U.S. R&D trends; and the composition of the U.S. science and engineering workforce. The report is prepared by the National Science Foundation’s (NSF) Division of Science Resources Statistics (SRS) on behalf of the National Science Board (NSB). Both the extensive fi nal report and a more easily usable Digest are online at www.nsf.gov/news. Science and Technology Agenda: Rep. Bart Gordon’s tenure as Chairman of the House Science and Technology Committee has been marked by an aggressive schedule of hear- ings and successes in getting bills passed in the House. The Committee won’t slow down this year, even though Rep. Gordon (D-TN) has announced his retirement from Congress at the end of the term. The signature piece will be reauthorization of the America COM- Please visit us at TMS 2010 booth 428. PETES Act which expires at the end of FY 2010, keeping the National Science Founda- tion (NSF), the National Institute of Standards and Technology (NIST), and the Offi ce of Science within the Department of Energy (DOE) on the path to doubling their research and development budgets. Chairman Gordon’s plan is to get a multi-year COMPETES reauthorization drafted and through the House by Memorial Day. In addition to its current provisions, the bill is also likely to include new programs and policy direction at NSF, NIST, and other agencies. Also by the end of May, the committee expects to draft and report out a multi-year reauthorization of NASA. In energy, the committee intends to draft legislation to authorize a broad research and development program on nuclear energy at the DOE. The new program is expected to address enrichment, reprocessing, generation, and the storage of spent fuel. The committee also will spur development of new energy Instruments & technologies through targeted, stand-alone bills. Testing Services Energy Information Hub Funding: Three Energy Innovation Hubs, to be funded at up to $122 million over fi ve years, will be established in FY 2010. The three hubs will Thermal Analysis focus on: · High Temperature & Classical  Production of fuels directly from sunlight DSC, DTA, TGA, DSC-TGA  Improving energy-effi cient building systems design · High Accuracy Specific Heat  Computer modeling and simulation for the development of advanced nuclear reac- · Coupling to FTIR & MS tors Conductivity & Universities, national laboratories, nonprofi t organizations, and private fi rms are eligible to compete for an award and are encouraged to form partnerships. Further information is Flash Diffusivity available at www.energy.gov/hubs. · –125 to +2000°C · New Guarded Hot Plate Public Access to Research Results: The Scholarly Publishing Roundtable convened last summer by the House Science and Technology Committee, in collaboration with the Dilatometers White House Offi ce of Science and Technology Policy. In mid-January, it released its · CTE & Sintering to +2800°C report, calling on federal agencies that fund research to develop and implement policies Contract Testing ensuring free public access to research results “as soon as possible after those results have been published in a peer-reviewed journal.” The Roundtable identifi ed a set of prin- Services ciples viewed as essential to a robust scholarly publishing system, including the need to NETZSCH Instruments preserve peer review, the necessity of adaptable publishing business models, the benefi ts 37 North Ave. of broader public access, the importance of archiving, and the interoperability of online Burlington, MA 01803 content. The group also affi rmed the high value of the “version of record” for published Call (781) 272-5353 articles and of all stakeholders’ contributions to sustaining the best possible system of Fax (781) 272-5225 scholarly publishing. The report can be found at www.aau.edu. e-mail: [email protected] Thermo-Calc Software Thermodynamic and Diffusion Simulation Software Thermo-Calc Thermo-Calc: powerful software for thermodynamic calculations for multi-component multi-phase systems. Thermo-Calc can be used for a wide range of applications including: 9 Calculating stable and meta-stable phase equilibria 9 Predicting amounts of phases (mass, volume and mole fractions) and their compositions 9 Liquidus and solidus temperatures, Scheil solidification 9 Thermochemical data such as enthalpies, heat capacity, and activities 9 Predicting driving forces and transformation temperatures 9 Phase diagrams (isothermal, isoplethal sections for multi-component multi-phase systems) 9 Databases for Fe-, Ni-, Al-, Ti-, Mg- based alloys, solders, refractory oxides, slags and more… 9 Alloy design: Modifying chemistry to improve properties or reduce costs 9 Amount and composition of phases varying with temperature or composition of the alloy 9 Quality control: How phase balance and certain properties vary within in-tolerance compositional variation DICTRA A unique engineering tool for simulating various types of diffusion-controlled phase transformations in multi-component systems. DICTRA provides valuable information on the time evolution of the microstructure. Applications include: 9 Homogenization 9 Diffusion controlled phase transformation kinetics 9 Carburizing & Decarburizing 9 Nitriding and carbonitriding 9 Microsegregation during solidification 9 Coarsening / Dissolution of precipitates 9 Interdiffusion 9 Databases for Fe-, Al- and Ni- based alloys Thermo-Calc Programming Interfaces Thermo-Calc Programming Interfaces are available for users who wish to integrate many of the Thermo-Calc functions and utilize the thermodynamic databases directly from their own in-house written software.

TQ-Interface: For programming in Fortran

TC-API: For programming in C++, Java, Microsoft Visual Basic and Excel etc.

TC-Matlab toolbox: For accessing Thermo-Calc from MatlabTM. Calculated using MICRESS(®) in conjunction with the TQ-Interface. TCC (Ver. S) / TCW (Ver. 5) / DICTRA 25 now available USA, Canada and Mexico: Rest of the World: Please visit us at Booth 315 Email: [email protected] Email: [email protected] PleaseMS&T visit us at Phone: (724) 731 0074 Phone: +46-8-545 959 30 TMS Pittsburgh,2010 booth PA 100. Fax: (724) 731 0078 Fax: +46-8-673 37 18 Oct 27-28, 2009 www.thermocalc.com www.thermocalc.com Feature Presidential Perspective TMS: An Essential Connection to an Evolving Profession

George T. “Rusty” Gray III

It is with excitement and a great sense tanium, and Mechanical-Behavior com- liance of the Americas initiative aimed of honor that I accept the Presidency mittees and served two terms on the at working with society partners from of The Minerals, TMS Board of Directors—fi rst as chair South America and Canada. To the Metals & Mate- of the Structural Materials Division and north, initiatives are underway with the rials Society for then as the Director of Publications. Metallurgical Society of the Canadian 2010–2011. Let As is undoubtedly the case with Institute of Mining, Metallurgy and me introduce my- many of you, TMS has been—and Petroleum, especially with our partici- self. My name is is—my central networking tie into the pation on Lead-Zinc 2010 during their George Thompson evolving world of materials and a wel- Conference of Metallurgists. Gray III, but no coming venue to meet twice a year with Next, leadership development is key one calls me that. research colleagues and friends. because leaders within strong, enduring To family, friends, and the materials My goals as TMS president are to organizations don’t happen by chance; community, I’m “Rusty;” a nickname work with you, the volunteers, along they are sought, most often through in- given to me by my grandfather. with TMS staff, to maximize the value formal and personal contacts and there- My education includes a B.S. and of TMS to its membership by strongly after developed and nurtured on an on- M.S. from the South Dakota School of supporting our core strengths. These in- going basis. While staff can help coach Mines, a Ph.D. from Carnegie-Mellon clude maintaining TMS as: the process, the volunteers are best po- University, and a three year postpoc- • The preferred source of and dis- sitioned to infl uence members’ engage- toral fellowship at the Technical Uni- semination venue for leading-edge ment in committees, boards, etc. versity in Hamburg-Harburg, Germany. technical information and knowl- Finally, the focus on our “value” Today, I’m a Laboratory Fellow at Los edge for its members proposition: TMS is committed to Alamos National Laboratory, where I • The home society for the manufac- demonstrating its ability to remain rel- have pursued both fundamental and ap- turing, engineering, research, and evant to meet the ever-changing needs plied research for the last 25 years. materials education communities of our diverse membership through When not at work, I enjoy a number and cultures, bridging science and innovations in our technical meetings of activities outside of materials science engineering technologies critical to and exhibits, publications, information including bee-keeping, weaving bas- industry, research, and academia technologies and content, and member kets, and designing and building stained • The society dedicated to excel- services while continuing to focus on glass lamps. I also have a passion for lence and growth in supporting the fostering an inclusive environment for volunteering to help those less fortu- evolving fi eld of materials science the very diverse community that is ma- nate. For each of the last eight years, for and engineering through education, terials science and engineering. example, I’ve spent a week in Mexico and the application of materials to As TMS continues to develop and as site-boss for 50–80 high school stu- benefi t societal needs grow new initiatives in many of these dents there to help build houses. As president I hope to positively im- areas, it is you, the membership, who I have been supported in all of these pact TMS’s value to members in three I challenge to both support our society activities, in more ways than I can specifi c areas: international liasions, and our materials community. Seize any name, by my wife Altana and our two leadership development, and the soci- opportunity to mentor a new engineer children, Alex and Krista. ety’s value proposition. or scientist, promote your fi eld through As a professional, I have been an ac- A focus on international liaisons is organizing a symposium, volunteer in tive member of TMS since 1986. As critical because TMS must remain agile your local schools promoting science a young professional, TMS served as in the rapidly changing materials fi eld; and math, or join a TMS committee. my primary conduit of materials sci- one avenue is through international We’ll all be richer for the effort. ence knowledge and the principle place teaming initiatives. This year’s jointly for me to present and receive technical organized conference with the Brazilian George T. Gray III is Laboratory Fellow in the Dynamic Materials Properties Section at Los feedback on my research. Within TMS, Metallurgical, Materials, and Mining Alamos National Laboratory, Los Alamos, New I have served on the Programming, Ti- Society (ABM) builds on the TMS Al- Mexico and the 2010 TMS president.

Vol. 62 No. 2 • JOM www.tms.org/jom.html 13 Feature Presidential Interview

The Journal Talks with 2009 TMS President Ray Peterson

TMS President Ray D. Peterson will and replaced that meeting with phone be there to foster your development and complete his term as president of the and web conferences. Some resources interchange. We’re going to grow with society at the TMS 2010 Annual Meet- were reallocated and sadly, we even our members. ing this month. Pe- had to freeze staff salaries to keep our As to the long-term future of TMS, terson is director economics in order. we’re a multifaceted, very broad orga- of technology for More importantly for the long term, nization, and we’re going to be adding Aleris Internation- though, because we can’t save ourselves new areas of concentration around the al, a global leader into prosperity we put more effort this periphery as the fi eld of materials science in the production year into working on the income side expands. We’re going to become more and sale of alumi- for the future. In particular, we looked inclusive, covering a larger fi eld. Growing num rolled and ex- at what we can do to grow our income, in these new directions should make us truded products, and concluded that we need to make stronger without taking anything away recycled aluminum, and specifi cation some long-term changes to encourage from our core strengths. alloys. Peterson is Past Chair of the growth, particularly in new areas, which Q. One of your broad goals, stated a TMS Light Metals Division, and has we will talk about more a little later on. I year ago, was to encourage greater also chaired the Aluminum Committee see these changes as being very positive participation of members in TMS and and the Recycling Committee. He has for TMS in the coming years. its many benefi ts. What improvements been a member of TMS since 1985. As Q. The board of directors decided in have you seen in member involve- Peterson neared the end of his term, late 2009 to build on TMS’s technical ment? JOM interviewed him to reminisce strengths with its initiative to “Grow about the challenges and accomplish- A. First I want to say I thought TMS staff TMS” in the areas of computational ments of his year as president. really took this challenge to heart, and I materials science and engineering, was very impressed by that. Our members Q. 2009 was a year of worldwide micro- and nanostructure process- have access to a great deal of content fi nancial instability. How has TMS ing and property relationships, and on-line, such as the AIME Transactions emerged from the downturn, and what materials and energy. How do you see archive, electronic subscriptions to Met- strategies were employed to keep the this plan changing the future profi le allurgical and Materials Transactions organization on track fi nancially? of TMS? A and B, new resources for consultants, A. 2009 was a diffi cult year, and TMS, A. The materials fi eld is not static; new an e-mentoring program, the Job and like so many businesses worldwide, areas are constantly developing and Financial Resources Center, the TMS felt the effects of the downturn. For evolving. For an organization it’s some- e-library—so many options for members example, participation at the annual times easy to maintain the status quo are now on-line that in the past weren’t. meeting dropped off when employers but TMS can’t—and won’t—do that. Not everybody has access to a university imposed travel restrictions to control We have to nurture our core areas so library, so these expanded on-line offer- costs and some of our usual attendees they remain strong but we have to grow ings are benefi cial, especially to members had to cancel due to the economic crisis. new areas, as well. We’re not going to employed outside of academia. That impacted our revenue side for the abandon our traditional areas, but we And to help members feel more per- meeting, which is signifi cant because the have members who may be moving from sonally connected, in JOM and on-line annual meeting is one of our big income the traditional areas TMS serves into we’ve begun to provide a monthly focus sources. But Executive Director Warren newer disciplines such as micro- and on our members, looking at them outside Hunt and his staff really worked hard nanostructuring, computational materials of their profession. We’re learning about on reducing costs through the rest of science and engineering, and materials their hobbies and their interests, and how year through careful management. As and energy. These are all growing areas, that makes them who they are. the recession dragged on, we cancelled and we want to be part of that growth. So A little bit peripheral to this, we’ve several small conferences and eliminated I guess you might say that as your inter- been working on improving the exhibit one in-person board of directors meeting est in careers changes, TMS is going to at the annual meeting. We’ve taken into

14 www.tms.org/jom.html JOM • February 2010 consideration the comments and sugges- whole process for our awards system. ago, and granted, may not be very excit- tions of both exhibitors and members to Unfortunately, it will be a year and a ing but it is very important we had our foster the best opportunities for exchange half before anyone sees the output of this bylaws in place. We continued to work on the exhibit fl oor. At the same time, we program update, because the nominations on membership rules, and how to keep understand that not everyone can attend process for 2011 is just now fi nishing members involved, even holding our the annual meeting, so we continue look- up. That’s another area of improve- membership dues constant for 2010. ing for ways to provide members more ment—where the awards process used The continuation of the Alliance of value with on-line products and other to take 18 months from nomination to the Americas was fruitful in 2009. We services. presentation, the cycle will be shortened met with the Metallurgical Society of the In the end we have to look at ourselves to 10 months. We’re really compressing Canadian Institute of Mining, Metallurgy as being just like any other seller of ser- the schedule, but that won’t happen until and Petroleum (MetSoc) at their annual vices or goods. If we don’t provide the after the 2011 meeting. meeting in Canada, and have been very services that somebody wants they’ll go With these improvements, I believe active with the Brazilian Metallurgi- elsewhere. It’s a competitive world and we’ve made it easier to recognize our best cal, Materials and Mining Association we have to realize our members are also and brightest members. And that’s really (ABM), with a joint conference sched- customers. one of the key functions of a professional uled for 2010. We hope this will be the fi rst society. step in continued collaborations among Q. Another goal of yours was to take the Alliance of Americas societies. a close look at the TMS honors and Q. The TMS Foundation announced We also worked at intersociety coop- awards program and identify areas for a new initiative to support materials eration in the area of extractive process- improvement. In December the board science and engineering service outside ing, coordinating meeting scheduling to agreed to adjust the awards program the classroom or lab, through a col- reduce confl icts between the Society for to better fi t the current TMS member- laboration with Engineers Without Mining, Metallurgy, and Exploration ship, as discussed in your December Borders USA. What about the Engi- (SME) and MetSoc. 2009 JOM article. How will TMS neers Without Borders program made In the areas of advocacy, undergradu- members benefi t from the changes to it a good fi t for TMS support? ate education, and KÐ12 education we this program? A. In 2009 our Materials and Society continued to work at better coordination A. That was an area I took a personal Committee was authorized and became with all the different materials societies. interest in and spent a lot of time and effort active, and their emphasis was how we, We explored new technologies like You- on because, when we reviewed the honors as engineers and scientists, can make a Tube and Facebook to keep our members and awards program, we could see that positive impact on society, both locally connected. Again we’re not giving up on our system seemed to be out of date with and globally. This Engineers Without our old ways of doing things but trying the current fi eld. As a result, many of our Borders project was a direct outgrowth to fi nd new ways to meet the needs of accomplished members were not being of that committee. The project was ideal our society, and, where possible, save recognized—we have a lot of talented because we involved our Material Advan- money as well. people but they may not be eligible for tage chapters. We’ve found that students Q. What was the greatest challenge of recognitions under the existing system. at this level are enthusiastic supporters of serving as TMS president? After some study we concluded that there the kinds of concepts Engineers Without were several reasons for this. First, it was Borders emphasizes, so it was a natural A. A big challenge was understanding rather hard to nominate people, and per- fi t between extending our outreach and the breadth of activities that go on within haps some people found the process too involving our student chapters. This is a TMS. Even though I was active with the intimidating and time consuming. Also, one-year trial, but we’re very hopeful that board of directors for fi ve years prior to not all members had awards to aspire to. it will be something we can continue to coming into the presidency, I’ve learned If you were in one the of the newer areas support through the Foundation. many new things about TMS activities of materials science, our old awards may since I became president, which is just Q. What are the key achievements of not have matched up to what you were amazing to me. the TMS Board of Directors during doing. A key fi nding was that many of our Also a challenge, frankly, was the your term? awards were really geared to our senior demand on my time, I knew this posi- members, which is nice, but we have other A. In no particular order, our accomplish- tion would require a lot of time, and the people who should be receiving awards ments include the honors and awards job is fun and exciting, but I still have a who are not at the end of their career. update; keeping the society fi nancially regular job that demands my presence. This led us to develop new categories of strong even through the economic As president, you just have to fi gure awards: a mid-career award and awards downturn; and planning for the future out how you’re going to fi t these other for processes, structure, and properties, growth of TMS. things into everyday life. I want to thank which would be open to anybody within We created and approved the Women my employer, Aleris, for permitting me the materials science fi eld. in Materials Science and Engineering to do this and giving me a little bit of a We also worked at strengthening the Committee, and completed the update break. It’s a very demanding job and it’s divisional awards for all technical divi- of our bylaws and administration policy. not all fun and glory, but very satisfying sions, and really tried to streamline the That process started almost two years in the end.

Vol. 62 No. 2 • JOM www.tms.org/jom.html 15 Your Professional Partner for Career Advancement

Introducing 4 more reasons to renew your TMS membership in 2010!

1. E-mentoring – Designed to assist colleagues at every career level with professional development.

2. Consultant Resources – A compendium of tools, templates, and tips to ensure your business infrastructure is sound.

3. TMS On Reserve– An anthology of 12 newly-digitalized, library archive volumes of foundational documents and select conference proceedings from the American Institute of Mining, Metallurgy, and Petroleum Engineers (AIME).

4. OneMine – A collaborative effort among societies to create a single global resource for mining and minerals research. TMS members receive a special subscription rate.

Enjoy these and the numerous other TMS benefi ts you’ve come to value by accessing our sleek, redesigned Members Only Web pages.

Discover how all of your TMS professional essentials are just a click away!

http://members.tms.org

LEARN • NETWORK • ADVANCE www.tms.org Member News Updates on friends and colleagues in the materials community

TMS e-Mentor Program TMS Welcomes New Members Continues to Grow At its October meeting, the TMS Buenos Aires, Argentina. Board of Directors approved a slate • Peter D. Moran: Michigan Tech- of 34 new members representing 11 nological University, Houghton, countries and an array of materials Michigan. science and engineering disciplines. • Renee Pedrazzani: University of Please join us in congratulating these Rochester, New York. individuals as they begin to explore the • Mihriban O. Pekguleryuz: McGill many personal and professional growth University, Montreal, Canada. opportunities offered to them as full • Charles A. Pitzer: Pitzer Consult- members of TMS: ing, El Segundo, California. Ben Poquette Nitin Chopra • Thomas L. Aldrich: ASARCO • Siba P. Ray: Alcoa, New Kensing- Two additional TMS members are LLC, Tucson. Arizona. ton, Pennsylvania. offering their insights on career devel- • Iyad Talal Alzaharnah: KFUPM, • Rodney G. Riek: Harley Davidson opment strategies as TMS e-Mentors. Dhahran, Saudi Arabia. Motor Co., Wauwatosa, Wiscon- Providing an industry perspective is • Lynn H. Baldwin: Fort Worth, Tex- sin. Ben Poquette, president of Keystone as. • Ralph R. Sawtell: Alcoa, Cleve- Materials LLC. Nitin Chopra, assis- • Richard S. Bradley Jr.: CQE Sys- land, Ohio. tant professor at the University of Ala- tems, Inc., Wheaton, Illinois. • Tania M. Slawecki: The Pennsyl- bama, can advise on career questions • Renee Downie: The Equity En- vania State University, University related to the academic sector. They gineering Group, Inc., Houston, Park. join the inaugural TMS e-Mentor team Texas. • Oleksandr Vasylyev: Institute for of Thomas Battle, senior metallurgist, • Michael Driver: McKeesport, Problems of Materials Science, Midrex Technologies and current chair Pennsylvania. Kyiv, Ukraine. of the TMS Extraction & Processing • Oladapo O. Eso: ATI Engineered • Hakon Viumdal: Tel-Tek, Porsg- Division Council; Chandler Becker, Products, Huntsville, Alabama. runn, Norway. materials research engineer at the U.S. • Janardhan Gangojirao: Shimoga • Hong Nguyen Vu: Institute of National Institute of Standards and Karnataka, India. Chemical Technology, Prague 6, Technology (NIST); Jeffrey Fergus, • Jicheng Gong: University of Ox- Czech Republic. professor at Auburn University; and ford, United Kingdom. • Mingxu Xia: BCAST, London, Robert D. Shull, group leader of the • Isaias Almaguer Guzman: Servi- United Kingdom. Magnetic Materials Group at NIST cios Indust Penoles SA CV, Tor- • Hideyuki Yasuda: Osaka Univer- and 2007 TMS President. reón, Mexico. sity, Japan. Launched in November 2009, the • Eugene E. Haller: University of • Jiaming Zhang: Ann Arbor, Michi- TMS e-Mentor program is a new California, Berkeley. gan. member benefi t designed to meet the • Vincent H. Hammond: Army Re- career development needs of students search Labs. Jeffrey Wadsworth and young professionals, while also • Xiaohua Hu: McMaster University, Receives 2009 National providing an additional networking Ontario, Canada. Materials Advancement opportunity to those more established • Mitsuo Ishii: Nippon Steel Corpo- Award in their careers. The latest addition ration, Tokyo, Japan. Jeffrey Wadsworth, Battelle Memo- • to the suite of resources available Jagannadham Kasichainula: North rial Institute president and chief execu- through the TMS Job and Financial Carolina State University, Raleigh. tive offi cer and TMS Fellow, received • Security Resource Center, the TMS James E. Krzanowski: University the 2009 National Materials Advance- e-Mentor program offers the opportu- of New Hampshire, Durham, New ment Award from the Federation of nity to conveniently and confi dentially Hampshire. Materials Societies at a reception on • pose questions to an experienced TMS Fiona Levey: Northborough, Mas- December 9. The award was conferred member on day-to-day or long-term sachusetts. on Wadsworth for the major role that • career issues. Xinping Liang: GCL Holding he has played in national science and To begin using the TMS e-Mentor- Company Ltd. technology policy, with a particular • ing Program, log on to the Members Freddy O. Martinez: KB Alloys emphasis on the role and contributions Only home page and select “e-Mentor LLC, Robards, Kentucky. of materials science and engineering. Program.” • Nelida Mingolo: Universidad de TMS Member Profi les

Meet a Member: Xingbo Liu: TMS’ “Energetic” 2010 Early Career Faculty Fellow By Lynne Robinson Xingbo Liu’s career path was re- energy. Since joining TMS in 2002, fuel cells. I want to encourage broad vealed to him in the blink of an eye. he has worked diligently to “broaden interest in the topic so that the audi- He was on a high school fi eld trip to awareness of new materials in new ar- ence will want to learn the details on the Institute of Metal Research, about eas, such as renewable energy sources their own.” a mile from his home in Shenyang, and fuel cells.” Currently the JOM ad- As for the Materials in Clean Power northeastern China, when he peered visor for the Energy Committee, Liu Systems V symposium, Liu said that he through a transmission electron mi- is also vice chair of the High Tem- is particularly pleased that this year’s croscope (TEM) for the fi rst time. “I perature Alloys Committee and was a program features a strong energy stor- remember seeing materials down to founding member of the TMS Energy age component, focused primarily on the atom scale and thinking, ‘That’s so Conversion and Storage Committee. energy conversion and generation. cool.’ Then, I thought, ‘I want to do This year marks two new milestones The larger context for Liu’s scien- that,’” he recalled. “It was amazing to in his TMS involvement. He is a lead tifi c interests is currently shaped by his see something so tiny.” symposium organizer for the fi rst time 18-month-old son, Ethan—“My new It was little wonder that Liu was at the TMS 2010 Annual Meeting, inspiration,” he notes. An avid reader taken with the view through the TEM. namely for Materials in Clean Power of philosophical works, Liu is particu- With his father’s being a mechanical Systems V: Clean Coal-, Hydrogen larly intrigued by the contributions of engineer and his mother a metallurgist, Based-Technologies, Fuel Cells, and Bertrand Russell, considered a found- the lure of materials science was al- Materials for Energy Storage. And, he er of analytic philosophy. “To me, phi- most a genetic predisposition. His pur- was named the 2010 TMS Early Ca- losophy is like reading a review paper suit of “the cool” initially took him to reer Faculty Fellow. about the universe,” he said. “Every the University of Science and Technol- As a requirement of the Early Career time I study it, I fi nd something new. It ogy, Beijing, where he was ultimately Faculty Fellow award, Liu is present- challenges me and keeps me from just drawn to research superalloys for jet ing the TMS Young Leader Tutorial focusing on what is going on today.” engine turbines. “It was just really ex- Luncheon Lecture at the Annual Meet- Occasionally, Liu also fi nds time citing for me to be able to help make ing and, not surprisingly, his topic is to play tennis, although since Ethan’s something that could fl y,” he said. “Energy Materials—Past, Present, and birth, he admits to “playing more Wii Now an assistant professor in the Future.” Said Liu, “I’m going to make tennis.” His wife, Li Zhao, fi rst taught Mechanical and Aerospace Engineer- this fun. I plan to present a picture him the sport in China in the late ing Department at West Virginia Uni- about how energy use has evolved, 1990s. “I started playing tennis with versity, Liu has turned his attention to from ancient India, to the invention my friends and they became frustrated something that he fi nds not only ex- of the steam engine, to the impact of with me because I was a beginner and citing, but also urgently needed—the high temperature, nickel-based alloys they wanted to play more competitive development of sustainable sources of and the development of high effi ciency games. They said, ‘We know someone who is more patient who can teach you,’ and that’s how I met my wife. She taught me for half a year and then we started dating. Then she lost her patience. “She never did teach me how to serve.” Lynne Robinson is a news and feature writer for TMS.

Figure 1: Liu with Each month, JOM profi les a TMS his son, Ethan— member and his or her activities both in “My new inspira- and out of the realm of materials science tion”—as they hike and engineering. To suggest a candidate Mary’s Peak on the for this feature, contact Maureen Byko, Oregon Coast. JOM editor, at [email protected]. Volume 5, Number 1

The Newsletter of the TMS YoungYoung Leaders Committee

MEET A YOUNG LEADER: BEN POQUETTE ON BUILDING A BUSINESS LESSONS LEARNED FROM THE EMERGING LEADERS ALLIANCE Ben Poquette can personally attest to the consumed by business considerations, such challenges of moving a new technology from as marketing and safety. Preparing future engineering leaders for the laboratory to successful commercial ap- the challenges that they will face in advanc- plication as president Q. What accomplishments with Keystone ing their organizations is the intent of the of Keystone Materials Materials are you proudest of so far? Emerging Leaders Alliance, launched in LLC, a company he A. We were able to bring our GraphiMetal™ 2008. A joint venture of leading engineering founded with the other coatings technology from laboratory to societies, the Alliance offers a comprehen- inventors of the patent pilot scale manufacturing within the fi rst 12 sive leadership development experience cul- pending GraphiMe- months of our existence with strictly private minating in a three-day Capstone Program. tal™ coating process. funding. Also, the ability of GraphiMetal™ As a partner organization of the Alliance, (A detailed case study coatings to facilitate high quality solder join- TMS was proud to send three of its young on GraphiMetal™ is ing to graphite foam, using traditional solder leaders to the Capstone Program in October posted in the Established Materials Com- fl uxes and alloys. has generated signifi cant 2009 (Figure 1). Said one of them, Don munity archive of Materials Technology@ interest from the private sector. Siegel, University of Michigan, “The most valuable aspect of my experience was learn- TMS.) Poquette offered the following per- Q. What TMS resources have you found ing how to acknowledge the varied social spectives and advice on building a technol- to be particularly helpful? styles of my peers.” ogy business. A. Networking, networking, networking. Amy Clarke, Los Alamos National Even as an undergraduate student, I realized Q. How did Keystone Materials begin? Laboratory, who also participated, said, “I A. While still a PhD candidate at Virginia that the TMS membership is populated by am more aware of my leadership strengths Tech, I was asked to co-advise a senior great people, and being involved is your and weaknesses and have identifi ed areas design team with my faculty advisor, Ste- ticket to meeting those people. for improvement that will result in personal phen Kampe. The team’s project eventually Q. What advice would you give to others growth, improve my ability to work effec- focused on highly adherent metal coatings on who are considering their own business? tively with others, and expand my future graphite foam. By graduation, the students A. Put together a solid team. Your product career opportunities.” had submitted the application for a provi- or idea may be outstanding, but it takes a “Participating in the Emerging Leaders sional patent. Roughly a year later, I joined Alliance truly exceeded my expectations,” great team to bring real success. It may be Dr. Kampe to create Keystone Materials in an echoed TMS’ third participant, Michele tempting to pull together a group of friends effort to scale up the GraphiMetal™ coating Manuel, University of Florida. “I know that or even a group made up of strictly technical process and make it available for commercial the relationships that were fostered at this folks. However, both of these approaches can and defense applications. meeting will result in long-term professional be a recipe for failure. Ideally, the team has partnerships.” Q. What have been some of the key chal- members who trust each other, work well Applications for the 2010 program start lenges in developing your business? together, and have a complimentary mix of being accepted in the spring. For more infor- A. Having a technical background, the key backgrounds and contacts. This, coupled mation, go to the Emerging Leaders Alliance challenge to me personally was the realiza- with a strong leader, can make nearly any Web site at www.emergingleadersalliance product successful. tion that the majority of our efforts would be .org.

THE YOUNG LEADER TMS YOUNG LEADER A Young Leader is any TMS profes- COMMITTEE OFFICERS sional member in good standing age 35 or under. The goals of the TMS Young Lead- Gregory Thompson, Chair ers Committee are to recognize young professionals, develop in them an appre- Alpesh Shulka, Vice Chair ciation and awareness for TMS activities, Ben Poquette, Secretary provide services specifi cally tailored to young members, and encourage network- Subhadarshi Nayak, Past Chair ing with TMS leaders and prominent so- Figure 1. TMS’ representatives at the 2009 ciety members. For more on TMS Young Emerging Leaders Alliance Capstone Leader activities, visit www.tms.org Program (left to right): Michele Manuel, /YoungLeaders/YoungLeaders.html. Don Siegel, and Amy Clarke. TMS ANNOUNCES 2010 YOUNG LEADER PROFESSIONAL DEVELOPMENT AWARD WINNERS

Congratulations to the ten young professionals selected to receive Paul Ohodnicki the 2010 TMS Young Leader Professional Development Award. Chosen Electronic, Magnetic, & Photonic Materials from each of the fi ve TMS technical divisions, the winners receive sup- Division port from the TMS Foundation to attend two TMS technical conferences Paul Ohodnicki is a research and development and are provided with opportunities to become more involved with the engineer with the vacuum coatings group at Society and to network with TMS leadership. PPG Industries, where he works on new prod- Frank Balle uct development of large area glass coatings. Light Metals Division He received a B.S. in both engineering physics A junior scientist at the Center for Mathemati- and economics at the University of Pittsburgh cal and Computational Modeling, University of and earned his Ph.D. in materials science and engineering at Carnegie Kaiserslautern, Germany, Frank Balle is cur- Mellon University. rently researching innovative solid state joining techniques for light metals, as well as ultrasonic Andre Phillion techniques for characterization of the monotonic Light Metals Division and cyclic deformation behavior of composites. An assistant professor with the School of Engi- He holds a M.Sc. in mechanical engineering and a Ph.D. in materials neering at the University of British Columbia, science from the University of Kaiserslautern. Andre Phillion previously worked as a post-doc- Weisheng Cao toral fellow with the Computational Materials Structural Materials Division Laboratory at the Ecole Polytechnique Fédérale As a materials scientist with CompuTherm de Lausanne, Switzerland. His current research LLC, Weisheng Cao is researching the devel- interests include developing new experimental opment of integrated computational tools for and mathematical modeling tools to understand the relationships multi-component phase diagram calculation between microstructure and physical properties of multi-phase materi- and materials property simulation, as well as als during phase transformations. thermodynamic, kinetic, and processing mod- eling for potential applications in industry. He Doug Spearot holds both M.S. and Ph.D. degrees in materials science and engineering Materials Processing & Manufacturing from the University of Wisconsin-Madison. Division Nitin Chopra Doug Spearot is an assistant professor in the Electronic, Magnetic, & Photonic Materials Department of Mechanical Engineering, Uni- Division versity of Arkansas, where his research focuses Nitin Chopra is an assistant professor in the Met- on computational materials science, nanoscale allurgical and Materials Engineering Depart- material behavior, and multiscale structure-prop- ment at the University of Alabama. His current erty relationships. He holds a B.S. in mechanical research interests include nano/microfabrica- engineering from the University of Michigan and earned his M.S. and tion, nanostructure growth, and nanoscale Ph.D. in mechanical engineering, with a focus on computational materi- heterostructures. Chopra earned a B.S. degree als science, from the Georgia Institute of Technology. in materials and metallurgical engineering from the Indian Institute of Technology and a Ph.D. in materials science and engineering from the Kinga Unocic University of Kentucky. Structural Materials Division Rachel DeLucas Kinga A. Unocic is using advanced characteriza- Extraction & Processing Division tion techniques to study high temperature ther- Rachel DeLucas is a process engineer in the mal barrier coatings and thermally grown oxides melting operations and SGQ plant at H.C. on metallic alloys as a research staff scientist Starck. She holds an M.S. in manufacturing with the Corrosion Science and Technology engineering from Boston University and a B.S. Group at Oak Ridge National Laboratory. She in archeology and materials from the Massachu- received her M.S. degree in metallurgical engi- setts Institute of Technology. neering from AGH-University of Science and Technology in Krakow, Poland, and her Ph.D. in materials science and engineering from Ohio State University. Sandip Harimkar Materials Processing & Manufacturing James Yurko Division Extraction & Processing Division An assistant professor at Oklahoma State Uni- James Yurko is the co-founder, process develop- versity, Sandip Harimkar holds a Ph.D. in mate- ment manager, and chief engineer of Electrolytic rials science and engineering from the Univer- Research Corporation LLC. He received his sity of Tennessee and an M.S. in metallurgical B.S. in materials science and engineering from engineering from the Indian Institute of Science. the University of Michigan and holds a Ph.D. His research interests encompass the processing in metallurgy, with a minor in business, from the and characterization of advanced materials. Massachusetts Institute of Technology. Volume 17, Number 1

The Newsletter for Student Members of The Minerals, Metals & Materials Society

GRADUATE STUDENT SPOTLIGHT: PENN STATE 2010 STUDENT AWARD WINNERS AND CASE WESTERN RESERVE UNIVERSITIES Each year, the TMS Foundation and TMS technical divisions fund Last fall, TMS established a Graduate Student Advisory Council scholarships and awards to help students pursue their academic goals. to address the professional needs of graduate students within the The application deadline for 2011 scholarships and awards is March society. The council held an initial meeting at the Materials Science 15, 2010. Scholarship applications and submission requirements and Technology 2009 conference in Pittsburgh, Pennsylvania, last can be accessed through www.tms.org/Students/Awards.html. The October and will reconvene at the TMS 2010 Annual Meeting in following student members received this year’s awards. Seattle, Washington, this month to further look at ways that TMS can 1. TMS J. Keith Brimacombe meet the unique needs of graduate students. Presidential Scholarship On campus, many schools have their own organizations to serve “I feel incredibly honored and privileged to receive the these students. This edition of Professional Preface looks at the J. Keith Brimacombe Presidential Scholarship Award. activities of the on-campus graduate student groups at Pennsylvania This award will generously help fund my fi nal year of State University and Case Western Reserve University. undergraduate education, and give me the opportunity to attend the TMS conference to network with profes- Pennsylvania State University sionals in my fi eld and further develop my research interests.” In the summer, when most undergraduates head home, student 1. Kelsey Stoerzinger, Northwestern University activities on campus often come to a halt. Last year, the graduate students at Pennsylvania State University (Penn State) decided to take 2. TMS Outstanding Student Paper, advantage of this downtime to plan a lecture series geared toward Graduate Division, First Place their own needs, according to James Saal, a doctoral student in Penn “TMS has been and will continuously be a very impor- tant part in my career development since it was fi rstly State’s materials program and vice chair of the TMS Graduate Student introduced to me by my advisor, Dr. Xinghang Zhang, Advisory Council. in 2005. I am very delighted and honored to be the The graduate students organized a lecture series that featured talks recipient of this prestigious award. I would like to express by faculty members and alumni from the materials science and my sincere gratitude to TMS for providing various good engineering department and focused on post-graduate career options 2. opportunities not only for me, but also for many other for students. The lecture series also incorporated presentations from young scientists.” faculty members in other disciplines, including a talk by one of Penn Engang Fu, Texas A&M University State’s business professors on entrepreneurship. The summer series ended with a tour of a steel plant outside Harrisburg, Pennsylvania. 3. TMS Outstanding Student Paper, For the plant tour, it made sense, Saal said, to coordinate the trip Graduate Division, Second Place with the undergraduate students at Penn State in order to benefi t as “It is a great honor for me to be recognized by an out- many students as possible and to help defray the cost of the trip. The standing society such as TMS with this award. I would undergraduate and graduate groups are mostly separate organizations, like to dedicate this award to my family and my advisors, but there is a lot of coordination among the different materials groups Profs. Peter K. Liaw and Hahn Choo, for their invaluable on campus. Last year, Penn State’s graduate group participated in the support and guidance throughout my Ph.D. study. I would also like to thank TMS for their continuous support American Cancer Society’s Relay for Life. In 2010, they plan to 3. to student members. This award has already boosted participate again in this competitive fundraising event and may even my confi dence and has encouraged me to continue to challenge the undergraduate Material Advantage club at their school pursue and achieve my education and career goals.” to form their own team. Soo Yeol Lee, University of Tennessee Other plans for Penn State’s graduate students in 2010 include activities as varied as movie nights and etiquette dinners. The graduate 4. Electronic, Magnetic & Photonic students also plan to arrange meetings and luncheons that encourage Materials Division Gilbert Chin student interaction with the distinguished speakers who regularly visit Scholarship the campus as part of the materials department’s seminar series. “I’m so honored to be the recipient of the EMPMD Gilbert Chin Scholarship. I have dedicated many hours Case Western Reserve University 4. to extracurricular research, and am so pleased to see that it has paid off. This scholarship also encourages me In 2009, the Graduate Materials Society (GMS) at Case Western to continue my education and remain an active part of Reserve University added a new event to its annual line-up: a graduate TMS in the future. Thank you EMPMD of TMS!” student research colloquium. In November, the GMS sponsored an Sarah Miller, Washington State University event in which three students gave 15-minute presentations on their 5. Extraction & Processing Division research. Scholarship It was a volunteer-based activity just for graduate students, “I am extremely grateful to be a recipient of the generous according to Lisa Deibler, treasurer of the GMS and secretary of EPD Scholarship for the 2009–2010 school year. It is TMS’s Graduate Student Advisory Council. The goal was to create nice to have the backing of such a great organization, as a casual environment where graduate students could share their 5. TMS is, to help me get through my fi nal year in college. (CONT. ON NEXT PAGE) (CONT. ON NEXT PAGE) 2010 STUDENT AWARD WINNERS (CONTINUED) The fi nancial, academic, and multiple other resources 12. Materials Processing & Manufacturing TMS has provided me with are above and beyond what Division Scholarship would normally be expected from a student’s standpoint. “I am honored to be a recipient of the MPMD Scholar- Thank you once again for your generous gift.” ship and to join the prestigious list of others who have Eric Young, South Dakota School of Mines received this award before me. As a second year recipi- ent, I feel that the previous support of TMS is making 6. Extraction & Processing Division me better all the time. Thanks to the support of TMS, Scholarship my future is looking brighter all the time.” “I feel very fortunate for the opportunities I have received Eric Young, South Dakota School of Mines 6. as a metallurgical engineering student and as a member 12. of TMS. I am grateful for the scholarship I have received 13. Materials Processing & Manufacturing from TMS and the ability to continue studying and being Division Scholarship a part of the up and coming technology that is currently “I am exceedingly grateful to TMS and the Materials being developed.” Processing & Manufacturing Division for honoring me Ryan Morrison, University of Utah with the MPMD Scholarship. This award is a valuable 7. Extraction & Processing Division contribution to my materials education, and it encour- Scholarship ages me to pursue greater endeavors in education, pro- “It is a great honor for my efforts to be recognized by fessional development, and materials research. I would the TMS. This award will help me to pursue graduate also like to thank the fantastic faculty at Washington studies in metallurgy.” State University for their unwavering support.” Kalen Jensen, University of Alberta Whitney Patterson, Washington State University 7. 13. 8. Extraction & Processing Division 14. Structural Materials Division Scholarship Scholarship “Thank you so much for the opportunity you have “As a student, I am extremely thankful for the com- provided me by choosing me as the recipient for the mitment of TMS in developing the next generation of Extraction & Processing Division Scholarship Award. professional engineers. The organization is not only I am very grateful to receive this award. Not only will a great technical resource, but it also provides me the this scholarship help to provide fi nancial assistance opportunity to interact with and learn from other people for the upcoming school year, it also gives me great within the materials community.” encouragement for my future in the materials science David Cole, University of Cincinnati fi eld. I will strive to do the very best I can and I hope to be an important asset to the study of metallurgy and materials science.” 14. 15. Structural Materials Division 8. Erin Diedrich, Washington State University Scholarship “I would like to thank you for awarding me the Structural 9. Light Metals Division Scholarship Materials Division Scholarship. This scholarship has “It is truly an honor to receive the TMS Light Metals Divi- given me the opportunity to continue to excel in my fi eld sion Scholarship. This scholarship will help to alleviate of study. It has assisted me in funding my education. the cost of attending The Pennsylvania State University Given this scholarship I will be able to focus more on my and to further my education in the fi eld of materials academics and continue to strive to keep my scholarship science and engineering. My research with magne- and to be honored for more scholarships. This will help sium and aluminum has helped me to understand the me in my achievements and provide encouragement to importance of light metals and I hope to contribute to this be more successful. I appreciate this scholarship very specifi c area of metallurgical sciences. The Department much, and thank you for the support.” of Materials Science and Engineering at Penn State 15. Mohamad Zbib, Washington State University and my research advisor, Dr. Zi-Kui Liu, have greatly 9. helped me to fully embrace the fi eld of materials. I am proud to be recognized by TMS and appreciate the GRADUATE STUDENT SPOTLIGHT (CONTINUED) encouragement to pursue excellence in the fascinating fi eld of materials science and engineering.” research progress with their colleagues. The three presenters Laura Jean Lucca, Pennsylvania State University represented various stages of the process, from the beginning stages 10. Light Metals Division Scholarship of master’s degree research to a nearly fi nished doctoral project. “I am very honored to receive such a prestigious award While some of Case Western’s GMS activities are exclusively for from one of the largest materials science and engineer- graduate students, Deibler thinks that coordination with their ing societies in the world. I am truly grateful to TMS undergrad counterparts is also valuable. for their generosity and encouragement. The provided “Undergraduates have a different focus,” Deibler said, explaining fi nancial support will help me pursue my studies and the need for separate groups. “But it’s good to do things together, too, 10. research and attend future conferences. I would also like so14. they can see what it’s like to be a grad student.” to thank the students and professors of the Materials The two groups come together for a weekly department lunch, an Engineering Department of McGill University for their event typically attended by 20–30 people. Everyone pitches in a few passion and enthusiasm.” dollars, and students take turns preparing lunch. Other GMS social Dmitri Nassyrov, McGill University and community activities in 2009 included a Grad Student Night Out at a local billiard hall, barbeques, and participation in the American 11. Light Metals Division Scholarship Cancer Society’s Relay for Life fundraising project. “This scholarship gives me the funds to continue my For 2010, the Case Western GMS is planning to repeat its successful education. It allows me to take the 18 credits I want to, research colloquium and other events, but also is developing new instead of working at a minimum wage job. This scholar- activities. One new project encourages materials students to share ship will allow me to devote more time to my classes, their favorite hobbies. Students will meet regularly to take turns meaning I will not only continue my education but I will teaching their fellow students about their outside interests, opening receive a better, more complete education because of these grad students to a whole new set of skills that can’t be found in 11. this scholarship. And for that, I thank you.” a materials classroom. Michael Fank, University of Wisconsin Feature Materials World

Enhanced for the Web This article appears on the JOM web site (www.tms.org/jom.html) in html ¯format and includes links to additional on-line resources. Advanced Materials

Figure 1. “A bobsled is a fascinating Are a Game vehicle,” notes Ulrich Suter, professor in ETH Zurich’s Department of Materials and CITIUS project coordinator. “It has no Changer in the engine and derives its kinetic energy solely from the gravitational acceleration in the ice channel and the initial thrust by the crew before they board the sled.” Winter Olympics (Courtesy ETH Zurich)

Lynne Robinson

It starts with a push. Figure 2. SBSV athletes and Propelled by the strength and col- the CITIUS endure the wind lective will of four powerful sprinters tunnel test, where the sled was challenged to withstand a 150 gripping its sides, an Olympic bobsled kilometer per hour headwind. launches from a frozen start ramp, ac- (Courtesy ETH Zurich) celerating from a standstill to 40 kilome- ters an hour in less than fi ve seconds. At about 50 meters into the race, the team hops into the sled one by one, with the pilot already preparing to hit the fi rst of 16 turns built into the ice-encased track of the Whistler Sliding Centre, site of the bobsleigh competitions for the 2010 Vancouver Winter Olympics. Like a fi - Figure 3. Dany Heatley, who will play for Team berglass bullet, the sled shoots down a Canada at the 2010 Winter vertical drop of 152 meters, subjecting Olympics, puts the high- its crew to crushing G-force and bruis- performance fi ber com- posites in his equipment ing jolts around the curves, while at- to the test. The adoption of taining speeds of nearly 150 kilometers advanced composite ma- an hour. Victory at the end is measured terials has revolutionized hockey stick design, both in in hundredths of a second, with almost terms of weight and perfor- imperceptible factors deciding the dif- mance. Since a composite ference between a medal run and a dis- structure is a laminate of individual layers, there is appointing fi nish. much greater control over For the 2010 Olympics, the Swiss a stick’s stiffness and mass Bobsleigh Federation (SBSV) wanted distribution, as well as the strength and toughness to make sure that those factors had profi le, according to Travis nearly nothing to do with the construc- Downing of Easton Hockey. tion of the sled. Under the project name The anisotropy of fi ber also enables the freedom to con- CITIUS—which echoes the Olympic trol the strength, stiffness, motto, “Citius, Altius, Fortius (faster, and torsion characteristics higher, stronger)—SBSV marshaled of a stick based on fi ber ori- entation in each layer of the the talents and resources of the Swiss laminate. (Courtesy Easton Federal Institute of Technology (ETH Hockey)

Vol. 62 No. 2 • JOM www.tms.org/jom.html 17 Zurich), a consortium of industrial part- as “an alloy of iron and carbon with an the CITIUS Dossier posted on the ETH ners, and current and former bobsleigh iron content of more than 50 percent,” Zurich Web site. athletes to work on “building a better while “rubber or rubber-like material” A WEIGHTY MATTER bob” (Figure 1). can be incorporated into the sled for the purpose of damping vibrations. Hulls Whether smoothing the performance THE SCIENCE OF THE SLED have no prescribed material at all. Suter of a bobsled careening down the Whis- “An initial question we asked our- said he was surprised, actually, at “how tler ice track, or giving an aerials skier selves was ‘what slows down a bob- many unspecifi ed areas were left in a an extra bounce into the Canadian sky, sled?’” said Ulrich Suter, professor in seemingly very complete—some might advanced materials technologies like ETH Zurich’s Department of Materials say overly complete—book of rules.” those used in the CITIUS project will and CITIUS project coordinator. “After Working within these parameters, a be on prominent display at this year’s a careful analysis, we decided to break team of more than 20 ETH Zurich sci- Winter Olympics. The quest for a com- the problem into a number of simpler entists and engineers from a wide range petitive edge through advancements in tasks. Aerodynamics is clearly impor- of disciplines applied their specifi c ex- engineering isn’t limited to these elite tant, especially in the very fast track pertise to different aspects of the CI- athletes, however. Said Travis Down- at Whistler. The shape and surface TIUS bobsled. A battery of tests re- ing, a materials engineer in research and structure of the runners are also very vealed the sources and nature of vibra- development at Easton Hockey, Cali- relevant. The hull and chassis materi- tions, which forces were exerted where fornia, “Once full composite hockey als, as well as the dynamics within the on the chassis, and the specifi c stiffness sticks were introduced onto the market system, play a crucial role in the me- and strength of certain materials. “Our about ten years ago, they immediately chanical dissipation of energy in the simulations indicated that several ap- became the ‘must have’ stick among bobsled. And, above all, safety is para- parent ineffi ciencies of the sleds might professional players. Subsequently, mount—the sled must survive a crash be avoided with an appropriate choice amateur players were quick to jump at full speed and protect the crew. We of materials,” said Suter. “However, on that trend, as well. The fact that it numerically simulated all key aspects almost all standard elastomeric materi- is nearly impossible to fi nd a wooden of the sleds and tried to isolate the most als were inadequate to the demands of hockey stick anymore speaks volumes important factors, and then experimen- the sport. And, many tested adhesives to the adoption rate of composite sticks” tally tested our assumptions to optimize proved insuffi cient under the (dynam- (Figure 3). the design.” ic) mechanical and thermal stresses of A key advantage of composite A particular challenge for CITIUS bobsled racing.” hockey sticks over wooden ones, said was developing a state-of-the-art bob- To address these issues, the research Downing, is weight. Wooden sticks sled in accordance with the rules for team created elastomeric damping routinely tip the scales at more than construction and design enforced by materials specifi cally for the CITIUS, 650 grams, while modern composite the Fédération Internationale de Bob- while also developing a composite hull sticks can be as low as 400 grams. By sleigh et de Tobogganing (FIBT), the strengthened with long-fi ber reinforce- being lighter, composite sticks are governing organization for competitive ment. Once the components were as- easier to control and provide for great- bobsledding. The runners, for instance, sembled into prototype sleds, athletes er swing velocity. Downing said that have to be constructed from a rigidly de- from the SBSV tested the CITIUS in composite hockey sticks have also fi ned steel procured from a prescribed the wind tunnel at the Audi works in been found to be superior to wooden source, with any type of modifying Ingolstadt, Germany (Figure 2). The ones in storing and rapidly releasing treatments “forbidden, including those real moment of truth for the project oc- strain energy, which enables players which even cause only a local variation curred when, bristling with sensors to to boost their shot velocity. From a of the physical characteristics and/or measure forces, driving dynamics, and manufacturing standpoint, composites of the composition and/or structure of aerodynamic properties, the CITIUS make it possible to reproduce a hock- the material.” Steels used elsewhere on hurtled down an ice track for the fi rst ey stick’s properties exactly each time, the sled are more generally described time near Innsbruck, Austria. Although while the organic nature of wood tends highly guarded about revealing de- to introduce inconsistencies from stick tails of the CITIUS’ development and to stick (Figure 4). performance, the research team was “For most sports, lowering the pleased with the results, according to weight of the equipment without re-

Figure 4. Easton’s newly introduced S19 hockey stick demonstrates the fl exibility that composites offer in designing sports equipment to save weight and improve control. The stick has a traditional rectangular handle section with a unique elliptical cross section in the lower section of the shaft, increasing torsional rigidity and ensuring the blade face stays square to the puck throughout the shooting motion. (Courtesy Easton Hockey)

18 www.tms.org/jom.html JOM • February 2010 ducing performance is highly desir- able,” said Katharine Flores, an associ- ate professor at Ohio State University and TMS member who was tapped by NBC to contribute to its Science of the Olympic Winter Games educational series. “This can be addressed both by improved materials and improved processing technologies. Beyond that, the design really depends on the sport. With skis, for instance, the most important consideration is probably stiffness. If the ski defl ects or twists a lot, that steals kinetic energy—and speed—from the skier” (Figure 5). “Perhaps the most interesting thing about sports equipment is what doesn’t drive change—cost,” Flores continued. “Even recreational users are willing to pay more for perceived improvements in performance.” Many of the advancements in sport- Figure 5. Skis have evolved from simple wooden slats to complex systems of layered materials, customized to a specifi c use. Stiffness is a highly desired characteristic for ing equipment owe their origins to the downhill skiers, who depend on it for control as they fl y down a slope. Joey Discoe, seen aerospace industry, particularly in the here in moguls training, needs a more fl exible ski to execute the jumps, fl ips, and spins application of composites and light- required in this competition. Specialty alloys, carbon fi ber, and even piezo-ceramics for vibration damping have been incorporated into modern ski design. (Courtesy Todd weight alloys. Said Downing, “The Shirman, U.S. Ski Team) entire sports equipment spectrum— from baseball and softball bats, to ten- nis rackets, to golf shafts, to bicycle Sports, of which Easton Hockey is man’ being the operative word. A lot and motorsports structures—uses es- part, has already had initial success of people feel that it really shouldn’t sentially the same carbon fi ber and adding functionalized multi-wall car- be about who has the latest and great- two-part epoxy that was originally de- bon nanotubes (CNTs) to strengthen est gadget. There is a fi ne line between veloped for military and aerospace ap- epoxy resin for bat, bicycle, and hock- engineering the equipment to highlight plications.” ey applications. “It has been diffi cult the capabilities of the athlete—freeing Even with space age technology, to realize the full promise that single- him or her from various physical con- the quest for optimum performance in wall CNTs has shown in the lab, be- straints—and engineering the equip- sporting equipment can be challeng- cause the processing technology need- ment to artifi cially improve the perfor- ing. “Because hockey is such a fast- ed to exploit their full potential has not mance of the athlete. paced sport with frequent collisions, it reached maturity,” said Downing. “The “For the recreational athlete, howev- becomes crucial to have the proper bal- next few years, however, will see addi- er, I think these ethical considerations ance of performance and durability in tional leaps forward in nanostructured go away,” she continued. “There’s a stick,” said Downing. “The unfortu- composites for sports applications.” nothing wrong with making equipment nate reality is those factors are largely As exciting and anticipated as these that improves the performance and, as in opposition to one another, given the potential advancements may be, Flores a result, the experience of recreational desire to maintain the lightest structure noted that concerns about technology players so they will enjoy the game possible. Composite suppliers tend to overshadowing technique in the elite more and stick with it.” address the problem by adjusting the sports world beg to be addressed, fu- ACKNOWLEDGEMENT epoxy formulation, with variation in eled by such recent controversies as fi ber properties also factoring into the the Speedo¨ LZR Racerª suit worn JOM would like to acknowledge equation through alteration of tensile by Michael Phelps in the 2008 Beijing Francis H. (Sam) Froes, retired direc- modulus, strength, and elongation to Olympics. tor and department head, Materials failure.” “There are some important ethical Science and Engineering Department, considerations at work in the use of ad- University of Idaho, and TMS member TECHNOLOGY VS. vanced materials and technological ad- for the insights and background infor- TECHNIQUE vancements in general,” she said. “The mation he contributed to this article. The next big advancement in sports Olympics, in particular, are supposed materials technology may be launched to be about reaching for the pinnacle of Lynne Robinson is a news and feature writer with by the power of small. Easton-Bell human athletic achievement, with ‘hu- TMS.

Vol. 62 No. 2 • JOM www.tms.org/jom.html 19 

By Land, Sea or Air Magnesium Elektron Alloys are there….

Magnesium Elektron is a world leader in the  production of high performance magnesium alloys and products. Magnesium Elektron began production of magnesium in1936 and subsequently developed a wide range of magnesium alloy systems and products. Over time, Magnesium Elektron has patented 21 magnesium alloys for a wide variety of applications. New magnesium alloys recently developed by Magnesium Elektron include Elektron®21 and Elektron®675, as well as Elektron®WE43 now offered in rolled form.

 Magnesium Elektron has 7 manufacturing sites located in US, Canada, United Kingdom and Czech Republic, all dedicated to producing high performance alloys in cast, wrought and powder form. In addition, Magnesium Elektron operates a fully integrated magnesium engraving plate business, while also producing zinc, brass and copper engraving plates and support chemicals which serve the Graphic Arts industry.

 The company is part of the Luxfer Group, an international manufacturer of high performance engineering materials including magnesium alloys, zirconium products and high pressure lightweight gas cylinders. Magnesium Elektron personnel will be presenting 5 technical papers at TMS 2010 139th Annual Meeting and Exhibition. Please visit us at Booth #437 on the exhibition floor of the Washington State Conference Center. Also visit our website at www.magnesium-elektron.com for contact information.   Magnesium Elektron North America 1001 College Street PO Box 258 Madison, IL 62060 Phone: 800-851-3145 Web: www.magnesium-elektron.com  Please visit us at TMS 2010 booth 437. SIMPLIFIND Tap into the incredible network of The Minerals, Metals & Materials Society with The Minerals, Metals & Materials Society Marketplace. Powered by MultiView, the Guide gives materials and metallurgy professionals a faster and easier way to find great vendors.

Simplifind your search today at tmsmarketplace.com. Visit the Knowledge Resource Center to reserve your copy today! REWAS 2008: Global Symposium on Recycling, Waste Treatment and Clean Technology by B. Mishra, C. Ludwig, and S. Das, editors The 2008 Global Symposium on Recycling, Waste Treatment and Clean Technology (REWAS 2008) is the third in a series of international conferences to address the continuing globaliza- tion of environment protection through progress in recycling technology, re-engineering of the production system, and clean technologies. The conference incorporates the Fifth International Symposium on Recycling of Metals and Engineered Materials and R’08 - the World Congress on the Recovery of Materials and Energy for Resource Effi ciency. Technical topics covered in the conference proceedings include:

• Aluminum By-product Recovery and • Ferrous Metals Secondary Production • General Recycling and Solid Waste • Automotive Recycling Processing • • Case Studies in the Development Heavy-Metal Containing Waste of Waste Treatment Technologies • Nonferrous Metals • Clean Technology and Re-engineering • Precious and Rare Metals of Current Processes • Radioactive Waste • Composite Materials • Refractory Metals • Design and Engineering of Waste • Remediation of Contaminated Soil Treatment Plants and Industrial Sites • Economic Evaluation of Waste Treatment • Resources, Monitoring and Strategies Characterization of Waste • Electronic Waste • Treatment of Liquid and Gaseous Effl uents • Environmental Issues Related to Waste Storage and Recycling • Waste Conversion and Reutilization

TMS Member price $64 TMS Student Member price $49

To order these or related publications, contact TMS: E-mail [email protected] • Phone (724) 776-9000, ext. 256 • Fax (724) 776-3770 Aluminum Overview Aluminum Recycling—An Integrated, Industrywide Approach

Subodh K. Das, John A.S. Green, J. Gilbert Kaufman, Daryoush Emadi, and M. Mahfoud

The aluminum industry is a leading it can be essentially recovered during MODELING THE URBAN proponent of global sustainability and recycling. Thus, from a sustainability ALUMINUM MINE strongly advocates the use of recycled viewpoint, aluminum recycling is tap- metal. As the North American primary ping into a convenient “urban mine” of Modeling, together with some life aluminum industry continues to move material that enables reuse while sav- cycle assumptions, has also been used offshore to other geographic areas ing energy and reducing environmen- to explore the impact of recycling on such as Iceland and the Middle East, tal impacts. This urban mine analogy the overall U.S. metal supply.10 This where energy is more readily available is becoming increasingly real as the “urban mine” model assumes that at lower cost, the importance of the amount of aluminum in vehicles con- the amount of scrap that will become secondary (i.e., recycled metal) market tinues to increase in order to improve available in a given time frame can in the U.S. will continue to increase. performance, reduce weight, increase be predicted based on the amount of The purpose of this paper is to take an fuel effi ciency, and enhance safety. material sold into specifi c markets and integrated, industry-wide look at the See the sidebar for background on the typical life cycle of material in recovery of material from demolished aluminum recycling. these markets. For instance, building buildings, shredded automobiles, and components, automobiles, and bever- aging aircraft, as well as from tradi- age cans have useful lives that are, re- How would you… tional cans and other rigid contain- spectively, ~50, ~15, and ~0.2 years, ers. Attempts will be made to assess …describe the overall signifi cance and the amount of metal that is sold of this paper? how the different alloys used in these to these markets is known and can be This paper introduces the concept separate markets can be recycled in the and suggests several implementable forecast into the future with some as- most energy-effi cient manner. ideas to design and commercialize surance. recycle-friendly aluminum alloys for Existing industry data has been INTRODUCTION key market sectors. used for the model. The model inputs In his recent book Aluminum Recy- …describe this work to a include primary production data from cling,1 Mark Schlesinger points out that materials science and engineering 1893, secondary production data from in traditional extractive metallurgy, the professional with no experience in 1913, and imports from 1911. These your technical specialty? raw material used to produce a metal data are merged with market informa- Most of the aluminum alloys used is mined from the earth and then is in the commerce today have been tion for the last 40 or 50 years in order separated from other gangue minerals designed and sold commercially using to project future U.S. metal supplies. and impurities. In aluminum recycling, primary and pure aluminum, pure The model also projects information however, the ore body consists of scrap alloying elements, and sometimes based upon the market size and growth special elements to meet perceived metal found on the ground rather than customers needs. This practice is not rates from 1990 to 2000. Specifi cally in it. He further notes that scrap comes economically and environmentally the percentage of the overall market with its own variety of gangue and im- sustainable as it leads to a larger and the average growth rate over the purities such as paint coatings and met- carbon footprint and lower recycling decade were, respectively: transpor- rates for the aluminum industry al attachments, as well as dirt and other products. tation (37%, 9.8%), containers and contaminants. So, while the recycling packaging (23%, 0.3%), consumer du- process still requires some refi nement …describe this work to a rables (8%, 4.8%), building and con- layperson? and expenditure of energy, the amounts struction (15%, 2.6%), electrical (8%, In order to achieve the sustainability of effort and energy involved are much goals of a manufacturing industry, 3.0%), machinery and equipment (7%, less than those in refi ning mined ores. each of the commercially sold 4.8%), and others (2%, 1.7%). The aluminum industry “Vision” product(s) to its intermediate or Considerable insight has been document,2 published in 2001, notes ultimate customer(s), must be gained from the macro-model of the designed and processed keeping the 10 that the electrical energy embedded recyclability in mind during the entire urban mine. A major conclusion is in metallic aluminum can be consid- phase of its product life cycle. that the automotive area, driven by the ered to be an “energy bank” because number of vehicles (~15 million per

Vol. 62 No. 2 • JOM www.tms.org/jom.html 23 year) and the growing amount of alu- drive toward “sustainable mobility” is Transportation Recycling—The minum used per vehicle, will dominate increasing, and the aluminum recycling Major Recycling Opportunity the future scrap source picture. potential from the automotive sector is As noted in the Modeling section enormous. RECYCLING and in Reference 10, the transportation Automotive recycling is relatively CHALLENGES BY MARKET area is the biggest market for alumi- complex due to the large number of APPLICATION num and the scrap generated from used alloys used. In a parallel paper,12 the Capturing the metal in the recycling autos and trucks now exceeds that from authors have explored an ideal automo- process represents only a part of the recycled beverage cans. Specifi cally, tive recycling scenario addressing these overall recovery process. A more sig- according to the Aluminum Statistical issues. Among the alloys widely used nifi cant challenge is preserving the Report11 the transportation market sec- are Al-Cu (2xxx) alloys 2008 and 2010 economic value of the alloying ingre- tor consumes ~8.7 million pounds or for body sheet; Al-Mg (5xxx) alloys dients by recycling specifi c wrought 33.9% of North American production. 5022 and 5754 for body sheet; Al-Mg- alloys into the same or similar wrought Over the period 2001Ð2005, the annual Si (6xxx) alloys 6111 for body sheet alloys. The ideal is to recycle a specifi c growth rate of aluminum in this market and 6063 for structural components, in- alloy back to the identical alloy, as with has been 5.4%. This growth is driven cluding frames; and Al-Zn-Mg (7xxx) beverage cans; the value of the alloy is by the need to save weight, increase alloys like 7029, 7116, and 7129 for maximized. fuel effi ciency, and enhance safety. The bumpers. Casting alloys like A356.0, 360.0, and A380.0 are used for engine BENEFITS OF AND BACKGROUND ON ALUMINUM RECYCLING and other mounting components. It is The conventional benefi ts that are cited for aluminum recycling are energy savings, apparent from an examination of the 12 emissions reductions, and the elimination of landfi ll wastes. A comprehensive life cycle compositions of these alloys that the analysis conducted by the industry under USCAR guidance, and which involved analysis high-Zn 7xxx alloys are not compatible of 15 unit processes in 213 plants worldwide, verifi ed these savings.3 Specifi cally, it was with the other alloys, and the relatively shown that the production of primary aluminum, when all electrical generation, trans- high Cu in the 2xxx alloys will not fi t mission losses, and transportation fuels are accounted for, requires ~45 kWh of energy well with the 5xxx and 6xxx alloys

and emits ~12 kg of CO2 for each kilogram of metal. On the other hand, the recycling Thus the potential importance of of aluminum requires only ~2.8 kWh of energy and emits only ~0.6 kg of CO2 for each pre-shred dismantling becomes clear. kilogram of metal. Thus, ~95% of the energy and ~95% of the environmental emissions Ideally, all vehicles would be recycled are saved when the metal is recycled. The energy banked in the metal during the initial and would be subject to a pre-shred smelting process is recovered, with minor losses, each time the material is recycled. disassembly process where large com- A newer and more critical benefi t of aluminum recycling is that it helps to maintain a viable North American aluminum fabrication industry. With the declining aluminum ponents of known alloy composition primary industry, the secondary industry based upon recycling now provides the bulk of (i.e., wheels, bumpers, engine blocks, domestic metal to the downstream fabrication industry. perhaps body sheet components like Historically, since the start of the aluminum industry in 1888, some amount of recy- hoods, deck lids, and door panels) cling has always occurred. It has been estimated that ~70% of all the aluminum that has would be removed and segregated for ever been made is still in active use.4 But, it was not until the advent of the aluminum remelting. Hand-held chemical analyz- can in 1965 and the buy-back programs of the Reynolds Metals Company that recy- ers could aid preliminary separation cling became part of the public awareness. The success of the beverage can spawned the and sorting of alloys. As one example, secondary, or recycling, industry. The large number (~10,000) of recycling centers that the batching of all bumpers togeth- emerged nationwide during the 1970s and 1980s has given way to fewer but more ac- er—grouping alloys 7029, 7116, and tive recycling companies with curbside collection of all recyclables that go to municipal 7129—would probably yield a melt recycling facilities (MRFs). There are now some 300 MRFs representing more than half the U.S. population and this phenomenon has generated a subsidized dealer market for that could be directly reused for bum- 12 aluminum cans, along with other recyclable items. pers as a 7xxx alloy. The remainder For the U.S. market, an increase of 1% in aluminum can collection is equivalent to a of the auto hulk would be shredded, saving of $12 million, so renewed effort is being put in that direction. For example, the subjected to laser-induced breakdown Aluminum Association and the Can Manufacturing Institute have formed the Aluminum spectroscopy (LIBS) sorting,8,9 and re- Can Council to develop educational and promotional programs. Other groups have also melted, using the most effi cient process- been active in this effort, notably Secat, Inc., the Center of Aluminum Technology at es to reduce dross losses and maximize the University of Kentucky, and the Sloan Industry Center for a Sustainable Aluminum recovery. Of course, if/when shredding Industry to focus on recycling issues. This consortium has already addressed some of the technology develops to the point where 5Ð7 cultural and societal factors that that infl uence recycling. individual alloys can be identifi ed and Another new development has been the advent of the industrial shredder. The shred- separated, pre-shred disassembly may ding of large industrial scrap such as automobiles, refrigerators, freezers, and other large machinery has been complemented by the developments by Huron Valley Steel Corpora- not be warranted. tion8,9 of color sorting and laser induced breakdown spectroscopy, making it possible to Another option is to revisit the idea analyze the emitted plasma light from a scrap sample and determine its chemical com- of developing some single versatile position. In the future, it should be economically feasible to separate and sort aluminum “unialloy” that could meet many of the scrap on an alloy-by-alloy basis by virtue of the developments on an industrial scale. requirements for a large number of au- tomotive components, thus simplifying

24 www.tms.org/jom.html JOM • February 2010 the scrap stream. This has proven dif- Table I. Nominal Compositions of Potential New Recycling Friendly Alloys Resulting from fi cult to achieve because of the diverse B&C Product Sorting and Remelting performance requirements of different components. However, some progress Alloy Al,% Si Fe Cu Mn Mg Cr Zn Ti has being made, as companies like 505X ~96.0 0.4 0.4 0.15 0.6 2.5 0.15 0.25 0.1 Toyota use alloy 6022 for both inner (- 606X ~96.0 0.5 0.5 0.2 0.12 0.8 0.15 0.15 0.12 O temper, greater formability) and outer (-T4 temper, greater dent resistance) and factories contain much more trations; a mass balance based upon the body components. aluminum, by 1Ð2 orders of magni- expected inputs by alloy and product tude, than residential buildings. would lead to more precise projections. Building and Construction— ¥ Residential construction in warm The projected average 505X compo- A Hidden Opportunity climates contains ~20 times more sition in Table I is similar to a 5052-type The third largest market sector for aluminum than in cold climates alloy, with higher Si and Zn impurities, aluminum, after transportation and con- (i.e., sun shades, refl ectors). Future plus a higher Mn content (0.6% vs. a tainers/packaging, is the building and architectural designs are anticipat- max. of 0.1 for 5052). Based upon the construction (B&C) market that also ed to increase this amount. known performance of commercial includes highway structures (e.g., lamp ¥ Recovery is enhanced when a 5xxx alloys, all of these three variances posts, signs, bridge decks). This mar- subcontractor is involved who may be acceptable for direct reuse of ket has not received much attention in wishes to reuse some components this composition in many products. The regard to recycling, primarily because (i.e., windows), directly. Si and Zn impurities are not likely to be the life cycles of buildings and highway An examination of the 2005 Alumi- troublesome, and the high Mn content structures are so long, ~30Ð50 years, num Statistical Review11 indicates that may lead to slightly higher strength. It plus the smaller amounts of aluminum the 3.7 billion pounds of metal in this would be practical to defi ne a new alloy previously used in B&C applications. market is almost equally divided be- based upon the likely output from B&C Newer architectural designs, however, tween extruded shapes and sheet prod- recycled sheet and plate that could be have emphasized the value of aluminum ucts. The workhorse alloys for B&C ex- reused directly in B&C applications, as to make buildings more energy effi cient trusions are Al-Mg-Si alloys 6061 and well as others perhaps. through the use of sun screens, cladding 6063.14,15 Alloy 6061 provides the best Similarly, the 606X-type represents and shades, curtain wall construction combination of strength, cost, weldabil- a melding of 6061 and 6063 composi- and advanced window designs, and the ity, and corrosion resistance, while alloy tions, with variation to be expected based potential value of recycling. The alumi- 6063 is more readily extrudable with upon the mix of the two at a given time, num components have increased. lower strength and lower cost. By con- which also appears potentially directly Recent studies at the Technical trast, for B&C sheet and plate, the most reusable; it would have useful structural University of Delft13 have provided a widely used alloys are the Al-Mg 5xxx strength when heat treated and aged and comprehensive evaluation of recycling series, ranging from 5050 with ~1.5% would be rather readily extrudable. during the demolition of nine types of Mg to 5456 with ~4.5% Mg. These Thus, if component pre-sorting based buildings located in six countries in alloys are characterized by excellent upon separating rolled and extruded Europe. In this study, conducted under corrosion resistance, toughness, weld- products is adopted during demolition, EU regulations and employing ISO- ability, and moderate strength. While it seems likely that both output remelt certifi ed protocols and contractors, the the 5xxx and 6xxx alloys can usefully compositions could be directly reused in deconstruction and demolition of the be remelted together, it is clearly desir- B&C applications, including highway buildings was closely monitored and able to separate the 5xxx alloys from the structures, with no further processing. comprehensive data were collected. 6xxx alloys prior to melting to most ef- Recycling Aerospace Alloys— The following conclusions were fi ciently reuse the alloying constituents. An Overlooked Opportunity among those drawn from the Delft study, It is fortuitous that the 6xxx alloys are all supporting the recycling of aluminum primarily extrusions and the 5xxx alloys For decades, thousands of obsolete components from B&C structures:13 are primarily sheet and plate. Thus a ru- aircraft have been sitting in the south- ¥ The average aluminum collec- dimentary sorting of 5xxx from 6xxx al- west desert “graveyards,” while the de- tion rate for all nine buildings was loys can be achieved by merely separat- mand for recycled aluminum increases. 95.7%. ing structural extrusions from sheet and The discarded aircraft provide a large ¥ Even though the aluminum content plate materials at the demolition site.16 potential source of valuable metal. How- of the buildings was <1%, there It is useful to speculate upon the alloy ever cost-effective recycling of aircraft was still a substantial mass of alu- content likely to result if the approach alloys is complex because aircraft alloys minum available for recovery. of separating the sheet and plate compo- are typically relatively high in alloying ¥ Much of the aluminum in a build- nents from the extrusions components is elements like Cu (2xxx series) and Zn ing is on the periphery (e.g., air adopted. Table I illustrates two potential (7xxx series), and may contain low lev- conditioners, sun shades and aw- output compositions from remelting els of minor elements to optimize frac- nings, etc.) and is easy to recover. these two separations, designated 505X ture toughness. ¥ Non-residential high-rise buildings and 606X. These are speculative illus- Again it would seem that dismantling

Vol. 62 No. 2 • JOM www.tms.org/jom.html 25 areas to more precisely determine the Table II. Potential Compositions of Some Recycled Aircraft Alloys Assuming Pre-Sorting realistic opportunities and likely remelt Alloy Al Cu Fe Mg Mn Si Zn Others compositions. Secat has proposed to R2xxx ~93 4.4 0.5 1.0 0.7 0.5 0.1 0.2 undertake such studies collaboratively R7xxx ~90 2.0 0.4 2.5 0.2 0.2 6.0 0.2 with organizations such as HVSC and other interested market leaders. and presorting would aid in maximizing have grain-refi ning elements such ACKNOWLEDGEMENTS the value of recycled aircraft aluminum.17 as Cr, Zr, and V present in small One technique would be to dismantle quantities (~0.1% or less), and the The authors acknowledge the partial aircraft into certain logical component potential buildup of such elements support of Qatar National Research groups, as these typically are made of in addition to Fe, Mg, and Si needs Fund through a grant from College of alloys of the same series. As examples, to be the subject of further study. the North AtlanticÐQatar, Doha, Qa- landing gears, engine nacelles, tail sec- This second factor will become tar. tions, and fl aps could be presorted, and increasingly important as newer wings separated from fuselages. Such aircraft alloys such as 2124, 2048, References separations may be desirable anyway to 7050, 7055, and 7085 enter the re- 1. Mark E. Schlesinger, Aluminum Recycling (Boca permit removal of non-aluminum com- cycle mix. Raton, FL: CRC Press, Taylor & Francis Group, 2007), p. 9. ponents before shredding. CONCLUSIONS 2. Aluminum Industry Vision – Sustainable Solutions Guidance in dismantling and pre- for a Dynamic World (Arlington, VA: The Aluminum sorting would be available from aircraft In several major product areas, op- Association, 2001), p. 17. 3. Roy F. Weston, Life Cycle Inventory Report for manufacturers who can identify the portunities to enhance the recycled met- the North American Aluminum Industry, AT-2 Report alloys used in various components. If/ al stream have been identifi ed. (Arlington, VA: The Aluminum Association, November when such information is not available, First, the potential secondary metal 1998). 4. Alcoa 2004 Sustainability Report (Pittsburgh, PA: it may be possible to identify the various supply in the automotive vehicle mar- Alcoa, 2004), p. 10; www.alcoa.com/global/en/about_ alloys with hand-held mobile spectrom- ket dwarfs all other product streams. Alcoa/sustainability _report_2004/images/sr_online_ eters. Non-aluminum components may The vehicle recycling process is com- fi nal.pdf. 5. S.K. Das, Light Metals 2006, ed. T.J. Galloway also be readily identifi ed using this tech- plex because of the large number of (Warrendale, PA: TMS, 2006), pp. 911–916. nique. If a useful level of discrimination both wrought and cast alloys involved. 6. John Green and Michael Skillingberg, Light Metal and separation of 2xxx and 7xxx series The primary opportunities to increase Age (August 2006), p. 33. 7. Fred W. Morgan and Margaret V. Hughes, JOM, 58 alloys is feasible, the compositions of the value of recycled automotive alumi- (8) (2006), p. 32. recycled metal are likely to represent num arises from the ability to disman- 8. Adam Gesing et al., Aluminum 2001, ed. S.K. Das, something like those shown in Table II. tle as many components of known alloy J.G. Kaufman, and T.J. Lienert (Warrendale, PA: TMS, 2001), pp. 31–42. Based upon this hypothesis, there classes (e.g., bumpers, wheels, body 9. Adam Gesing et al., Aluminum 2002, ed. S.K. Das would appear to be opportunities for sheet, castings) prior to shredding, and and M.H. Skillingberg (Warrendale, PA: TMS, 2002), direct reuse of the recycled metal in segregate them before remelting. With- pp. 3–17. 10. William T. Choate and John A.S. Green, Light new 2024-like and 7075-type alloys. out such strategies, the recycled metal Metals 2004, ed. A.T. Tabereaux (Warrendale, PA: Upon solution heat treatment and pre- stream will be limited to the production TMS, 2004), pp. 913–918. cipitation aging, the properties of these of casting alloys, much of it overseas. 11. Aluminum Statistical Review for 2005 (Arlington, VA: The Aluminum Association, 2005). remelt alloys are likely to be similar to Next, in the area of building and 12. S.K. Das, J.A.S. Green and J.G. Kaufman, JOM, 59 those of 2024 and 7075. Subject to more construction, the aluminum compo- (11) (2007), pp. 47–51. thorough performance evaluation, there nents of buildings being demolished 13. Collection of Aluminum from Buildings in Europe, Delft University of Technology (Brussels, Belgium: is reason to conclude that such metal are readily recoverable for recycling, European Aluminum Association, 2004), www.eaa might be utilized in non-fracture-criti- and can readily be sorted by extruded .net. cal aerospace components or in other or rolled products, effectively separat- 14. J. Randolph Kissell and Robert L. Ferry, Aluminum Structures—A Guide to Their Specifi cations and structural market application. If, on the ing Al-Mg-Si (6xxx) alloys from Al- Design (New York: John Wiley & Sons, Inc., 1995), p. other hand, 2xxx and 7xxx alloys cannot Mg (5xxx) series alloys. Remelting 33. be sorted before remelting, the combi- these as segregated lots will provide 15. J. Gilbert Kaufman, Introduction to Aluminum Alloys and Tempers (Materials Park, OH: ASM International, nation of Cu and Zn is likely to require compositions suitable for direct reuse 2000), pp. 87–118. some type of benefi ciation of the remelt in B&C applications. 16. S.K. Das and J. Green, Light Metals 2008, ed. before being very useful, even in non- Finally, if aircraft structures can be David H. DeYoung (Warrendale, PA: TMS, 2008), pp. 1113–1118. aerospace applications. dismantled and presorted by Al-Cu 17. S.K. Das and J. Gilbert Kaufman, Light Metals Two other factors should be noted at (2xxx) and Al-Zn (7xxx) alloys, the 2007, ed. Morten Sørlie (Warrendale, PA: TMS, 2007), this stage. segregated remelt compositions may pp. 1161–1165. ¥ First, at least small quantities of be directly reusable in non-fracture- Subodh K. Das, John A.S. Green, and J. Gilbert several other wrought aluminum al- critical components. Kaufman are with Phinix, LLC, P.O. Box 11668, loys like 2219 and 6061 and cast al- More in-depth analysis and mass fl ow Lexington, KY 40577-1668; Daryoush Emadi is loys like 201.0, A356.0, and A357.0 balances are needed to more precisely with Qatar University, Doha, Qatar; and M. Mahfoud is with the College of the North Atlantic–Qatar, may go into any such recycle mix. determine the likely compositions of re- Doha, Qatar. Dr. Das can be reached at skdas@ ¥ Second, aircraft alloys typically cycled metal from each of these product phinix.net.

26 www.tms.org/jom.html JOM • February 2010 Aluminum Overview Aluminum Industry and Climate Change—Assessment and Responses

Subodh K. Das and John A.S. Green

It is now possible to assess the im- the Aluminum Association when the study was the North American indus- pact of the production processes of alu- industry was challenged by the auto- try, since it included the chain of plants minum on the environment and to de- motive producers, under the United that supply the big automotive compa- scribe some of the ongoing responses States Automotive Materials Partner- nies, it was in fact global in nature, as and opportunities for improvement. ship (USAMP) initiative, to develop a it included data from 213 plants located This is compared with the benefi ts of life cycle inventory (LCI) for the metal. in Australia, Africa, Brazil, and Jamai- aluminum in transportation, where The results of the LCI were fi rst pub- ca as well as from many operations in the growing usage in various forms of lished in 1998,2 and now have been in- North America. transport due to its low density, high corporated into an ASM International The responsibility for continuing strength, and ability to be recycled Sourcebook.3 this database and extending it globally enables reduced mass, increased fuel While the domain of this initial to Asia and the Far East has now been effi ciency, reduced emissions and in- passed to the International Aluminum creased safety. It is the purpose of this Institute (IAI);3 see Chapter 3 for de- How would you… paper to compare and contrast the tails of the initial LCI study; see Chap- …describe the overall signifi cance emissions generated in the production of this paper? ter 4 for more recent IAI surveys and of aluminum with the benefi ts accruing Consistent with earlier predictions, information; see Chapters 5 and 6 for from its increased use in transporta- this assessment of the aluminum discussions of materials fl ow modeling tion. industry suggests that the reduction of aluminum based on this informa- of carbon dioxide emissions tion; see Chapter 7 for discussion about INTRODUCTION through the use of aluminum in transportation are growing at a rate the benefi cial impact of recycling; and Aluminum is a latecomer to the greater than the emission generated lastly see Chapter 10 and related ap- suite of industrial metals. It was only by increased production. Thus the pendices for information about carbon industry will have a negative carbon relatively recently, in 1886, when com- footprint likely by 2020. dioxide and other emissions. For cur- mercial application of the metal began rent information about industry prog- …describe this work to a 4 following the discovery of the Hall– materials science and engineering ress and goals, see the IAI Web site. Héroult reduction process. By contrast, professional with no experience in With this body of information, it is metals like copper, tin, lead, zinc, and your technical specialty? now possible to assess the impact of the iron have been known for centuries. This paper describes the actions production processes of aluminum on However, by virtue of its light weight being taken by the global industry the environment and to describe some to decrease its environmental and other unique properties, aluminum footprint such as process effi ciency of the ongoing responses and opportu- has now surpassed most metals in terms improvements and recycling and nities for improvement. This is com- of global production and is second only contrasts with emission savings pared with the benefi ts of aluminum to iron (steel). A survey of primary alu- through its use in transportation (i.e, in transportation, where the growing automotives). The paper outlines the minum smelters of the world lists a life cycle fl ow model being used to usage in various forms of transport due total global capacity of about 41 mil- assess the situation. to its low density, high strength, and lion metric tons, although some 7% of …describe this work to a ability to be recycled enables reduced this amount has been shut down on a layperson? mass, increased fuel effi ciency, reduced temporary basis.1 With this large global This paper assesses the response emissions, and increased safety. It is production, it is logical to ask about the of the aluminum industry to the purpose of this paper to compare environmental challenges and impact of the metal on the global envi- climate changes. Overall, it is and contrast the emissions generated ronment. suggested that the use of aluminum in the production of aluminum with the The industry fi rst began a compre- is benefi cial to the environment, benefi ts accruing from its increased use hensive accounting of the energy used, especially through its use in the in transportation. transportation industry, where its and the emissions generated, during all light weight makes vehicles more See the sidebar for background on facets of production in the mid-1990s. fuel effi cient. assessing aluminum production emis- This initial effort was undertaken by sions.

Vol. 62 No. 2 • JOM www.tms.org/jom.html 27 nology is being upgraded to the more More radical energy effi ciency im- FIVE INDUSTRY energy effi cient fl uid bed technology. provements have been limited by the RESPONSES In the area of smelter technology, the extremely corrosive and aggressive cell cell amperage is being increased and, environment and by the lack of materi- Process Effi ciency in development activities, cells to op- als to withstand exposure to the cryolite Improvements erate at 500,000 A are being explored. electrolyte at the operating temperature The process technology of the indus- Higher cell amperages proportionally of ~950C. Two signifi cant cell design try is continuing to evolve, especially in increase the productivity of the cell. improvements have been pursued in- the regions of low-cost energy. An ob- Along with increasing amperage, the termittently with limited success by vious means to further reduce the CO2 use of slotted anodes better enables the various companies over the past three footprint of the industry is to use en- escape of the CO2 gas from the spacing decades, but the costs and operat- ergy from the most effi cient generating between the electrodes, and improves ing lifetime of available materials has plants and to use a greater proportion operation of the cell. Also, better cell limited their economic effectiveness. of renewable energy. Recently, both Al- controls, the use of magnetic compen- These technologies are the use of wet- coa and Century Aluminum have estab- sation to stabilize cell operation, and table / drained cathodes and the use of lished smelters in Iceland, where hydro the adoption of alumina point feeders inert anodes. Both of these technolo- and geothermal power is available. Ad- all have helped to increase effi ciency of gies promise to improve effi ciency by ditional sources of hydro power are be- cell operations. Using these and other reducing the ohmic resistance within ing developed in China and Brazil and developments over the last 50 years, the the anode cathode distance (ACD) in this trend is expected to continue. average amount of electricity needed to the cell. The ACD is typically 4–5 cm With regard to existing plants in the make a pound of aluminum has been re- and constitutes about 40% of the over- United States and globally, plants are duced from ~12 kWh to ~7 kWh. These all cell resistance. For a more complete being upgraded with new technology. signifi cant improvements are expected discussion of these electrode systems For example, in some alumina refi ner- to continue to gradually improve cell and the complex issues involved see ies, the older rotary calcinations tech- effi ciency. Reference 5.

ASSESSMENTS OF ALUMINUM PRODUCTION EMISSIONS The energy to produce aluminum in the United States has been There are two signifi cant emissions from the fuel used in the reduced by 64% over the past 45 years. This reduction has come process. In alumina calcination, the fuel used is combusted to car- about by technical progress (22%) and by the growth of recycling bon dioxide. In anode preparation, the fuel and the excess carbo- (42%). Globally, these energy saving trends are similar, and in naceous material within the coke and pitch likewise is eventually

some countries the energy savings may be accentuated because converted to CO2. In fact, in most anode baking furnaces, the fuel newer technology smelters with better energy effi ciency are being residual in the pitch is fully utilized in the process to assist with built overseas. heating the baking furnace and minimize the need for additional

Despite this signifi cant progress, aluminum remains one of the fuel. Again, the process offgas is eventually lost as CO2. energy-intensive materials to produce. The U.S. aluminum indus- In the smelter cell itself, the cell reaction produces oxygen try directly consumes 42.3 × 109 kWh of electricity annually, or which reacts with the carbon anode and generates a mixture of

1.1% of all the electricity consumed by the residential, commer- mostly CO2 with a small amount of carbon monoxide. There is cial, and industrial sectors of the U.S. economy.5 also some additional burning of the anode materials, independent

The energy values discussed in this paper are tacit energy val- of the electrochemical reaction, which also yields CO2. All cell

ues; in other words they include the value for the fuel to create gases are captured and treated in a gas handling system but CO2 is the energy in the fi rst place as well as the value for the actual en- eventually released to the atmosphere. ergy used in the process itself. Many different fuels are used in the Other gases, called perfl uorocarbons (PFC), are released from overall production process throughout the world depending on the the cell during upset conditions when the supply of alumina is de- region—for an extensive discussion on fuel usage, see Reference pleted in the cell and an alternative anode reaction becomes domi-

6. In all cases, the industry has sought to use hydro power where nant. These gases are tetrafl uoromethane (CF4) and to a much less

available and the carbon footprint of hydro power is generally con- extent hexafl uoroethane (C2F6). These gases are signifi cant since sidered to be zero. Currently, the industry use of hydro power in they have considerable global-warming potential and are equiva-

the United States is ~40% and globally is ~50%. Nuclear power lent to 6,500 and 9,200 times that of CO2, respectively. In other

also has a low carbon footprint, but its use by the industry in the words, 1 kg of CF4 released to the atmosphere is equivalent in its

United States is less than 2%. warming capacity to 6,500 kg of CO2. On average, in a modern

Specifi cally, the most energy intensive phases of aluminum cell the PFC emissions account for 2.2 kg CO2 / kg Al (see Table production are the electrolysis or reduction step, followed by the D4 in Reference 3). anode preparation for the reduction process, and then the alumina The industry has been proactive in seeking a reduction of the production process. In total, it requires 60.5 kWh/kg to produce PFC emissions. In fact, while worldwide production of aluminum primary metal ingot whereas it only requires 2.8 kWh/kg to pro- has increased ~24% since 1990, the emissions of PFCs have de- duce secondary, or recycled metal. Thus, ~95% of the energy em- clined ~39% from the 1990 baseline. This is due to better cell con- bedded in the metal during the reduction process is saved when the trol and the introduction of point feeder systems that enable a more metal is recycled. Also, virtually the same proportion of process precise addition of the amount of alumina that is added to the cell. emissions is eliminated by recycling. This benefi cial trend is expected to continue.

28 www.tms.org/jom.html JOM • February 2010 The carbon cathode is made “wet- automotive aluminum). However, if a 7xxx alloy containing table” by the addition of small quanti- With regard to UBC material, the zinc and copper additions and normally ties of titanium diboride which is baked rate of recycling in the United States is used for high-strength aerospace appli- into the cathode. TiB2 is expensive and a lackluster 52% as compared to rates cations is used as a bumper in a vehicle is slowly consumed during the process of ~95% for Brazil and Norway. This application, the zinc and copper alloy which is a further drawback. However, means that the balance is being lost in additions enormously complicate the it is estimated that a wettable cathode landfi lls around the country. It has been recycling procedures and reduce the can improve the cell performance by estimated that there is an accumulated value of the metal. Different streams of ~15% and it is thought that several total of 20 million tons of UBC material recycled metal will be needed to fully companies are currently testing this in the United States. This equals the an- benefi t from the metal that is recov- technology. The combination of a wet- nual production of three new aluminum ered. table cathode, together with an inert smelters! This has led to the sugges- There is a need to make automotive anode, which would enable a decrease, tion to mine the landfi lls to recover the designers aware of the complexities of and more precise control of the ACD, is buried cans.8 An attractive alternative recycling and the impact that certain de- estimated to improve cell performance is not to bury the material in the fi rst signs may have on the quality and eco- by ~20%. Alcoa, Hydro, and Moltech place—possibly through the sorting of nomics of recycled metal. Also, there are companies known to be exploring municipal waste prior to land fi lling. In is a need to refi ne dismantling and pre- inert anode technology.7 this regard, it is noteworthy that Alcoa sorting procedures prior to shredding From the environmental viewpoint, has now set a goal and called upon the and to refi ne separation and sorting of the inert anode is the Holy Grail of the industry to raise the UBC recycling rate the non-magnetic metallic concentrate industry at present since its use would to 75% by the year 2015.9 after shredding. The recent demonstra- eliminate the need for carbon anodes For the future, the more important tion of laser based post-shredder sort- and the related anode baking complete- area for recycling will be automotive ing technology (e.g., LIBS) by the Hu- ly, and would generate oxygen, instead aluminum, as the volume of recycled ron Valley Steel Corporation11 has been of CO2, as the cell off gas. This would aluminum from automotive compo- a great tool to sort material, but as yet have a huge impact on the industry, nents exceeded the metal coming from it is just being used to benefi ciate 3xxx 10 eliminating more than 60% of its CO2 UBC for the fi rst time in 2005. With alloy feedstock even though in concept footprint. Of course, it would cost some the growth of automotive aluminum it could enable individual alloy separa- energy to make the inert anodes but this for light weighting, fuel effi ciency and tion. Equipment cost and process time is diffi cult to estimate since it is not safety considerations, this area is antic- limit the application. known whether the best electrode ma- ipated to become even more dominant, Further, there is a need to explore the terial will be ceramic, metallic or cer- especially with the recent increases in development of recycle-friendly alumi- met. Regardless, the benefi cial impact fuel prices (see the next section). num alloys for certain market segments would be huge. With regard to processing of auto- of the industry. Conceptually, these Finally, if these electrode materials motive wastes, the important recent alloys would have alloy additions and can be proven on an industrial scale, development is the introduction of the impurity contents that likely would fi t it is a relatively easy step to visualize industrial shredder. This has automated recycle metal streams and would not additional improvements in cell design the process of vehicle recycling and, require any, or at least a minimum, of from the exclusively horizontal elec- unlike the case of an individual beer or post processing for reuse. These issues trode assembly used in industry today beverage can, it has taken the decision are discussed at length in Reference 10. to a more compact, energy-dense, ver- away from the hands of individual con- Another way to improve the quality of tical electrode assembly. However, this sumers. Virtually all used vehicles are recycled metal is to employ a rudimen- is probably some two or three decades shredded, and the U.S. Environmental tary alloy separation during recycling. in the future. Thus, there are many op- Protection Agency (EPA) estimates For example, most aluminum used tions for future improvements, though that ~90% of all automotive aluminum in the building and construction mar- the best options still require extensive is now recovered and recycled. ket is either the 5xxx alloy, mostly in research and development. To capitalize on the benefi ts of the sheet form, or the 6xxx alloy, generally shredder, there is now a need to improve used as an extrusion. Accordingly, if Recycling and extend the dismantling and presort- the dismantlers segregate the different As stated previously, recycling of ing of vehicle components. A key issue shapes at the job site an effective alloy aluminum saves ~95% of the energy in recycling is the control of the alloy- separation can be achieved and the re- and emissions as compared to primary ing additions and impurities to retain sultant secondary melt quality will be production. While the industry always the value of the secondary metal. Some improved. These and other issues are has been proactive in emphasizing the alloys can be recycled better than oth- detailed in Reference 12. Lastly, for need for recycling (i.e., in the case of ers. For example, the 3xxx, 5xxx, and the most effi cient recycle process, it used beverage cans (UBC), the need 6xxx alloys generally used in automo- is important to adopt the best melting now is to reemphasize recycling of tive applications have similar alloying practice which limits the exposure of UBC but in addition to push for all additions of magnesium and silicon and the molten metal to oxygen and reduces other types of recycling, especially of can be readily commingled for melting. metal loss through dross formation.

Vol. 62 No. 2 • JOM www.tms.org/jom.html 29 is consumed to make 1 kg of aluminum. 350 5% CAGR Accordingly, a carbon trading scheme Europe 300 259 would strongly encourage development 250 196 200 to make the technical breakthrough to 150 112 achieve a viable inert anode material. 100 50 Success here would eliminate the need 0 for carbon anodes completely and have 1990 2000 2006 a huge impact on the greenhouse gas 3% CAGR (in pounds) 350 Japan Figure 1. Growth footprint of the industry for the reasons 300 251 in aluminum con- 250 212 cited above. tent in Japanese Another result expected from a car- 200 135 and European 150 bon trading scheme is that the recycling 100 light vehicles. 50 (Charts: Ducker of aluminum (secondary production) 0 Worldwide.) would be favored over primary produc- 1990 2000 2006 tion. This follows from the fact that only life cycle inventory and materials fl ow 5% of the energy is required for recy- Promote Aluminum Use in modeling of the industry. First, each cling as compared to primary produc- Transportation to Reduce pound of aluminum replacing two tion and no carbon anodes are involved Emissions pounds of traditional materials (iron in the recycling process. An additional The unique properties of aluminum or steel) in a vehicle can save a net 20 benefi t is that recycling facilities (re- make it an ideal material to reduce the pounds of CO2 emissions over the life- melters) are smaller facilities and only mass of transportation vehicles. Reduc- time of the typical vehicle. Second, a require about 10% of the capital costs of tion of mass directly converts to a sav- fuel savings of 6–8% can be gained for a new smelter installation. Thus, carbon ing of fuel. This applies to all forms of every 10% weight reduction resulting trading might favor more remelters con- transport such as aircraft, railcars, ships, in fewer greenhouse gas (GHG) emis- struction in consumption-rich countries and especially cars and trucks and bus- sions. like China and India where the supply of es. Less fuel usage in turn reduces CO2 Further, according to the EPA, nearly scrap is more likely to be available. It is emissions. 90% of automotive aluminum is recov- unlikely to modify production technol- Properties that are key are its light ered and recycled. It is estimated that ogy in the energy-rich areas of Iceland weight (roughly a third of the density of about 57% of the aluminum content in and the Middle East where scrap avail- steel), high strength, and corrosion re- North American vehicles was sourced ability is limited. sistance. Light weighting of the vehicle from recycled material. The alloy speci- ADDING IT ALL UP structure reduces fuel requirements, and fi cations for cast aluminum alloys, used enables bigger, lighter structures with for items like engine blocks and motor Figure 3 illustrates a material fl ow “crumple” zones and other features to housings, are quite tolerant of recycled model for the global aluminum industry improve passenger safety. material. For additional discussion, see throughout its life cycle that has been The growth of aluminum usage has Reference 13. developed by IAI for the year 2004— been ongoing for more than 30 years for additional details, see Reference 3, Carbon Trading and is a global trend (see Figure 1). Re- p. 95. The values cited are in millions of cent increases in the price of oil and gas Although the outlines of a carbon metric tons. and concerns for the carbon footprint of trading scheme are not yet developed The area of the circles illustrates the vehicles will almost certainly enhance and far from clear, such a system prob- relative volume of the fl ows. The total this trend. In North America in 2007, ably would have a signifi cant impact on products in use, an estimated 538 mil- the average light vehicle contained 327 the aluminum industry. For example, a lion metric tons, represent more than lb. of aluminum and there are no obvi- little more than 0.4 kg of carbon anode 70% of all of the aluminum that has ever ous factors, other than the metal’s price, to constrain this trend (see Figure 2). 2015: 374 lbs. Earlier problems with the metal’s 350 2007: 327 lbs. formability have been reduced as de- 300 signers have become familiar with alu- 250 Projected minum alloys and fabricators gain more 200 experience with forming it. Future high- 150 mileage vehicles will need to contain 100 large quantities of light metals such as 50 aluminum, magnesium, and titanium to attain their high-performance goals. 74 76 78 80 82 84 86 88 90 92 94 96 98 00 02 04 06 Two additional factors are important Calendar Year with regard to aluminum use in vehicles Figure 2. North American light vehicle content (in lbs.): trends and projection. (Ducker Worldwide) and directly follow from the extensive

30 www.tms.org/jom.html JOM • February 2010 Material Flow Metal Flow Bauxite3 153.7 Alumina3 57.3 Total Products Stored in Use Since 1888 538.5 Fabricated and Finished Finished Products Products Transport (Input) (Output) Buildings 28% Primary Ingots 61.8 60.5 36.3 31% Aluminum Automotive Used 30.2 Net Addition 2004 21.1 18% Figure 3. A material Remelted fl ow model for the Aluminum 31.6 Other Packaging 1% Old Other global aluminum in- Fabricator Applications3 Scrap Scrap2 16.5 11% dustry throughout its Recycled 1.1 7.4 Engineering Aluminum and Cable 29% life cycle, developed Traded New Traded New by IAI for the year 15.1 Scrap1 1.3 Scrap3 7.7 2004. For additional details, see Refer- ence 3, p. 95. Values Bauxite Residues 56.7 cited are in millions of and Water 39.7 Metal Losses 1.3 Not Recycled in 20043 3,4 Under Investigation4 3 metric tons.

been produced. With this database it is to aluminum use in transportation will 7. Ioan Galasiu, Rodica Galasiu, and Jomar Thons- possible to assess current energy and surpass the global industry’s production tad, Inert Anodes for Aluminum Electrolysis (Düssel- dorf, Germany: Aluminium-Verlag, 2007). emissions intensity and project ahead emissions by 2020 (Reference 3, see 8. S. Das et al., “Recovering Aluminum from Used for future years the energy use and GHG Chapters 5&6). The 2007 Alcoa Annual Beverage Cans —The Dilemma of 900,000 Annual footprint of the industry. Report contains the statement “the use Tons” (Paper presented at TMS 2007 Annual Meet- ing, Orlando, FL, February 25–March 1, 2007). Based on the growth of recycling and of aluminum in planes, trains, and au- 9. Alcoa Annual Report 2007 (Pittsburgh, PA: Alcoa, the reducing energy intensity of the pri- tomobiles is projected to make the en- 2007), p. 18. mary production and the signifi cantly tire aluminum industry greenhouse gas 10. S.K. Das, J.A.S. Green, and J. Gilbert Kaufman, JOM, 59 (11) (2007), p. 47. lower PFC emissions, the CO2 emissions neutral by the year 2025.” Whatever the 11. Adam Gesing et al., Aluminum 2002, ed. S.K. Das footprint of the industry is decreasing precise date, this is indeed an encourag- and M.H. Skillingberg (Warrendale, PA: TMS, 2002), on a per ton basis, although the industry ing sign! pp. 3–15. 12. S. Das, J. Gilbert Kaufman, and J.A.S. Green, continues to expand production. On the References “Aluminum Recycling—An Integrated, Industry-Wide other hand, the savings of CO2 emissions Approach,” in this issue, pp. 23–26. from the transportation use of alumi- 1. R.P. Pawlek, Light Metal Age, 66 (2) (2008), p. 6. 13. “Aluminum in Autos: Driving toward a Sustainable 2. Life Cycle Inventory Report for the North American Environment” (Arlington, VA: The Aluminum Asso- num are continuing to grow as more and Aluminum Industry, Report AT-2 (Arlington, VA: Alumi- ciation, Aluminum Transportation Group Media Kit), more metal is used and the resultant fuel num Association, 1998). www.autoaluminum.org. savings accumulate and the amount of 3. John A.S. Green, editor, Aluminum Recycling and Processing for Energy Conservation and Sustain- recycled material is increased. Assum- ability, ASM Sourcebook (Materials Park, OH: ASM ing the global industry can be updated International, 2007). Subodh K. Das is CEO and Founder of Phinix, LLC, to the best available technology as of 4. Internatioinal Aluminum Institute, London, www P.O. Box 11668, Lexington, Kentucky, 40577-1668; .worldaluminium.org. and John A.S. Green, consultant, is retired from The 2003, the model indicates that the fuel 5. In Ref. 3, Chapter 10, p. 158. Aluminum Association. Dr. Das can be reached at effi ciency and emissions savings due 6. In Ref. 3, pp. 71, 77, 160, 232, etc. [email protected].

Visit the Knowledge Resource Center to reserve your copy today! Converter and Fire Refi ning Practices by Alistair Ross, Tony Warner, and Ken Scholey, editors These proceedings will provide readers with an update on research, development, and operating practice in the fi elds of converting, fi re refi ning, and casting in the base metal pyrometallurgy industry. Materials and supply vendors, service pro- viders, and researchers and plant operators in the pyrometallurgy industry will fi nd this book to be a valuable resource. From the 2005 TMS Annual Meeting held in San Francisco, California, February 13–17, 2005. TMS Member price $121; Non-member price $173; TMS Student Member price $96

To order these or related publications, contact TMS: E-mail [email protected] • Phone (724) 776-9000, ext. 256 • Fax (724) 776-3770 Aluminum Research Summary Adaptive Control of Feed in the Hall– Héroult Cell using a Neural Network

K.D. Boadu and F.K. Omani

A linear neural network is proposed High concentrations of alumina lead slope calculations to the estimation for estimating alumina concentration to sludge formation which lower metal problem has proved inadequate in the in an aluminum reduction cell. Bath production.7–11 Therefore feed control face of a search for higher current ef- resistance/alumina concentration data strategies aim at maintaining the con- fi ciency. This is due to the noise in the from a simulated 140 kA Center-Break centration of alumina in the bath with- resistance data that often causes false Hall–Héroult cell were used as input in the limited range of 2.5 to 6 wt.%,12 or belated anode effect alarms and vectors to train a two-layer neural net- to avoid underfeeding or overfeeding. thus causes overfeeding or underfeed- work constructed with six constraints The application of simple resistance ing problems. More robust methods and six degrees of freedom. Results have, however, demanded impractical from simulated and real data tests us- computing power from cell microcon- How would you… ing the derived estimation algorithm trollers. …describe the overall signifi cance are presented. Also, neural network is of this paper? The extended Kalman fi lter solution 5,6 compared with extended Kalman fi lter The overall signifi cance of this applied to the estimation problem and shown to have a superior perfor- paper is to increase the current, has been seen to have three major mance in the estimation problem. Fi- energy, and production effi ciencies problems: nally, the paper claims robustness for of an aluminum reduction cell. 1. Convergence, stability, and opti- It simultaneously reduces gas the neural algorithm against changes emissions, and for that matter, mality of the estimator cannot be in resistance due to cell events like environmental pollution by a guaranteed due to the nonlinear tapping and anode change. reduction cell. nature of the relationship between …describe this work to a resistance and concentration.5 INTRODUCTION materials science and engineering 2. The algorithm is dependent on professional with no experience in In recent years several research pa- your technical specialty? regular estimates of anode move- pers have been published on the con- The key to maximizing the ments, to be compensated for in- trol of alumina concentration in the effi ciencies of the Hall–Héroult dependently, in order to provide Hall–Héroult cell. Resistance slope process for aluminum reduction is stable estimates.1 calculations have been used with con- a good prediction of the alumina 3. The continuous relinearization of concentration. This paper proposes siderable success in the estimation of the use of a neural network as the system at each sampling in- alumina concentration in the bath.1,2 a model-free algorithm for the terval makes it diffi cult to obtain In the search for improved accuracy in prediction problem. It abandons the a steady Kalman gain. Thus in the prediction of concentration, other usual model-based approaches due implementation, the Kalman gain to the time-series, polynomial nature methods have been used and reported. of the resistance-concentration curve has to be adapted on-line, giving These include the use of the ARMAX in the reduction cell. rise to a more complicated algo- 3 4 model, Runge-Kutta methods, and …describe this work to a rithm. Consequently, the need to the extended Kalman fi lter.5,6 layperson? reset the covariance matrix after The main focus of such research is Aluminum reduction cell control every feed leads to the Kalman to improve on the current and energy computers need to keep a careful fi lter detuning and giving rise to balance between under- and over- effi ciencies of the Hall–Héroult pro- feeding a cell with alumina, the initial poor estimates, before sub- cess. The literature abounds with the primary input in the production of sequent convergence to more ac- relationship between the alumina con- aluminum. The aim is to maximize curate estimates (see Figure 9). centration in an electrolytic cell and effi ciency and reduce cell conditions In this paper we present a neural which throw more than permitted current and energy effi ciencies. Low gases out to pollute the atmosphere. network approach to the estimation levels of alumina concentration often This paper proposes neural networks problem. A set of the fi ve most re- lead to an increase in the anode over- as a better method for the control cent measurements of cell resistance, computers in predicting alumina potential which increase anode effect concentration and maintain better and the previous estimate of alumina frequency and thus lower the energy control over alumina feed. concentration, are presented to the and production effi ciencies of the cell. network as input vectors for each mea-

32 www.tms.org/jom.html JOM • February 2010 sured alumina concentration (provided as the trainer signal). The neural net- 19.6

work learns the functional relationship μΩ ) between resistance and concentration. 19.4 After the off-line training, the network is easily implemented on-line, using the 19.2 parameters obtained through training to Figure 2. Bath resistance/ estimate the concentration of alumina. Bath Resistance ( 19 concentration curve. In general, the approaches adopted 1.5 2 2.5 3 3.5 4 so far to solve the estimation problem Alumina Concentration (wt.%) have depended on scientifi c rationalism to look for a tractable function relating bath resistance to the alumina concen- networks are constructed with many work output, Xt. The linear error is then ε tration, through the use of models. “But simple nonlinear summary junctions given by t = dt – Xt. For the next itera- how close can a low-order polynomial connected together by connections tion, an adaptation algorithm adjusts W ε 2 come to the time-series footprints left of varying strength to approximate in a way that minimizes t . Learning by a chaotic system?.”13 unknown functions with raw sample stops when the sum of the squared er- We therefore wish to propose the ap- data.14 rors is less than an error goal15,16 (see proach of scientifi c empiricism (neural Figure 1). The Adaline networks, in this case) to solving this Neurocomputing is currently in use estimation problem. In this approach, The ADALINE (adaptive linear el- in the industry for solving problems in we let the resistance data tell their own ement) is a neural network which lin- signal processing, speech recognition, story via model-free algorithms. Our early combines its input to derive an visual perception, adaptive control, and simulated and real data test results show output. robotics. that the trained network has a superior In training the ADALINE, an input For this study we used supervised T performance to the extended Kalman pattern vector Rt = [ro, r1, r2, r3, ..., rn] learning as a stochastic approximation fi lter. and a desired response (trainer signal), of the error surface of the unknown d , are presented to the network at each resistance/concentration function. We NEURAL NETWORK t iteration. Then the components of the applied the Widrow–Hoff learning rule THEORY α input vector, Rt, are weighted by the ( -LMS) to an adaptation of the ADA-

Neural networks are model-free esti- weight vector Wt = [wo, wl, w2, w3, LINE which embodies the minimal T mators which employ numerical algo- ..., wn] . The weighted inputs are then disturbance principle. The unknown re- T → rithms to transduce numerical inputs to summed, the inner product RtWt , to sistance/concentration function f: X numerical outputs. Fashioned after the produce a linear input for the output Y is estimated from observed random

“cognitive computations” of the human layer of the network, which is a transfer vector samples (x1, y1), . . . , (xm, ym) by brain, albeit a gross abstraction, neural function that fi nally produces the net- minimizing an unknown expected error functional.14–16 THE WIDROW–HOFF DELTA Bias Input Input Vector Weight Vector RULE (α-LMS) r0 = +1 Rt Wt The α-LMS was used as the weight w Bias Weight w1 0 adaptation algorithm due to the untrac- r1 table nature of the problem of predict- w2 ing alumina concentration in a chaotic r2 Linear Output system. The weight update equation w3 Xt r3 Σ F can be written as:

Transfer Function ttR WWt1  t 2 (1) R wn t r n Where α is the network learning rate. It can be shown that the error is reduced ΔW ε − t t by the factor α for each adaptation LMS Σ Error α Rt Algorithm + cycle, thus making control stability and speed of convergence during learn- dt ing.15 Desired Response (Trainer Signal) THE REDUCTION PROCESS

Figure 1. The adaptive linear element (ADALINE). The nerve center of aluminum reduc- tion is the electrolytic cell. The primary

Vol. 62 No. 2 • JOM www.tms.org/jom.html 33 1 reaction in the cell can be represented by: Al O (soln) + 3/2C(s) 0 2 3 = 2Al(l) + 3/2CO2(g) (2) This reduction is carried out in cryo- −1 Figure 3. Partial derivative lite bath using carbon electrodes, at a of resistance with respect to concentration/concentration temperature around 960C and at an −2 curve. effi ciency ranging between 85 and 95% δ( Resistance) / δ (Concentration) 0 12345 of Faraday’s theoretical production rate Alumina Concentration (wt.%) of aluminum.5,7,9 The main loss in effi - ciency is caused by the reoxidation of 5 the aluminum11 through the reaction: Al(soln) + 3/2CO (g) 4 2 = 1/2Al2O3(soln) + 3/2CO(g) (3)

3 The process is quite chaotic because while the primary reaction is continu- Figure 4. Bath concentration/ ous, the removal of the aluminum pro- feed time curve. 2 Alumina Concentration (wt.%) Alumina Concentration 0 20 40 60 80 100 120 140 duced (tapping) is discrete. Other noise Time (min.) inducing events in the course of the re- action include anode consumption and 4 change, unstationary bath volume, in- terpolar distance and temperature, and unscheduled addition of alumina. 3 Resistance Concentration Curves 2 Recent studies (on both point-feeder Simulated Estimated and center-break cells) on the shape 17

Alumina Concentration (wt.%) Alumina Concentration of the resistance-concentration curve 1 0 50 100 150 200 typify the nonlinearities in the ohmic Time (min.) resistance reported in the literature. See Figure 5. Simulated concentration and neural network estimated concentration Figures 2, 3, and 4.

curves (input pattern: Rt, . . ., Rt–4). It is clear from Figure 2 that the ohmic resistance does not identify the alumina concentration. However, a par- Bias Input Resistance Vector Weight Vector tial derivative of the resistance w.r.t. the r0 = +1 Rt Wt alumina concentration (Figure 3) does.5 w0 Bias Weight (This is the reason for the choice of the w1 r1 ADALINE since it obtains an implicit w2 Estimated Concentration numerical differentiation of the resis- r2 of Alumina tance at the fi rst layer synaptic junc- w3 Xt r3 Σ tion.) The characteristic behavior of alumina concentration over time is also w4 Purelin r4 shown in Figure 4. w5 r5 Rate of Alumina Dissolution w6 r6 Alumina dissolution in cryolite melt Xt−1 has also been the subject of various re- search work in the industry. The main ΔW ε − t t factors that impact the dissolution rate Widrow-Hoff Σ R Estimation Error t Delta Rule + include the alumina concentration, bath temperature, feeding method,18 and cell d t operating conditions.19 The rate of dis- Measured Concentration (Trainer Signal) solution is reported as 0.027 wt.%/min. for the center-break cells studied in Ref- Figure 6. Optimal neural network learning machine. erence 17, and 0.0625 wt.%/min. for the point-feeder cells in the same study. The

34 www.tms.org/jom.html JOM • February 2010 rate of dissolution was needed for the present study in order to build a simula- 5 Feed tor to test the neural-based feed strategy Real Estimated developed. To get this value, the concen- tration-time curve (see Figures 7 and 8) Figure 8. Real concentra- tion and neural network derived from the real cell data collected estimated concentration from 2 center-break cells were differen- Anode Effect curves (input pattern: Rt, . . ., R , X ). tiated. This yielded a dissolution rate of 0 t–4 t–1

Alumina Concentration (wt.%) Alumina Concentration 0 20 40 60 80 100 120 0.0253 wt.%/min. which was used for Time Since Last Feed (min.) the simulation in this paper. 5 The Learning Process Feed Real Estimated To train the neural network to learn the resistance/concentration relation- ship, a two-layer Widrow–Hoff super- Figure 9. Real concen- vised learning network was built. Data tration and extended from a simulated 140 kA center-break Kalman fi lter estimated Anode Effect concentration curves. 6 0 cell was used for the training. 0 20 40 60 80 100 120 140 Resistance data was regressed as (wt.%) Alumina Concentration Time Since Last Feed (min.) the input pattern vector using the fi ve most recent resistances (R). The Wid- row–Hoff delta rule (α-LMS) was used using the fi ve most recent resistances THE ESTIMATION to generate the weight vector (W). A (R , …, R ) as the input. The network’s t t-4 ALGORITHM network bias was allowed (w0t) with an performance looked good (Figure 5) input of unity (r0t). The purelin transfer but came up with an unacceptable sum Over 1,000 iterations were run function16 was used, and for each itera- squared error of 9.6317 (against a goal through the network using a 5×134 tion a trainer signal (dt) was provided of 0.5) over a training cycle of 1,000 bath resistance data input matrix and a as the measured alumina concentration iterations. An auto-regressive-moving- 1×134 measured alumina concentration scalar at time t. The network output, average (ARMA)-type input pattern vector (trainer signals), with each itera- T F(RtWt ), was then the estimated con- was therefore considered next (Rt, …, tion changing the network weights to centration of alumina, given the resis- Rt–4, Xt–l). Then the performance of the improve the estimation. Network learn- tance vector R. network signifi cantly improved (Figure ing ceased at the following output pa- The estimation error is calculated as 7), with a fi nal sum squared error of rameters: Weight Vector W = [–0.2564,

εt which is used via the Widrow–Hoff 0.8770. Performance, however, started –0.2046, –0.0537, +0.0919, +0.3402, delta rule to adjust W for the next itera- deteriorating on the next ARMA-type +0.9646]; Network Bias = +1.6799; Fi- tion (see Figure 1). input of Rt, …, Rt–4, Xt–l, Xt–2. The opti- nal sum squared error = 0.8770. There- mal training input pattern was therefore fore the algorithm for estimating the SELECTING THE OPTIMAL selected as R , …, R , X giving the concentration of alumina (X ) in a cell, NETWORK ARCHITECTURE t t–4 t–l t fi nal architecture (Figure 6) of: given the 5 most recent resistances (Rt) The basic concept adopted in this pa- is: X = F(R W T + Bias) (4) per for the training of the network was t t t n5 a moving average input, to enhance ef- XRW t6kk   fective resistance-concentration pattern where Rt = Rt, Rt-l, Rt-2, Rt-3, Rt-4, Xt-l, and k1 recognition. Initial simulations were run W = W , W , W , W , W , W t 1 2 3 4 5 6 Xt1 W n1 1.6793 f (5) δ 5 where f = rate of alumina dissolu- Simulated Estimated tion. The neural algorithm defi ned in 4 Equation 5 was tested on a 140 kA center-break simulated and real cell 3 data. Results are shown in Figures 7 and 8. (The real cell data used for the 2 test was collected with anode move- ments frozen.) Alumina Concentration (wt.%) Alumina Concentration 1 An extended Kalman fi lter algo- 0 50 100 150 200 rithm6 was tested on the same set of Time (min.) data used for the neural network test Figure 7. Simulated concentration and neural network estimated concentration curves (input pattern: R , . . ., R , X ). in Figure 8. Performance of the neu- t t–4 t–1 ral network (Figure 8) was superior

Vol. 62 No. 2 • JOM www.tms.org/jom.html 35 no small way. In a bid to keep compu- Simulated Tap/Anode Change tations simple, the massively parallel 20 topology of neural networks was not used. With more computing power, μΩ ) 19.5 a fully connected single neuron, or more parallel neurons may be tried for still better performance in predict- 19 ing alumina concentration in the cell Anode for higher current and production ef- Bath Resistance ( Tap Tap Change 18.5 fi ciencies. 0 50 100 150 200 Time (min.) References 1. M. Farrow, Journal of Metals, 36 (11) (1984), pp. Figure 10. Simulated resistance and resistance vectors during tap/anode change simula- 33–34. tion. 2. Y. Macaudiere, Light Metals 1988, ed. L.G. Boxall (Warrendale, PA: TMS, 1988), pp. 607–612. Simulated Estimated 3. T. Moen, J. Aalbu, and P. Borg, Light Metals 1985, 5 ed. H.O. Bohner (Warrendale, PA: TMS, 1985), pp. 459–469. 4. Laszlo Tikasz et al., Light Metals 1988, ed. L.G. Boxall (Warrendale, PA: TMS, 1988), pp. 583–588. 5. Einar Arne Sörheim and Peter Borg, Light Met- als 1989, ed. P.G. Campbell (Warrendale, PA: TMS, 1989), pp. 379–384. 6. Joan S. McKenna, F.K. Omani, and Thomas Nyadziehe, JOM, 45 (11) (1993), pp. 44–47. 7. K. Grjotheim and B.J. Welch, Aluminium Smelter Anode Tap Tap Change Technology (Düsseldorf, Germany: Aluminium-Ver- Alumina Concentration (wt.%) Alumina Concentration lag, 1988). 0 50 100 150 200 8. K. Grjotheim et al., Aluminium Electrolysis (Düs- Time (min.) seldorf, Germany: Aluminium-Verlag, 1982). 9. R.A. Lewis, “Technical Fundamentals of the Alu- Figure 11. Simulated concentration and neural network estimated concentration curves minum Reduction Cell Process” (Foothill Ranch, CA: (tapping and anode change simulated). Kaiser Aluminum and Chemical Corporation, 1994). 10. Claude A. Wilson and Alton T. Tabereaux, Light Metals 1983, ed. E.M. Adkins (Warrendale, PA: TMS, 1983), pp. 479–493. to that of the extended Kalman fi lter shows the corresponding performance 11. B. Lillebuen and Th. Mellerud, in Ref. 3, pp. 637– (Figure 9). Late convergence of the graph of the neural network algorithm, 646. extended Kalman fi lter algorithm after demonstrating the robustness of the al- 12. K.R. Robilliard and B. Rolofs, in Ref. 5, pp. 269– 273. each feed makes it unsuitable for use gorithm during cell events like tapping 13. Bart Kosko, Neural Networks for Signal Process- on short feed cycle cells (e.g., point and anode change. ing (Englewood Cliffs, NJ: Prentice-Hall, 1992). feeders). 14. Bart Kosko, Neural Networks and Fuzzy Systems CONCLUSION (Englewood Cliffs, NJ: Prentice-Hall, 1992). TEST FOR ROBUSTNESS 15. Clifford Lau and Bernard Widrow, editors, Special Algorithms for estimating alumina Issue on Neural Networks I (Piscataway, NJ: The In- Finally, the neural algorithm was concentration in the Hall–Héroult cell stitute of Electrical and Electronic Engineers (IEEE), September 1990). tested for robustness during a simula- have been saddled with the search for 16. Howard Demuth and Mark Beale, Neural Network tion of anode change and tapping by a resistance/concentration model. This TOOLBOX (Natick, MA: The MathWorks, Inc., 1992). random increases in the cell resistance paper has shown that the model-free 17. E.R. Owusu-Sechere, “Characteristics of Cell Re- sistance and Alumina Concentration in Bath (Mead & over random time ranges in the feed generalization mechanism employed Tacoma Prebake Cells” (Foothill Ranch, CA: Kaiser cycle. The estimated concentration of in neural networks for pattern recog- Aluminum and Chemical Corporation, DTC, 1994). alumina continued to be correct during nition can well be applied to the un- 18. H. Maeda, S. Matsui, and A. Era, in Ref. 3, pp. 763–780. these simulations. This performance certain and imprecise environment in 19. G.I. Kuschel and B.J. Welch, in Ref. 5, pp. 299– of the algorithm was in line with sci- the cell, through the use of a black- 305. entifi c expectation because of the box approximator from input to out- K.D. Boadu, formerly an automated systems pro- network’s numerical differentiation put. The estimation algorithm is very grammer at Kaiser Aluminum & Chemical Cor- of the resistance w.r.t. concentration simple and yet robust, and can be used poration–Valco Works, Tema, Ghana, is currently the Chief Executive Offi cer of Arrow Network which linearizes the resistance input effectively in a feed control program Systems, a wireless telecommunications com- before the concentration is estimated. to keep the concentration of alumina pany operating in West Africa. F.K. Omani is the Figure 10 shows the changes in bath within preferred limits. The sense of Head of the Research, Consultancy & ICT Depart- ment of The Regional Maritime University, Accra, resistance during the simulated tap- mystery and beauty that characterize Ghana. Dr. Omani can be reached at fkomani@ ping/anode change period. Figure 11 neurocomputing overwhelmed us in hotmail.com.

36 www.tms.org/jom.html JOM • February 2010 Aluminum Research Summary The Hot Formability of an Al–Cu–Mg–Fe–Ni Forging Disk

M. Aghaie-Khafri and A. Zargaran

The high temperature formability of limit of conventional 2000 alloys. The AA2618-T61 forged disk was studied How would you… addition of a small amount of Fe and by means of tensile test over tempera- …describe the overall signifi cance Ni enhances the microstructural stabil- of this paper? tures and strain rates ranging from 100 ity of the alloy for applications involv- to 400ºC and 3 × 10–5 –3 × 10–3 s–1, re- The high temperature formability of AA2618-T61 forged disk was ing high-temperature exposures up to 6 spectively. The constitutive equations studied by means of tensile test over 300C. The presence of stable inter- of the material were calculated based different temperatures and strain metallic particles, such as the Al9FeNi on an Arrhenius-type equation and the rates. The main softening mechanism particles, helps to control grain size of the AA2618-T61 forged disk ductility of the material was evaluated deformed at high Zener–Hollomon and impede dislocation motion. The considering elongation and percent (Z) parameter value (100–200ºC, AA2618 alloy is widely used for the reduction of area. The results showed s–1) is dynamic recovery. Dynamic production of forged components such that both kinds of softening mecha- recrystallization is the primary as engine parts for both automotive and softening mechanism when the alloy 7,8 nisms, dynamic recovery and dynamic deformed at low Z value (350–400ºC, aircraft applications. The convention- recrystallization, occurred during high s–1). Dynamic recovery and dynamic al processing route consists of heating temperature deformation of the alloy. recrystallization play a role at the and maintaining the billet at the hot Strain rate sensitivity of the material same time at medium Z value (250ºC, working temperature for a suffi ciently s–1). was evaluated in all the deformation long time followed by forging. Later, conditions and the obtained values …describe this work to a the component is solution-treated and were used to calculate the apparent ac- materials science and engineering artifi cially aged to produce the required professional with no experience in 9 tivation energy. your technical specialty? balance of mechanical properties. The ageing sequence of this alloy INTRODUCTION AA2618, a heat-treatable AlCu– Mg–Fe–Ni forging alloy, was and several aspects of its mechanical Formability of a material is the ex- developed originally to raise the response at high temperature are well tent to which it can be deformed in a temperature limit of conventional known.10–12 The warm formability of a 2000 alloys. This work evaluates the particular process before the onset high temperature tensile formability solution-treated AA2618 aluminum al- of failure. There are many variables of AA2618-T61 forging disk over a loy was investigated by means of tor- that infl uence sheet metal formability, temperature range of 100–400ºC. sion tests at temperatures between 150 which can be broadly classifi ed as pro- The infl uence of deformation and 300 C. Dynamic precipitation took temperature, strain rate and strain  1 cess variables and material variables. degree, and softening mechanisms place during deformation raising the Process variables determine the nature such as dynamic recovery and fl ow stress and lowering ductility. The of external loading on the work piece, dynamic recrystallization on the fl ow stresses were substantially higher hot formability of AA2618-T61 are whereas material variables determine studied. than those observed by testing the same the kind of response the material exhib- alloy in the as-extruded state.9 its to that loading. Previous investiga- …describe this work to a The hot and warm formability of layperson? tions have shown that hot formability AA2618 aluminum alloy, in the as- AA2618 is one of the aluminum of aluminum alloys are determined by alloys originally developed for solutioned condition, was also inves- process parameters, such as deforma- aircraft components and automobile tigated. The effect of the precipitation tion temperature, strain rate, and strain industries. This alloy was of second-phase particles, occurring degree. Softening mechanisms such as successfully used as the primary during deformation, on the fl ow curve structure of the supersonic Concorde dynamic recovery and dynamic recrys- airplane. Fabrication of components shape and on the stress level was evalu- tallization are two important material is dependent on the mechanical ated.13 Other studies have analyzed the variables controlling the hot formabil- properties of material such as effect of high-temperature creep on the ity of aluminum alloys.2–5 strength, ductility, and formability. precipitate structure. Several concur- The aim of the present work is to AA2618 a heat-treatable Al–Cu– study the mechanical properties of ring phenomena, such as the loss of Mg–Fe–Ni forging alloy was devel- this alloy at high temperature. coherency of precipitates, the change oped originally to raise the temperature in particle size and distribution, and

Vol. 62 No. 2 • JOM www.tms.org/jom.html 37 the transformation of the metastable into the stable phase, can be strain- enhanced, and lead to the progres- sive softening of the structure during creep.14 To the best of authors’ knowl- edge, there have been few reports on the high-temperature tensile properties of the AA2618-T61 forging disk. The aim of the present work is to study the infl uence of temperature on the tensile formability of AA2618-T61 forging disk over a temperature range of 100– 400C. See the sidebar for experimental pro- cedures. RESULTS AND DISCUSSION

Stress-strain Curve and Figure 1. Nominal strain-stress Microstructure curves of the AA2618-T61 forged disk at different strain Figures 1 and 2 show the nominal rates and temperatures. stress-strain curves of the forging disk at various temperatures and strain rates which show the effects of deformation temperature and strain rate on the fl ow behavior of the material. At the low- the curves increase up to the peak fol- more and more pronounced, and it temperature regime the stress-strain lowed by a moderate decrease in nomi- transforms in a continuous reduction in curves monotonically increase up to nal stress. At the high-temperature stress, approaching zero. fracture; at intermediate temperatures regime the decrease in stress becomes The variations of fl ow stress with temperature for different strain rates EXPERIMENTAL PROCEDURE are shown in Figure 3. It is clear that fl ow stress decreased with increasing The aluminum wrought alloy AA2618-T61 was delivered as forged disk with chemical composition of 0.3Si-1.06Fe-1.92Cu-1.18Mg-0.9Ni-0.07 Ti and Al balance (all values in temperature. However, the slope of the wt.%). The optical micrograph of the material was shown in Figure A, which consists of curves increased by increasing tem- perature and tends to a constant value uniformly distributed fi ne Al2CuMg intermetallic particles and larger precipitates embed- ded in aluminum solid solution phase. In the present investigation the energy dispersive at 400C and strain rate of 3 × 10–4 s–1. x-ray spectroscopy (EDX) point analysis of the particles confi rmed that the large precipi- The dependence of the fl ow stress on

tates consist of Al9FeNi intermetallic. temperature can be evaluated by Seeger The specimens used in the tensile tests were cut in radial direction from the center theory of fl ow stress.15 The stress re- of the forging and had a diameter of 4 mm and a length of 25 mm in the gauge-length quired for deformation can be divided section. Hot tensile testing was performed in a servo-hydraulic machine, Instron 8800, into two parts: σ*, which is dependent equipped with a three-zone split furnace. The tests were conducted in laboratory air envi- on the temperature and σ , which is the ronment over the temperature and initial strain rate range of 100–400 C and 3 × 10–5 –3 G  athermal component of the fl ow stress: × 10–3 s–1, respectively. The temperature was controlled with the aid of a thermocouple fi xed on the test specimen’s surface. Maximum temperature variation was well within σ = σ* + σ (1) 2C, of the set-point temperature over the entire duration of the test. Before each test, the G specimen was maintained or soaked at the test temperature for 15 min, so as to achieve stability with the environment. It can be seen that the athermal compo- σ ≈ Samples for optical microscopy were sec- nent of stress is G 32 MPa at a strain tioned from the gauge area, mounted, me- rate of 3 × 10–4 s–1. chanically polished according to the standard Deformation at 100–250C results in procedure and then etched with a solution of a microstructure with elongated grains, 10 ml Hf, 5 ml HNO and 85 ml H O. 3 2 which are typical of dynamic recovery. One such microstructure is shown in Figure 4, which corresponds to a speci- Figure A. Optical micrograph of the micro- men deformed at 200 C and strain rate structure of AA2618-T61. The microstructure  of 3 × 10–3 s–1. Figure 5 shows that consists of uniformly distributed fi ne Al2CuMg precipitates and larger intermetallic particles. the elongated grains decompose dur- 100 μm ing high temperature deformation and some fi ne and equiaxed grains are ob-

38 www.tms.org/jom.html JOM • February 2010 partially recrystallized regions becomes unstable causing severe necking and subsequent ductile fracture observed at 300C. The steep increase in ductility at higher temperatures (350–400C) implies that a complete softening due Figure 2. Nominal strain-stress to dynamic recrystallization has taken curves of the AA2618-T61 forged place, and the recovered structure, disk at strain rates of 3 × 10–3 s–1 which is expected to be the cause and different temperatures. of the observed premature failure, has been removed. Furthermore, at higher temperature metastable or stable precipitates coarsened and the served along the big angle boundaries at is the product of a complex interaction precipitates become less and less 250C with the strain rates of 3 × 10–3 among the kinetics of generation effective in pinning dislocations; new s−1, illustrating that dynamic recrystal- and annihilation of dislocations, dislocations are generated and recover lization occurs during tensile deforma- nucleation of new phases and zones, in larger subgrains.9 It can be concluded tion at the deformation conditions. The and growth of precipitates.9 Freshly that dynamic precipitation retards microstructure of forged disk deformed nucleated fi ne particles very effectively at 300C and 3 × 10–4 s–1 is shown in pin dislocations, thus reducing Figure 6, which exhibits typical dy- recovery and lowering ductility. The namic recrystallization features and the interaction between particles and absence of annealing twins. dislocations involves pinning and diffusion along dislocations. Such High-temperature Ductility effect, if temperature is suffi ciently The ductility of the forged disk in high, provides suffi cient driving hot and warm conditions, measured as force and accelerates the transition total elongation is shown in Figure 7. toward the dynamic recrystallization. It can be observed ductility of the ma- Consequently, deformation in the μ terial increased over the temperature 150 m Figure 4. Microstructure of the sample range of 100–200C mainly due to dy- deformed at temperature of 200C and namic recovery process. However, after strain rate of 3 × 10–3 s–1. an initial increase, the total elongation to fracture gradually decreased over a temperature range of 200–300C fol- lowed by an abrupt increase at 400C. Furthermore, in the case of higher strain rate, deformation proceeds to larger strains. Figure 8 shows the ten- sion specimens deformed at 200 and 300C. Although the elongation of the specimen deformed at 300C de- creased, the reduction of area severely 200 μm increased with temperature (Figure 9). Figure 5. Microstructure of the sample deformed at temperature of 250C and In general, in several alloys dynamic strain rate of 3 × 10–3 s–1. recovery and dynamic recrystallization are the main softening mechanisms; however, in aluminum alloys softening mechanisms are characterized by the ability of dislocation to cross-glide or to climb easily, under annihilative at- traction.9 The observed loss of ductil- ity at 250C is an expression of the low level of dynamic recovery which is a consequence of the dynamic nature of the precipitation processes occurring 200 μm during high-temperature deformation. Figure 3. Variation of fl ow stress at strain Figure 6. Microstructure of the sample of 0.02 with temperature at different deformed at temperature of 300C and In this case, the microstructure does strain rates. strain rate of 3 × 10–4 s−1. not depend only on applied stress, but

Vol. 62 No. 2 • JOM www.tms.org/jom.html 39 from 0.007 at 200C to 0.2 at 350C. and substructural changes occurring The phenomenon can be attributed to during deformation.21 The unexpected the dynamic recrystallization at higher value of activation energy in the temperatures that produces higher present investigation is a consequence strain rate sensitivity.18 of the precipitation process occurring during high-temperature deformation. Constitutive Equation In this case, the microstructure does The hot and warm forming behavior not depend only on applied stress, but of AA2618 aluminum alloys can be is the product of a complex interaction Figure 7. Variation of total elongation as a modeled by correlating fl ow stress to among the kinetics of generation and function of temperature at different strain strain rate and temperature according to annihilation of dislocations, dynamic rates. the well known constitutive equation:19 recovery and dynamic recrystallization, nucleation and growth of precipitates. mmmQ It can be concluded that dynamic Bexp &   BZ (3) RT precipitation restricts dislocation and grain boundary mobility and delays μ Q 3521 m where Zexp&  is recovery. This leads to high values of RT apparent activation energy related to the Zener–Hollomon parameter the deformation mechanisms taking representing the temperature modifi ed place at high temperature. strain rate, Q is the activation energy The conventional procedure for μ a 3,000 m related to the deformation mechanism calculating constitutive equations for taking place in the material during high temperature deformation based the process, R is the universal gas 2412 μm constant, T is the absolute temperature, & is the strain rate, m is strain rate sensitivity coeffi cient, and B is a material parameter. From Equation 3, the apparent activation energy Q, may be defi ned by:

μ b 3,000 m Rln Figure 8. Tensile test specimens deformed Q  (4) m1T at 200C and 300C, respectively.   &, Figure 9. Variation of reduction of area Plots of lnσ versus 1/T for a strain as a function of temperature at strain –5 −1 recovery and stimulates subsequent of 0.02 and different strain rates rates of 3 × 10 s . dynamic recrystallization at the higher are shown in Figure 12. It is clear temperatures. The latter at suffi cient the two lines that were fi tted to the high temperatures (350–400C) experimental data represent two promotes ductility. different softening mechanisms at low Figure 10 shows the variation of fl ow and high temperatures. stress with strain rate on a ln–ln plot at The calculated apparent activation different temperatures and at plastic energy decreased with temperature strains of 0.02. The strain rate-fl ow mainly due to greater values of strain stress profi le is clearly linear between rate sensitivity at higher temperatures, 150 and 400°C. Strain rate sensitivity shown in Figure 13. The mean value of Figure 10. Variation of ln(σ) versus ln( & ) at a strain of 0.02 and different tempera- coeffi cient (m) as an important material the apparent activation energy at high tures. property can affect formability of temperature region (247 ± 14 kJ/mol) is metals and can be defi ned as:16,17 greater than the value of one calculated for self-diffusion in aluminum alloys ln m  (2) (143.4 kJ/mol).20 It is worth noting ln & T, the activation energy determined from The calculated strain rate sensitivity Equation 4 is ‘apparent’ because the (m) at constant true plastic strain and microstructural and substructural temperature based on Equation 1 is changes accompanying deformation shown in Figure 11. The m value is infl uence Q through their effect on σ and nearly constant up to 200 C and shows m. Thus, given the same deformation  Figure 11. Plot of strain rate sensitivity an abrupt increase at 250–300C. As it is rate controlling mechanism, Q will versus temperature. observed, the mean value of m changes vary depending on the microstructural

40 www.tms.org/jom.html JOM • February 2010 on Equation 3 assumes that, in steady ing conclusions can be made. state, the activation energy and m value Strain rate sensitivity is believed to of samples tested at different tempera- infl uence the post-uniform deformation tures appear approximately identical of the material at high temperature de- for a fi xed stress. Concerning these formation conditions. Maximum elon- assumptions, if Equation 3 adequately gation and percent reduction of area represents the constitutive behavior of about 18% and 98 %, respectively, was AA2618-T61 aluminum alloy, then a obtained at strain rate of 3 × 10–3 s–1 and log-log plot of the fl ow stress versus 400C, where the strain rate sensitivity the Zener–Hollomon parameter should (m) was at maximum of 0.22. Figure 13. Variation of apparent activation yield a straight line with a slope of m. The main softening mechanism of energy as a function of temperature at The mean value of apparent activation the AA2618-T61 forged disk deformed different strain rates. energy for high temperature region at high Z value (100–200C, 3 × 10–3 (247 ± 14 kJ/mol) was used to calcu- s–1) is dynamic recovery. Dynamic re- late the Zener–Hollomon parameter crystallization is the primary softening and is plotted versus the fl ow stress in mechanism when the alloy deformed at Figure 14. This plot shows a good cor- low Z value (350–400C, 3 × 10–5 s−1). relation between the fl ow stress and the Dynamic recovery and dynamic recrys- Zener–Hollomon parameter which im- tallization play the role at the same time plies that the fl ow stress follows the ex- at medium Z value (250C, 3 × 10–5 – 3 pected trend with respect to strain rate × 10–3 s–1). and temperature. The observed loss of ductility over Figure 14. Variation of fl ow stress versus a temperature range of 200–250C is Zener–Holomon parameter at a strain CONCLUSIONS of 0.02 and temperature range of 300– believed to be an expression of the low 400C. Based on the present study on level of dynamic recovery which is a the high temperature formability of consequence of the dynamic precipita- 64–71. AA2618-T61 forging disk, the follow- tion processes occurring during high- 6. I. Özbek, Mater. Charac., 58 (2007), pp. 312–317. temperature deformation. Subsequent 7. J. Wang et al., Mater. Charac., 59 (2008), pp. 965– 968. decrease in ductility at 250–300C is 8. J.C. Williams and E.A. Starke, Acta Mater., 51 attributed to the severe strain localiza- (2003), pp. 5775–5799. tion at partially recrystallized regions. 9. F. Bardi, M. Cabibbo, and S. Spigarelli, Mater. Sci. Eng., A334 (2002), pp. 87–95. Flow stress was found to increase 10. ASM Handbook vol. 2 (Materials Park, OH: monotonically at low temperature, American Society for Metals, 1990). while at the higher temperatures of 11. K. Yu et al., Mater. Sci. Eng., A368 (2004), pp. 88–93. the investigated range the stress–strain 12. D. Bobrow, A. Arbel, and D. Eliezer, J. Mater. Sci., curves exhibited a peak, followed by 26 (1991), pp. 2045–2049. softening. 13. P. Cavaliere, J. Light Metals, 2 (2002), pp. 247– 252. The mean apparent activation energy 14. R. Singer and W. Blum, Z. Metallkde., 68 (1977), for tensile deformation of the alloy is p. 328. 247 ± 14 kJ/mol, which is higher than 15. A. Seeger, Work Hardening, vol. 46, ed. J.P. Hirth and J. Weertman (New York: TMS-AIME, 1968), pp. that one calculated for self-diffusion in 27–63. aluminum alloys. The appreciate differ- 16. M. Aghaie-Khafri and R. Mahmudi, JOM, 50 (11) ence between the two activation energy (1998), pp. 50–52. 17. M. Aghaie-Khafri and N. Golarzy, Mater. Sci. Eng., values is a consequence of the dynamic A486 (2008), pp.641–647. precipitation process occurring during 18. F. Moneteillet and J.J. Jonas, Metall. Mater. Trans., high-temperature deformation. 27A (1996), pp. 3346–3348. 19. M. Aghaie-Khafri and N. Golarzy, J. Mater. Sci., 43 (2008), pp. 3717–3724. References 20. O.D. Sherby, R.J. Klundt, and A.K. Miller, Metall. Trans., 8A (1977), p. 843. 1. G.E. Dieter, editor, Workability Testing Techniques 21. C. Girish Shastry et al., Mat. Sci. Eng., A465 (Metals Park, OH: American Society for Metals, (2007), pp. 109–115. 1984). 2. H. Yamagata, Scr. Met. Mater., 27 (2) (1992), pp. 201–203. M. Aghaie-Khafri, Associate Professor, and A. 3. B. Ren and J.G. Morris, Met. Mater. Trans., 26A Zargaran, graduate student, are with the Faculty of (1995), pp. 31–40. Mechanical Engineering, K.N. Toosi University of Figure 12. Plot of fl ow stress versus 1/T 4. H.J. McQueen, Mater. Sci. Eng., A387–389 (2004), Technology, P.O. Box 19395-1999, Tehran, Iran. Prof. at a strain of 0.02. pp. 203–208. Aghaie-Khafri can be reached at +98 09123088389; 5. H.E. Hu et al., Mat. Sci. Eng., A488 (2008), pp. fax +98 21 88674748; e-mail [email protected].

Vol. 62 No. 2 • JOM www.tms.org/jom.html 41 Aluminum Research Summary The Use of Friction Stir Welding for Manufacturing Small-scale Structures

T. Hirata and K. Higashi

Friction stir welding (FSW) is a most applications of FSW are limited How would you… relatively new joining process and is to major enterprises, and practical use being used commercially in several …describe the overall signifi cance has not been advanced in small and of this paper? industry sectors. In small and medium medium enterprises. This is because In this paper, several approaches enterprises, however, this novel tech- for developing practical friction stir additional value is high even if an only nology has not been applied despite welding (FSW) processes in medium linear weld is possible when large-scale its remarkable advantages because of and small enterprises are introduced. structures are manufactured by major the drawbacks of FSW. A database has Friction stir welding has attracted enterprises. However, the cost cannot a great deal of attention as a high- been assembled and drawbacks have quality welding technology and is be justifi ed when small and medium been analyzed then solved appropri- currently being used commercially enterprises make small products using ately in the Industry-Government-Aca- in several industry sectors. However, FSW. In addition, it is very diffi cult for practical use has not advanced demia Collaboration Project. As an in small and medium enterprises small and medium enterprises to con- outgrowth, the optimum and individual because of some technical problems. duct needed three-dimensional (3-D) know-how of practical FSW technolo- This article describes experiments welding by FSW. gy could be transferred to medium and to determine how practical FSW The City Area Project—an industry- technology could be transferred to small enterprises. those medium and small enterprises. government-academia collaboration project at Osaka-East-Urban Area in INTRODUCTION …describe this work to a materials science and engineering Osaka Prefecture, by the Ministry of Friction stir welding (FSW) is a professional with no experience in Education, Culture, Sports, Science relatively new joining process. It was your technical specialty? and Technology Japan (MEXT) was developed initially for aluminum al- Friction stir welding (FSW) is a established in 2004. This project main- loys, by The Welding Institute (TWI) relatively new joining process. The ly conducted research and development 1 process advantages result from the of the United Kingdom. In FSW, a fact that the FSW process takes place on FSW to promote product upgrades cylindrical, shouldered tool with a in the solid phase below the melting and strengthen the price-competitive profi led probe is rotated and slowly point of the materials to be joined. edge of the major industries such as plunged into the joint line between two However, this novel technology metalworking machinery in Osaka Pre- cannot be applied in small and pieces of sheet or plate material, which medium enterprises because it is very fecture. The schematic of the procedure are butted together. Frictional heat is diffi cult to make small products by used to carry out this project is shown generated between the wear-resistant using FSW. A new FSW system has in Figure 1. been developed in this work so that welding tool and the material of the practical FSW technology could be A new FSW machine was developed workpieces. This heat causes the latter transferred to the medium and small during this project with two types of to soften without reaching the melting enterprises. heads: a high-output linear head with point and allows traversing of the tool …describe this work to a two-axes movement (X and Y axes) along the weld line. It leaves a solid layperson? and a 3-D drive head with fi ve-axes phase bond between the two pieces. Welding technology is important for movement (X, Y, Z, a, and c axes). The advantages result from the fact manufacturing, with fusion welding Therefore, this machine is capable of the most popular technology in metal that the FSW process takes place in the welding. Friction stir welding is a welding a material in a wide range of solid phase below the melting point of relatively new joining process that, shapes. the materials to be joined. In addition, like fusion welding, does not melt the Several approaches were carried out the process offers low distortion and materials. The process advantages to develop a practical FSW process for are low distortion, energy effi ciency, 2,3 is energy effi cient, environmentally environmental friendliness, and medium and small enterprises. The friendly, and versatile. versatility. In this work, research and drawbacks that occurred during the Friction stir welding is currently be- development on FSW was carried out FSW experiments were analyzed and ing used commercially in several sec- to promote the products upgrade and solved appropriately. As a result, the strengthen the price-competitive edge tors, such as the railway, automotive, in the medium and small enterprises. optimum and individual know-how on and aerospace industries. However, practical FSW could be transferred to

42 www.tms.org/jom.html JOM • February 2010 medium and small enterprises. Images of the supply device of metallic parts and the water-cooling plate used in FSW are shown in Figure 2. It was pos- sible to produce in less time and at low cost. In addition, the newly developed tool is shown in Figure 3. This tool is highly suitable for use in a 3-D FSW process. In this paper, a database con- structed in this project is introduced, and several approaches for developing practical FSW processes for medium and small enterprises are introduced. a See the sidebar for details on the FSW database. DEVELOPMENT APPROACH

Supply Device of Metallic Parts The supply device of metallic parts and a schematic of the chassis of the Figure 2. Images of (a) supply device are shown in Figure 4. In the device of metallic parts and 2,3 conventional method, the top and bot- (b) water-cooling plate. b tom of the column or the column and the base were connected by bolts. It is parts can be manufactured using FSW. The material used was a 6063-T5 possible to reduce the number of mod- In addition, it is expected to be able to aluminum alloy sheet with a thickness ules, duration of the process, and cost manufacture products with a good de- of 10 mm. Single-pass friction stir butt if the supply device made of metallic sign. welds were produced using an FSW tool with a shoulder diameter of 20 mm and a probe length of 9.8 mm. The cross section of the weld joint under the optimum FSW condition is shown in Figure 5. The weld did not exhibit cracks and porosity, indicating that the weld was of good quality. Moreover, the joint effi ciency under the optimum FSW condition was approximately 70%. This value was considered to be very high because the 6063-T5 alumi- num alloy was a heat-treated aluminum alloy. During FSW, it is necessary to hold the material fi rmly in place. In par- ticular, it is important to fi rmly hold the material along the direction verti- cal to the welding direction from the

Figure 3. The tool developed in this proj- Figure 1. A schematic of the procedure used to carry out this project. ect.3

Vol. 62 No. 2 • JOM www.tms.org/jom.html 43 FRICTION STIR WELDING DATABASE Friction stir welding (FSW) is consid- its alloys have a high oxygen affi nity, they pin tool was developed using relatively low- ered by many to be the most signifi cant cannot be welded easily; therefore, tungsten cost metallic materials such as heat-resistant development in metal joining in the last inert gas welding (TIG) is usually used for alloys, with or without surface treatment.30 decade. In addition, friction stir process- welding titanium alloys.29 FSW is suitable An image of pure titanium welded using ing (FSP) has been developed based on the for titanium and its alloys. In this study, a FSW is shown in Figure B. The maximum basic principles of FSW. This technique is used to induce material modifi cation. To Table A. Parameters of Aluminum and Magnesium Alloys used in This Study along put the FSW technique to practical use in with the Welding Conditions of the FSW (FSP) Techniques4–25 medium and small enterprises, it is impor- tant to analyze various parameters. In this Tool Welding Conditions project, we investigated numerous param- Shoulder Probe Rotational Welding eters such as welding parameters, weld Material Diameter Diameter Speed Speed joint properties, and operation accuracy. (Thickness) (mm) (mm) (rpm) (mm/min) Ref. Aluminum and Magnesium Alloys 1050-H24 (5t) 15 6 600 ~ 2,000 100 ~ 800 4 5052-O (3t) 9 3 2,000 ~ 4,000 1,000 ~ 2,000 5 Table A lists the parameters of alumi- 5052-O (3t) 15 3 4,000 500 ~ 1,000 5 num and magnesium alloys used in the 5083-O (1t) 9 3 250 ~ 1,600 130 ~ 350 6 study along with the welding conditions of 5083-O (3t) 12 4 500 ~ 1,000 100 ~ 200 7 the FSW (FSP) techniques.4Ð25 It has been 2017-T351 (5t) 15 6 1,500 25 ~ 200 8 reported that 2000 and 7000 aluminum al- 2024-T351 (6t) 16 6 850 120 9 loys have been joined using FSW because 2024-T6 (6t) 16 6 850 120 9 this technique has the ability to join ma- 2024-T351 (7t) 23 8.2 215 ~ 360 77 ~ 267 10 terials that cannot be fusion welded. Fric- 6061-T6 (6t) 15 ~ 21 6 1,200 75 11 tion stir welding conditions infl uence the 6061-T6 (3.2t) 12.7 3.18 1,000 76 12 properties of a weld joint. The relationship 6056 (4t) 14 6 500 ~ 1,000 40 ~ 80 13 between the heat input parameter RtD3 / V 6082-T6 (3t) 10 3 715 ~ 1,500 71.5 ~ 200 14 and the grain size d in the stir zone of a 7075-T6 (7t) 20 6.5 700 160 15 5083 aluminum alloy is shown in Figure 7075-T6 (6t) 18 6 350 120 16 A.7 Rt is the rotational speed; V, the weld- 7449-T3 (22t) 23 8 350 175 ~ 350 17 ing speed; and D, the shoulder diameter 7449-T79 (22t) 23 8 350 175 ~350 17 of the tool. It is evident that RtD3 / V is AZ31B (0.64t) 19 6.3 800 ~ 1,000 60 18 closely related to d. Further, a refi nement AZ31B (4t) 12 4 375 ~ 2,250 20 ~ 375 19 in the grain size of the weld zone improved AZ31B (1.9t) 13 5 1,400 300 20 the formability of the weld joint.7 There- AZ31-H24 (4.95t) 19.05 6.35 500 ~ 1,000 60 ~ 240 21 fore, RtD3 / V was considered to be a key AZ61 (6.3t) 18 8 1,220 90 22 parameter in the manufacture of industrial AM60B (2t) 19 6.3 2,000 120 23 machines using FSW. AZ91D (2t) 8 3 800 ~ 1,800 90 ~ 750 24 Titanium LA141 (2t) 12 4 1,500 200 25 Titanium and its alloys, which have high specifi c strength, excellent corrosion resistance, and good biocompatibility, are widely used in aircrafts and artifi cial bones.28 Unfortunately, since titanium and

ab

Figure A. Relationship between RtD3 / V and d in the stir zone of 5083 aluminum alloy.7

cd Figure C. Allowable operational deviation in (a) parallel inclination against direction of welding, Figure B. Pure titanium welded using (b) vertical inclination against direction of welding, (c) butt line/tool-travel line deviation, and (d) FSW. gap spacing.3,31,32 The material used was 5083 aluminum alloy.

44 www.tms.org/jom.html JOM • February 2010 joint effi ciency was about 90%, and it was infl uenced the deformation behavior of the materials at a low strain rate. However, at a suggested that it could be welded using the FSPed Zn-22Al alloy. Figure Fa and b shows high strain rate, the value of the elongation low-cost pin tool. the variations in fl ow stress and elongation to failure of the FSPed material was higher to failure of the Zn-22Al alloy as a function than that of the base material. Therefore, Operational Accuracy of strain, respectively. The value of the elon- FSP is considered to be suitable for improv- For three-dimensional (3-D) FSW, it is gation to failure was almost the same for all ing the mechanical properties of materials. most important to improve operational ac- curacy.31,32 Therefore, the objective of this study is to elucidate the following: allow- able deviation (deviation is defi ned as the misalignment of the butt and tool-travel lines); allowable gap spacing (gap spacing is defi ned as the separation between the butt faces); allowable tool angle (tool angle is de- fi ned as angle at which the tool is inclined in the parallel and perpendicular directions to the welding direction). The allowable operational deviations are ab shown in Figure C. An FSW tool with a conventional shape, made of die steel, was used in this study. The butt line/tool-travel line was allowed to deviate up to half the probe diameter distance when the butt line was placed on the advancing side of the tool rotation. So we have the one-fourth of probe diameter allowance on both sides by setting the butt line from the probe center in a quar- ter of the probe diameter to advancing side. The allowable gap spacing was three-eighths of the probe diameter when the butt line was cd placed at the center of the tool-traveling line. Figure D. Images of samples produced by using this control system. In the case of the inclination angle, the plus and minus angles indicate the angles of ad- vancing and retreating and of the right and left, respectively. The allowable tool incli- nation in the direction perpendicular to the welding direction is 2¡ in the right and left directions. The allowable tool inclination in the direction parallel to the welding direc- tion is 3¡ of the advancing angle and 0.4¡ of the retreating angle, because material loss ab was signifi cantly high and defects occurred Figure E. Inverse pole on the surface of the weld line when the re- fi gures of (a) Al-rich phase treating angle was greater than 0.4. These of base material, (b) Al- results were important as they were used for rich phase of friction-stir- the development of a 3-D control system.33 processed material, (c) Zn-rich phase of base The samples produced by using this control material, and (d) Zn-rich system are shown in Figure D. In this way, phase of friction-stir- cd this machine is capable of applying the weld processed material.37 to the materials in a wide range of the shapes by using this control system. Friction Stir Processing The microstructures of several materi- als change to fi ne-grained structures due to dynamic recrystallization during FSP. In this study, many FSPed materials were in- vestigated.34Ð37 For example, FSP was used to effectively produce a Zn-22wt.%Al alloy with a very fi ne-grained microstructure. The initial texture of the Zn-rich phase became random due to FSP, as shown in Figure E. ab This microstructural evolution due to FSP Figure F. Variations in (a) fl ow stress and (b) elongation-to-failure as a function of strain rate.37

Vol. 62 No. 2 • JOM www.tms.org/jom.html 45 ability and high cost. If the waterway is formed in the aluminum block and the cover by aluminum board is welded at the edge of the waterway on the ba- sis of a lap FSW design, a reduction in both the duration of the working pro- cess and production cost is expected. Materials used to develop the wa- ter-cooling plate were 5052-H112 and Figure 4. (a) Image of 5052-H34 aluminum alloy sheets with supply device of metallic thicknesses of 30 mm and 2 mm, re- parts and (b) schematic spectively. The diameters of the probe ab of the chassis of the device. and the shoulder of the FSW tool were 4 mm and 12 mm, respectively. The length of the probe was 2.9 mm. The viewpoint of measurement accuracy. addition, considering the fact that dis- macrosection of the water-cooling plate When applying FSW for manufactur- tortion is caused after FSW, welded joined using lap welds under optimum ing a large-sized product, large-scale blanks and jigs were created, as shown FSW condition is shown in Figure 9. In structures are often linear welded for in Figure 6. Columns A1 and A2 were the fi gure, the left and right sides repre- long distances. Therefore, both sides of fi xed by blocks, and Column A1 was sent the retreating (RS) and advancing the welded blank are often parallel, and pushed by an air cylinder. Since one sides (AS) of the weld, respectively. it is comparatively easy to fi rmly hold side of Column A2 was not parallel to The weld did not exhibit cracks and the material along the direction verti- the welding direction, it was diffi cult cal to the welding direction. However, to push Column A2 from this side us- when applying FSW for manufactur- ing the air cylinder. Therefore, it was ing a small-sized product in small and assumed that Column A2 was held in medium enterprises, it is predicted that a fi xed position using a fi xing pin. Fur- both sides of the welded blank will of- ther, it was easy to push the columns ten not be parallel. In other words, it in a direction perpendicular to the is necessary to investigate the possibil- welding direction, and the product was ity of using another method for fi rmly manufactured with high-dimensional holding the welded blank with a spe- accuracy. cial shape in place. Therefore, another The base and column were welded welding method was developed. using the same approach as were the From the design considerations, the columns, as shown in Figure 7. Blocks Figure 7. Method using FSW technique for welding the base and column.2,3 right and left side of the column of the were used for fi xing the column; the welded surface must be placed inside base was pushed by the air cylinder. the supply device of metallic parts. In In addition, the column was fi xed by using the fi xing pin. Welding the base and column using the above-mentioned method resulted in the successful man- ufacture of the supply device of metal- lic parts at low cost, as shown in Figure 2a. Therefore, a technique employing the fi xed pin was used in manufactur- ing structures with specifi c design con- siderations, and it was expected that Figure 5. Cross section of the weld joint this technique should be applied to under optimum FSW conditions.2,3 manufacture other products. Figure 8. Conventional method for de- Water-cooling Plate veloping water-cooling plate. The conventional method to develop a water-cooling plate made of alumi- num alloy is shown in Figure 8. A gun drill was used to create holes in the aluminum block. Next, the holes were Figure 9. Macrosection of the plate Figure 6. Method using FSW technique blocked in order to form a waterway. joined using lap welds under optimum for welding columns.2,3 Therefore, the development of the wa- FSW condition.3 ter-cooling plate resulted in low work-

46 www.tms.org/jom.html JOM • February 2010 voids, indicating that the weld was of fectively. In addition, the water-cooling good quality. However, the joint inter- plate successfully passed the air and Cutting Work face remained horizontal on the RS of pressure tightness tests. Therefore, it the weld. was possible to develop the water-cool- Specimens were extracted from ing plate with lap weld and fi ve axes a welds, and shear tensile tests were per- movement using a rapid and inexpen- formed on them under two welding con- sive manufacturing process. FSW ditions. Rotational speed of 1,200 rpm Tool for 3-D FSW and welding speeds of 400 mm/min. and 100 mm/min. were maintained. One of the features of FSW is that b The shape of the specimen is shown in the tool used for welding is kept in a Figure 12. Position of tools used in (a) Figure 10. Because the joint interface fi xed position. The position of the tools working machine and (b) FSW machine. remains horizontal on the RS of the used in the working and the FSW ma- weld, welding was performed in direc- chines is shown in Figure 12. During Welding Direction tions A and B. Since all the specimens cutting, the tools are kept in a fi xed po- Interference fractured at the upper board, it implied sition. On the other hand, during FSW, Gap that weld joints were of very good the tools are inclined at the advancing quality. However, the strength of the angle along the direction vertical to the joints varied, and the maximum shear planar direction of the material. This Figure 13. Geometric interference be- strength was obtained at the rotational is because it is easy to supply a steady tween the work machine and the tool. speed of 1200 rpm, welding speed of heat input during FSW. In addition, it 400 mm/min., and in the welding direc- is possible to inhibit the occurrence of tion B. burrs, and a tool with a complex shape is 3-D FSW, and the development of a We challenged to develop a simple not required. The welding quality is ex- tool which can weld in a wide range of water-cooling plate. The schematic of pected to deteriorate remarkably when inclination angles has been expected. the simple water-cooling plate is shown FSW is performed using the conven- Thus, the new tool required for 3-D in Figure 11. This plate was developed tional FSW tool, which is not inclined FSW has been developed and is intro- under the optimum FSW condition. at the advancing angle. However, it is duced in this section. Further, the welding route with circu- diffi cult to maintain the inclination of As shown in Figure C in the side- lar arc was partially employed from the tool at the advancing angle during bar, a high joint strength was obtained the viewpoint of practical use. There- 3-D FSW. Moreover, it is necessary to along the range of the advancing angle; fore, the 3-D drive head with fi ve axes consider the geometric interference be- however, the joint strength decreased movement was used in the welding of tween the FSW machine and the tool, with an increase in the retreating angle the water-cooling plate. The image of as shown in Figure 13. Therefore, not when the conventional FSW tool was the water-cooling plate is shown in Fig- only the advancing angle but also the used. The joint strength decreased sig- ure 2b. The plate could be welded ef- retreating angle is very important for nifi cantly when the retreating angle was greater than 0.4, because mate- rial loss was very high and defects oc- curred on the surface of the weld line. Therefore, the development of a new tool was needed so FSW could be per- formed when the retreating angle was greater than 0.5. We have paid attention to a part of the shoulder for developing the new Figure 10. Shape of speci- tool, and the design of the new tool has men before performing the been discussed. The newly developed shear tensile test. FSW tool is shown in Figure 3. The probe of the new tool was almost the same as that of the conventional tool. On the other hand, a unique design was adopted to develop the shoulder of the new tool, which differed from that of the conventional tool. FSW was per- formed using the new tool. The image of the tensile specimen extracted from Figure 11. Schematic of the simple water-cooling the welds, and the mechanical proper- plate.2,3 ties of the weld joint are shown in Fig- ure 14. 6063-T6 aluminum alloy was

Vol. 62 No. 2 • JOM www.tms.org/jom.html 47 450–454. a 5. Y.S. Sato et al., Mater. Sci. Eng. A, 369 (2004), pp. 138–143. 6. Y.S. Sato et al., Mat. Sci. Eng. A, 354 (2003), pp. 298–305. 7. T. Hirata et al., Mater. Sci. Eng. A, 456 (2007), pp. 344–349. 8. H.J. Liu et al., J. Mater. Process. Tech., 142 (2003), pp. 692–696. 9. C. Genevois et al., Acta Mater., 53 (2005), pp. 2447–2458. 10. M.A. Sutton et al., Mater. Sci. Eng. A, 354 (2003), pp. 6–16. 11. K. Elangovan and V. Balasubramanian, Mater. De- sign, 29 (2008), pp. 362–373. 12. D. C. Hofmann and K. S. Vecchio, Mater. Sci. Eng. A, 402 (2005), pp. 234–241. 13. P. Cavaliere et al., J. Mater. Process. Tech., 180 (2006), pp. 263–270. 14. L. Fratini and G. Buffa, Int. J. Machine Tools & Manufacture, 45 (2005), pp. 1188–1194. Figure 14 (a) Image of the tensile 15. P. Cavaliere and A. Squillace, Mater. Character., specimen extracted from welds 55 (2005), pp. 136–142. and (b) tensile strength of the 16. J.-Q. Su, T.W. Nelson, and C.J. Sterling, Mater. Sci. weld joint.3 Eng. A, 405 (2005), pp. 277–286. 17. M. Dumont et al., Acta Mater., 54 (2006), pp. b 4793–4801. 18. J.A. Esparza et al., J. Mater. Sci. Lett., 21 (2002), pp. 917–920. 19. W. Xunhong and W. Kuaishe, Mater. Sci. Eng. A, used as the material. It was possible to Practical Use of Friction Stir Technol- 431 (2006), pp. 114–117. perform welding in the range of the ad- ogy” (established by the Osaka In- 20. M.B. Kannan et al., Mater. Sci. Eng. A, 460-461 vancing angle without the occurrence dustrial Promotion Organization) was (2007), pp. 243–250. 21. N. Afrin et al., Mater. Sci. Eng. A, 472 (2008), pp. of burrs. In addition, it was possible formed as a successor to the Urban 179–186. to perform welding in the range of the Area Industry-Government-Academia 22. S.H.C. Park, Y.S. Sato, and H. Kokawa, Scr. Mater., retreating angle range without pen- Collaboration Project (2004Ð2007). 49 (2003), pp. 161–166. 23. J.A. Esparza, W.C. Davis, and L.E. Murr, J. Mater. etrating the plate material, and a defect We aim at enhancing the applicability Sci., 38 (2003), pp. 941–952. was not observed on the surface. Good of the FSW technique via this associa- 24. S.H.C. Park, Y.S. Sato, and H. Kokawa, J. Mater. welding quality was obtained until the tion in the future. Sci., 38 (2003), pp. 4379–4383. 25. M. Tsujikawa et al., Mater. Trans., 47 (2006), pp. retreating angle of 1 . The position of  ACKNOWLEDGEMENT 1077–1081. the tool used in FSW has an infl uence 26. K. Colligan et al., Proc. 3rd Int. FSW Symp. (Cam- on various aspects. For example, the This work is a part of the Osaka bridge, U.K.: The Welding Institute, 2001). 27. J.K. Kristensen et al., Proc. 5th Int. FSW Symp. design of the 3-D FSW becomes sim- East Urban Area Industry-Govern- (Cambridge, U.K.: The Welding Institute, 2004). ple if the tool inclination is not consid- ment-Academia Collaboration Project 28. G. Luetjering and J.C. Williams, Titanium (Berlin: ered. In addition, the load on the ma- on FSW “CITY AREA” sponsored by Springer, 2003). 29. M.J. Donachie Jr., Titanium; A Technical Guide chine is expected to reduce, especially MEXT, 2004Ð2007. The authors would (Metals Park, OH: ASM International, 1988), pp. when the tool is inclined in the range of like to especially acknowledge Shi- 131–141. the retreating angle. The miniaturiza- monishi Seisakusho Co., Ltd, Nakata 30. M. Ikeda et al., Proc. 6th Int. FSW Symp. (Cam- bridge, U.K.: The Welding Institute, 2006). tion of the FSW machine is possible. Co., Ltd, and Isel Co., Ltd. They also 31. S. Oki et al., Mater. Sci. Forum, 539-543 (2007), Therefore, the development of the new would like to express their sincere ap- pp. 3838–3843. tool is considered to be cost effective. preciation to Associate Prof. Masato 32. S. Oki et al., Proc. 6th Int. FSW Symp. (Cambridge, U.K.: The Welding Institute, 2006). From the above-mentioned results, it Tsujikawa, Prof. Sachio Oki, Prof. Ma- 33. Y. Okawa et al., Welding Tech., 55 (2007), pp. is apparent that this new tool is highly sahiko Ikeda, Associate Prof. Yorinobu 71–75 (Japanese). suitable for use in 3-D FSW, and the Takigawa, and Prof. Yoji Marutani. 34. M. Tsujikawa et al., Mater. Trans., 46 (2005), pp. 3081–3084. applicability of FSW employing this They would like to express their thanks 35. M. Tsujikawa et al., Mater. Trans., 48 (2007), pp. tool can be enhanced in the future. to the parties concerned in Technology 618–621. Research Institute of Osaka Prefec- 36. J. Kobata et al., Mater. Lett., 61 (2007), pp. 3771– CONCLUSION 3773. ture. 37. T. Hirata et al., Scr. Mater., 56 (2007), pp. 477– Many problems remain in applying References 480. FSW for manufacturing products in 1. C.J. Dawes and W.H. Thomas, Weld. J., 75 (1996), T. Hirata is with the Metals and Machinery Depart- small and medium enterprises. More- pp. 41–45. ment of the Technology Research Institute of Osa- over, it is very important to solve these 2. T. Tanaka et al., J. Japan Inst. Light Metals, 57 ka Prefecture, 2-7-1 Ayumino, Izumi-city, Osaka problems to enhance the applicabil- (2007), pp. 549–553 (Japanese). 594-1157, Japan; K. Higashi is with the Graduate 3. T. Hirata et al., J. JFJA, 7 (2008), pp. 25–30 (Japa- School of Engineering, Osaka Prefecture Univer- ity of the FSW technique. In Osaka nese). sity, 1-1 Gakuen-cho, Sakai-city, Osaka 599-8531, Prefecture, “Research Association for 4. H.J. Liu et al., Sci. Tech. Weld. Join., 8 (2003), pp. Japan.

48 www.tms.org/jom.html JOM • February 2010 Nanocomposite Materials Commentary Nanocomposite Materials—Leading the Way in Novel Materials Design

Jonathan E. Spowart

Note: A nanocomposite is defi ned as a multiphase the processing science needed to take properties via control of nanocrystalline solid material where one of the phases has one, full advantage of the continually widen- structure, through the successful use of two, or three dimensions of less than 100 nm, or structures having nanoscale repeat distances be- ing materials design space. For example, an emerging processing technique. The tween the different phases that make up the mate- two of the papers describe a promising materials of interest are oxide nanocom- rial.1 emerging technique for nanocomposite posites, designed to exhibit very dif- The Nanocomposite Materials sym- material processing: spark-plasma sin- ferent functional behaviors—magnetic posium which was held at the TMS 2009 tering. The technique uses a combina- and optical—therefore emphasizing the Annual Meeting tion of electric fi eld, direct and/or indi- broad applicability of nanocomposites comprised around rect heating, and pressure to densify the beyond structural materials. 70 technical pre- material to increased levels compared The third paper, “Nanostructured sentations submit- with conventional sintering. Likewise, Metal Composites Reinforced with ted from more than additional advances continue to be Fullerenes” by F.C. Robles-Hernández a dozen different made in the area of materials character- and H.A. Calderon is an in-depth in- countries. The re- ization, both in terms of understanding vestigation of the synthesis and char- sulting special em- of the underlying physics of nanoscale acterization of metallic-matrix struc- phasis topic should phenomena, and in the development of tural nanocomposites. The paper gives convey to JOM readers the breadth and new experimental techniques. Scanning insight into the application of standard depth of the technical efforts currently probe microscopy and high-resolution techniques such as TEM, x-ray diffrac- underway in this rapidly developing (scanning) transmission electron mi- tion, and micro-indentation for charac- fi eld. Included are three papers that rep- croscopy (TEM) are rapidly becoming terizing microstructureÐproperty rela- resent this diversity both in material sys- the tools of choice for microstructural tionships. tems (metallic, polymeric, and ceramic) characterization, along with focused ion The future certainly looks bright and applications (structural and/or mul- beam (FIB) milling for specimen prepa- for nanocomposite materials, with the tifunctional). ration and micro-machining. joint ASM/TMS Composite Materials As the fi eld of nanomaterials has ma- The fi rst paper, “Fatigue and Frac- Committee of the Structural Materials tured, nanocomposites have emerged as ture Toughness of Epoxy Nanocompos- Division of TMS recently deciding to a promising and practical path for novel ites,” by I. Srivastava and N. Koratkar organize biennial symposia along the materials design. Although the ability provides a thorough overview and in- same theme. However, one outstanding to combine different material classes to troduction to the major issues affect- issue is raw material cost, which is re- provide a useful balance of properties is ing fatigue and fracture toughness in a lated to quality and reliability of supply nothing new—this is the basis of all ra- major class of structural nanocomposite for large-scale synthesis. It is hoped that tional composite materials design—the materials. The addition of nanofi llers to future TMS symposia will address these ability to combine disparate materials at existing epoxy systems has been shown issues as well as emphasizing the tech- the nanometer scale has led to a plethora to enhance fatigue and fracture behav- nical aspects of nanocomposite materi- of new materials with increased func- ior, even at relatively low levels of rein- als research. Stay tuned. tionality beyond that predicted by the forcement. The article reviews the major Reference “rule of mixtures.” These effects can be technological developments in the fi eld, attributed primarily to the size scale of while describing the underlying, inter- 1. P.M. Ajayan, L.S. Schadler, and P.V. Braun, Nanocomposite Science and Technology (Weinheim, the phases. At the nanometric scale— related mechanisms responsible for the Germany: Wiley VCH, 2003). where surface and interfacial phenom- improvements in performance. ena dominate over volume effects—ma- The second paper, “The Current Ac- Jonathan E. Spowart is a Senior Materials Research Engineer and Research Leader within terial designers need to be aware of such tivated Pressure Assisted Densifi cation the Metals, Ceramics and NDE Division of the quantities as specifi c surface area, sur- Technique for Producing Nanocrystal- Materials and Manufacturing Directorate, Air Force face energy, particle morphology, and line Materials” by J.E. Alaniz, J.R. Mo- Research Laboratory, Wright-Patterson Air Force Base, OH. He is also the secretary and current particle spatial distribution. According- rales, and J.E. Garay, highlights some of JOM Advisor for the joint ASM/TMS Composite ly, there has been a steady maturation in their group’s ongoing work in tailoring Materials Committee.

Vol. 62 No. 2 • JOM www.tms.org/jom.html 49 Nanocomposite Materials Overview Fatigue and Fracture Toughness of Epoxy Nanocomposites

I. Srivastava and N. Koratkar

The fatigue and failure mechanisms propagation rate per cycle (da/dN) is of epoxy composites have been re- How would you… directly proportional to the stress in- searched extensively because of their …describe the overall signifi cance tensity factor range ΔK and two con- of this paper? commercial importance in fi elds de- stants, C and n, which depend on the The mechanical properties of manding materials with high specifi c epoxy nanocomposites have been testing parameters including moisture, strength. Particulate, sheets, short and extensively researched because of temperature, loading frequency, and long fi bers with dimensions in the mi- their commercial importance in the stress ratio.8 The faster the rate of crometer and nanometer range are the fi elds demanding materials with high crack propagation, the lower the fatigue specifi c strength. Various fi llers like major fi llers which have been studied nano-ceramic, nano-rubber, carbon resistance. Smaller values of the Paris for enhancing the fatigue resistance nanotubes, and nano-clay sheets exponent signify high material resis- of epoxies. The nano and micro scale have been studied for enhancing tance to FCP. The Paris exponent, and dimensions of the fi llers give rise to un- the fatigue resistance and fracture hence the fatigue properties of the ma- toughness of epoxy composites. This expected and fascinating mechanical article is a review of the signifi cant terial, have been found to be strongly properties, often superior to the ma- developments to enhance fracture infl uenced by their fracture toughness. trix including fracture toughness and toughness and fatigue resistance of An increase in toughness of the matrix epoxy nano-composites. fatigue crack propagation resistance. has been associated with an increase Such properties are dependent on each …describe this work to a in FCP resistance of the material.9,10 materials science and engineering other (e.g., the fatigue properties of the professional with no experience in The S-N curve is another classical polymer composites have been found your technical specialty? approach used for studying fatigue be- to be strongly infl uenced by its tough- Thermoset polymers like epoxy are havior of a material. The curve gives ness). This article is a review of the important for the aerospace and an estimate of the material’s fatigue various developments in this fi eld and automobile industry due to their life under controlled amplitude of cy- high specifi c strength. However, they the underlying mechanisms which are lack damage tolerance, a critical clic stress. Hysteresis loss measure- responsible for performance improve- attribute for advanced applications. ments are also utilized to determine ments in such composites. Introduction of a second phase in the resistance of polymers to crack the epoxy matrix may improve the initiation and crazing. INTRODUCTION mechanical properties, especially fracture toughness and fatigue The fatigue process has been tradi- Fatigue is a mode of progressive resistance. Epoxy nanocomposites tionally divided into three stages: crack failure in solids under cyclic loading at utilizing carbon nanotubes, nano- initiation, crack propagation, and fi nal ceramics and nano-rubber exhibit 11 stresses lower than the material’s ulti- superior properties compared to the failure. Surface fl aws and inclusions 1 mate tensile stress. Due to the increas- matrix. are the most common sites for fatigue ing use of polymers and their com- crack initiation. In most engineering …describe this work to a posites in engineering applications, layperson? materials cracks are inherently pres- especially the aerospace industry, the Epoxy composites have greatly ent on the component surface. An- fatigue properties of composites have infl uenced the aerospace and other area prone to crack generation inspired much work.2–5 The character- automobile industry by allowing is the second phase-matrix interface. ization of material behavior under cy- new designs and innovations. Two Due to high local stress concentration important material properties critical clic loading is done either by fatigue life for advanced applications are high and the discontinuity in mechanical (S-N) measurement or by studying the fracture toughness and resistance to properties at the particle-matrix inter- fatigue crack propagation (FCP) rate cyclic loading. Epoxy has poor crack face crack initiation becomes easy. In propagation resistance because of in the material. The growth of a crack low ductility and hence second phase polymer matrix composites (PMC), in the stable region under cyclic load- nanoparticles are introduced in the the failure mechanisms are not as well ing (Figure 1) is governed by the Paris epoxy matrix for improvements. understood/defi ned as in metals. In da This article describes how fi llers of the failure of a PMC multiple mecha- law and is given by CK()n .7 various geometries and chemical dN nature affect the matrix. nisms might contribute, such as matrix According to the Paris law, the crack crazing, shear yielding, delamination,

50 www.tms.org/jom.html JOM • February 2010 ered silicate membranes,33,34 and their combinations.35,36 FILLER PROPERTIES AFFECTING TOUGHNESS AND FATIGUE The use of fi llers like elastomers with lower modulus than that of the matrix results in composites with in- creased toughness37 due to high duc- Figure 1. An FCP diagram tility, but modulus of the composite for a neat matrix in com- decreases,18 whereas for inorganic parison to a modifi ed and improved system.6 fi llers the modulus of the composite is observed to be more than that of the polymer matrix.38,39 The size of the fi llers also plays a central role. Use fi ber breakage, debonding, plastic thermoset polymers, studies have been of nano-fi llers (NF) has proved more void growth, and matrix cracking.12 done on the introduction of a second effective than micro-fi ller (MF) in en- Some scanning electron micrographs phase in them. Since the toughening hancing the strength of the polymer to convey a broad picture of failures in mechanism in thermoset polymers is matrix, without compromising ductil- polymer composites are shown in Fig- not a well-understood phenomenon, ity. The size of the NF (a few tens of ure 2. it led to the study of various kinds of nm in diameter) being smaller than the See the sidebar on page 52 for a de- fi llers to obtain optimum mechanical micro-sized fi ller reduces the stress scription of polymer failure modes. properties of the composite. Some concentrations in comparison to MF of the second phases that have been (a few tens of μm in diameter).40 The THERMOSETTING studied are micro- and nano-size par- high stress concentration at the surface EPOXIES ticulates like rubber (e.g., carboxyl makes MF more prone to cracking as Epoxy resin (two or more groups terminated butadiene acrylonitrile compared to NF. of epoxide per molecule) and a curing (CTBN)),21,22 thermoplastics (e.g., Another advantage of having nano- agent comprise an epoxy system. Ep- polyethersulphone),23–26 ceramic par- sized fi llers is that for the same volume 27 oxy composites are used for aircraft, ticles (e.g., Al2O3), some reactive di- fraction of fi llers added, the particle automobiles, ship-building, resin cast- luents;28 long and short fi bers includ- density will be higher for NF than for ing, slide bearings, and many other ing carbon nanotubes;29,30 laminates of MF. Therefore, the distance between applications because of their good fi bers31,32 in different orientations, lay- two neighboring particles will be chemical and corrosion resistance, high specifi c strength, and multiple curing options.2 Despite these advan- 2 μm tages epoxy lacks a crucial attribute: damage tolerance. Offering a fracture energy of less than 0.3 kJ/m2, they can be classifi ed as highly brittle.18 As specifi ed by Johnston,19 the fracture 1 μm energy for a resin to be used in aircraft structures should lie in the range of 1.9–3 kJ/m2. Epoxy possess aromatic rings, which increase the polymer ab chain stiffness and has high cross link density; hence the plastic deformation in front of the crack tip is highly local- ized. This extremely localized plastic deformation causes little absorption of energy leading to catastrophic brittle failure. Enhancement in strength and μ stiffness of the polymer matrix can be 5 m obtained by increasing the cross-link density, but it also increases the brit- cd tleness of the polymer leading to re- Figure 2. Scanning electron micrographs of common mechanisms contributing to 13 14 20 polymer composite failure (a) shear yielding, (b) crazing, (c) fi ber delamination fracture duced toughness. In order to control surface,15 and (d) particle pull out and plastic void growth.16 the toughness and hence the fatigue of

Vol. 62 No. 2 • JOM www.tms.org/jom.html 51 BASIC FAILURE MODES IN POLYMERS Reactive functional elastomers (e.g,. CTBN (carboxyl terminated Butadi- Depending upon the chemistry and the curing agent used, the failure mode of the ene),21,59–61 HTBN (hydroxyl terminated polymer matrix could be either ductile or brittle. The ductile mode is often identifi ed by 62 shear yielding. An increase in brittle nature of the polymer is seen as a shift from shear Butadiene), ATBN (amine terminated 63,64 yielding to micro-voiding followed by crazing. In case of networked polymers like epoxy Butadiene), VTBN (Venyl termi- highly brittle failure without crazing is common. nated Butadiene),65 and ETBN (epoxy terminated Butadiene)66,67) increased Shear Yielding the fracture toughness of the polymer Shear yielding is an energy absorption mechanism associated with polymer failure. It matrix when added between 5–20% occurs when localized plastic fl ow starts in response to an applied stress at approximately weight fraction.18 45 to the applied load. Such plastic fl ow might spread shear bands in the whole sample To obtain rubber dispersion in epoxy, (Figure 2a) absorbing a signifi cant quantity of energy or might lead to localized yielding rubber is dissolved in epoxy resin fol- resulting in isolated shear bands. lowed by curing with appropriate hard- Crazing ener. After curing, a dispersed phase of rubber in epoxy matrix is observed. Crazing is another common failure mode that is observed in glass and amorphous The degree of phase separation and the thermoplastic polymers.17 During crazing, microcracks form under tensile load and are held together by polymer fi brils called crazes (Figure 2b). Formation and plastic dispersion size can be controlled by deformation of such crazes involves absorption of energy, which leads to enhanced controlling curing parameters and fi ller toughness. When the applied stress is high enough to cause failure of the fi brils, the volume fraction.18,68 Elastomer addition microcracks start growing. by either phase separation at adequate temperature or through addition in the form of fi ne solid powder was found smaller for nano-sized fi llers, which fi cient load transfer and the increased to improve the ductility of the thermo- is an important parameter in crack toughness of the composite.52–54 set, increasing its toughness to around bowing and pinning. Lack of transpar- 2–4 kJ/m2.18 Shown in Figure 4a is a PARTICULATE–POLYMER ency is another issue associated with scanning electron micrograph of the COMPOSITE micrometer-sized fi llers in polymer dispersed phase of nano-rubber in ep- composites.41 Transparency of the oxy. Treatment of rubber with silane Rubber – Thermoset composite is improved for nano-fi ll- coupling agents, plasma oxidation, and ers as their size is less than the wave- Around 1971, the Sultan and McGar- acid treatment has been found to en- length of light, especially when a good ry group initiated research on the effect hance interfacial interaction, leading to dispersion is achieved.42 The presence of rubber in thermoset polymers.55–58 an increase in fracture toughness.70–72 of fi llers affects the crystallinity, mo- bility, and other structural aspects of 100 5 polymer chains. 90 (a) Fillers also act as nucleating 80 (b) sites for crystallization of polymer 70 (c) chains.43,44 The effectiveness of load 60 transfer between the polymer and the 50 Figure 3. A comparison

Stress (MPa) 40 fi ller is a function of wettability and 47 30 (a) Neat Polymer of (a) toughness and 45,46 (b) FCP resistance of adherence. The two plots in Figure (b) 10 phr NT-Al O /EP 20 2 3 neat, unmodifi ed Al O 3 demonstrate the difference in the 2 3 10 (c) 10 phr APTES-Al2O3/EP and surface modifi ed static mechanical and fatigue prop- 48 0 Al2O3 in epoxy. 024681012 erties of the neat epoxy, Al2O3 rein- a Strain (%) forced epoxy, and APTES (3-amino- propyltriethoxysilane) modifi ed Al O 2 3 10 x 10–1 reinforced epoxy. The surface modi- fi cation of fi llers improves their dis- Paris Law n = 7 10 x 10–2 persion, which could also be attained da Δ n n = 10 = Co · K by physical means like ultrasonica- dN l –3 49,50 10 x 10 tion. n = 5 Another signifi cant fi ller property –4 is the aspect ratio.51 Layered silicates, 10 x 10 with thickness in nanometers and da/dN (mm/cycle) Neat Polymer 10 x 10–5 10 phr NT-Al 0 /EP length and width in micrometers, and Rate, Growth Crack Fatigue 2 3 carbon nanotubes, with diameter in 10 phr APTES-Al203/EP 10 x 10–6 nanometers and length in micrometers, 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 have much higher aspect ratios than Δ √ b Applied Stress Intensity Range, Ki (MPa · m) particulates. This leads to a more ef-

52 www.tms.org/jom.html JOM • February 2010 a 500 nm b 500 nm c 500 nm Figure 4. Transmission electron micrographs of dispersed phases: (a) nano-rubber, (b) nano-silica, and (c) organoclay C30B in nylon-6 matrix.69

The crack propagation velocity in ~ 20.2 wt.% of nano SiO2 fi llers, with an mode and the fracture surface roughens rubber-reinforced polymers has been increase in elastic modulus from 2.96 (Figure 6c) as an outcome of increase in measured73 and it was observed that GPa to 3.85 GPa.10 fracture toughness.85–87 crack blunting in the rubber phase of the Numerous phenomena are associ- Theoretical models for such defl ec- composite signifi cantly decreased the ated with the toughness increase in par- tion studies have been analyzed by Faber crack propagation velocity. In the pres- ticulate-reinforced composites, some of and Evans.84 Crack pinning is a process ence of rubber particles the crack tip which are crack defl ection, crack pin- similar to pinning of dislocations by im- blunts due to localized shear yielding ning, and debonding followed by plastic purity or second phase particles in the around it. Figure 5a is a plot of FCP rate void growth. Crack defl ection is the pro- matrix. First proposed by Lange,88 the in rubber reinforced epoxy composite cess of tilting and twisting of cracks in process describes the bending of the with increasing fi ller content. the matrix around the fi ller particles, as crack front in the presence of pinning The plot shows improvement in the shown schematically in Figure 6a and b. elements, which obstruct its smooth fatigue resistance of the composite During the defl ection process the crack motion. The crack pinning is studied by with increasing rubber phase fraction. tip propagates under mixed I and II observing the bowing of the crack front, Fatigue property studies of elastomer- epoxy composites have shown the early 10−1 initiation of crazes around the rubber particles, proving useful in the late life of −2 the material by improving the toughness 10 )

of the polymer matrix via microcraz- −1 ing, cavitation, and shear yielding.74,76 10−3 The rubber phase undergoes signifi cant strain and fi nally acts as a ligament in 10−4 the polymer craze.77 However, this duc- da/dN (mm cycle tility is achieved at the cost of modulus, strength, and impact resistance of the 10−5 matrix.18,71 10−6 Ceramic – Thermoset 0.5 1 a ΔK (MPa m1/2) The reinforcement of epoxy using ceramic particles with materials such 100,000 as glass beads,27,78 silica,79–81 and alu- 49,82 mina, from micrometer to nanometer 10,000 size, is found to increase the toughness of the matrix without compromising its rigidity. Surface-coated inorganic fi llers 1,000 have been found to improve mechanical properties of the composite under both Figure 5. (a) FCP rate 100 tensile and cyclic load.49,83 In Figure 4b in neat (†) and rein- da/dN (nm/Cycle) forced epoxy with 1%(x), a scanning electron micrograph of the 5%(+), and 10%(O) rub- 10 dispersed phase of nano-silica in epoxy ber content;74 (b) FCP matrix is shown. The fatigue studies of rate in neat (z), and 1 ceramic-fi lled polymer have shown sig- reinforced epoxy with 1%(), 3%(Δ), 5%(†) 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 nifi cant improvement,50,83 increasing K b ΔK (MPa m1/2) IC and 10%(O) Al2O3 con- from ~0.5 MPa1/2 to 0.88 MPa1/2 with tent.74

Vol. 62 No. 2 • JOM www.tms.org/jom.html 53 as shown schematically in Figure 7a. void growth is another phenomenon Al2O3 particles due to enhanced shear Crack pinning is observed only when contributing to the fracture toughness in- yielding of the polymer. Functionalized the fi ller size is greater than the crack- crease of the polymer matrix infi ltrated nano-Al2O3 fi llers which had better ad- opening displacement; hence it is an ef- with micro- and nano-fi llers (Figure 8). hesion with the epoxy matrix than un- fective pinning mechanism mostly for The debonding of the fi ller relaxes the treated fi llers, were found to have 39% micrometer size fi llers for epoxy matrix stress state at the crack tip and causes higher strain to break as compared to whose crack opening size has been cal- deformation of polymer matrix by void the neat polymer, whereas unmodifi ed μ 85 85 culated to be ~1.7 m. The crack front growth process. nano-Al2O3 fi llers showed only 6% in- bows in between the particles, leaving Plastic void growth of the polymer crease at 10 phr (parts per hundred) behind tail-like structures (Figure 7b). has been reported to be the major cause fi ller concentration. It was realized that

Particle debonding and local plastic of toughening for nano-sized SiO2 and the plastic void growth mechanism is

FATIGUE BEHAVIOR IN COMPOSITES clay epoxy composite has been reported to increase with an increase Clay Nanosheet Polymers in fi ller content.95 The fracture surfaces of such composites have Exfoliated clay sheet reinforced polymers have been considered been found to be much rougher than plain epoxy, indicating higher for improved fatigue resistance of polymers. When the intercala- energy absorbtion by the system prior to fracture. Crack defl ection tion of clay layers through polymer chains is not possible, phase has been implied by the presence of steps and microcracking on the separation occurs and a micro-composite instead of nano-compos- fracture surface.96 The fracture toughness and FCP resistance of the ite is obtained, making the dispersion of nano-clay in the polymer composite have been found to increase by chemical modifi cation of critical for composite properties. Dispersion of clay nano-sheet clay sheets like surface treatment of the clay layers with 3-amino- in polymer is diffi cult due to the hydrophilic nature of clays and propyltriethoxysilane or long chain alkylammonium salt and using hydrophobic nature of polymers. The dispersion barrier is over- physical means like forcing polymer in between clay layers using a come using organic surfactants like vinylbenzyldimethyldodecy- twin screw extruder.97,98 lammonium chloride (VDAC), which makes the clay surface or- Nanofi bers ganophillic.89 Forcing polymer chains through the exfoliated layers of clay sheets (a few nm thick) leads to a fi nely dispersed compos- Nanofi bers are whiskers with diameters in the range of a few hun- ite (Figure 4c).34,90–92 The high aspect ratio of exfoliated clay layers dred nanometers and lengths of a few hundred micrometers. Vapor provides large interfacial contact area with the matrix.93 Studies grown carbon nanofi bers (VGCNFs) obtained from chemical vapor on the mechanical properties including fatigue of such nano-layer deposition of hydrocarbons, have been used to reinforce the epoxy reinforced composites have shown improvement in neat and tra- matrix due to its excellent mechanical properties.99 Vapor grown ditional fi ber reinforced polymers. An addition of 5 wt.% nano carbon nanofi bers consist of a few-nanometer-diameter carbon silicate clay in polypropylene has been found to increase fatigue tubes surrounded by concentric circles of carbon layers increasing strength coeffi cient (true stress which causes failure in one rever- its outer diameter to around 50–200 nm. sal) by 13.3% with respect to the neat polymer and at the same Dispersion of the VGCNFs is diffi cult to achieve as the fi bers get time resulting in an increment in modulus and yield strength by twisted and intertwined. Using surface treatments like functional- 90% and 5%, respectively.94 The stiffness and toughness of nano- ization and plasma treatment, better dispersion and interfacial ad-

10−3 Theory - PURE EPOXY Theory - 0.10% MWNT Theory - 0.25% MWNT Theory - 0.50% MWNT 10−4

a d

da/dN (mm/cycle) 10−5

10−6 0.3 0.35 0.4 0.45 0.5 0.55 b c ΔK(MPa√m) e Figure A. (a) Scanning electron micrograph of the side view of the fatigue crack. The microstructure of the sample in the vicinity of the crack is shown. (b) Close to the crack tip, fi ber bridging zone is shown, where the nanotubes are pulled out of the matrix but are effectively bridging the crack interface. (c) At a small distance behind the crack tip, some nanotubes are bridging the crack and some are pulled out of the matrix. (d) Far behind the crack tip, nanotubes are completely pulled out of the matrix. (e) Lowering of FCP rate in MWNT reinforced epoxy with increasing weight fraction of MWNT.29

54 www.tms.org/jom.html JOM • February 2010 the main mechanism resulting in the erties have poor transverse properties these traditional composites is by us- rise in toughness.47,48 Figure 5b is a plot due to lack of fi bers in the thickness ing nanofi llers like CNTs to improve 105 of FCP in Al2O3 epoxy composite with direction. Lack of fi bers through the the interlaminar strength of the com- varying volume fraction of fi ller. See thickness of the composite also leads posite.108 The CNTs could either be the sidebar for fatigue behavior in vari- to poor fatigue behavior. Weaving of randomly distributed or vertically ori- ous composites. fi bers in the thickness direction (i.e., ented on the fi bers.109–111 The effect of stitched composites) can be used to im- functionalization of the CNTs in such THREE-PHASE prove the transverse properties of lami- three phase composites was observed COMPOSITES nated composites but only at the cost of to further enhance the transverse and Oriented fi ber reinforced compos- longitudinal properties.106,107 Another longitudinal toughness of traditional ites with excellent longitudinal prop- approach to improve the properties of epoxy/fi ber composites.112,113

hesion between the polymer matrix and carbon nano-fi bers (CNF) crack suppression mechanism is crack bridging. The fatigue crack is achieved.100 Since random dispersion of VGCF leads to isotropic is bridged by high aspect ratio nanotubes generating a fi ber-bridg- properties in the composite, to obtain desired alignment extrusion101 ing zone in the wake of the crack tip. As the crack advances energy and magnetic fi eld102 are employed. Vapor grown carbon nanofi bers is dissipated by the frictional pull-out of the bridging nanotubes have resulted in an increase in the fracture toughness of epoxy with from the epoxy matrix which slows the crack propagation speed. increasing fi ber content.103 The fatigue life and the fatigue strength of However this crack bridging effect loses effectiveness at high ΔK nanofi ber-reinforced epoxy were also found to improve substantially due to progressive shrinkage in the size of the fi ber-bridging zone on introduction of ~2% VGCF.104 The increase in fracture toughness as ΔK is increased.29 The fact that such behavior is not observed is attributed to fi ber debonding and the pull out mechanism. in nanoparticle fi lled epoxy composites indicates that crack bridg- ing phenomena are not playing a dominant role in the fracture and Carbon Nanotube Polymers fatigue of nanoparticle fi lled polymer systems. Reducing the CNT Carbon nanotubes have been demonstrated to increase the fatigue diameter and increasing its length has been shown to increase the life of PMCs. The fatigue crack propagation velocity is signifi cantly effectiveness (Figure B) of the crack bridging mechanism.30 reduced through the action of crack bridging by the CNTs and the Functionalizing CNT with amido-amine groups has also been subsequent dissipation of energy that occurs due to the frictional pull shown to enhance the epoxy’s resistance to fatigue crack propaga- out of the bridging CNT fi bers as shown in Figure Aa–d. In the low tion by initiating craze formation (Figure 2b) in the epoxy.14 The stress intensity factor amplitude regime a decrease in the crack propa- origin of the crazing behavior was traced to a signifi cant amount of gation rate by over ten-fold is reported,30 with strong dependence on unreacted epoxy that was kinetically trapped in the crosslinked ma- the weight fraction of CNTs. Figure Ae shows the increase in resis- trix structure that is formed at the CNT/epoxy interface. Such local tance to fatigue crack propagation with an increase in the weight frac- heterogeneity in the curing may be caused by a variety of factors tion of multiwalled carbon nanotubes (MWNTs) in the epoxy matrix. such as, for example, the fact that the chemistry may be modifi ed The amount of CNT required for such a reduction in FCP rate is less locally due to the presence of amido-amine groups. Epoxy chain than 1% which is small in comparison to the amount of particulate alignment, which is known to infl uence the cross-linking density, nano-fi llers (~5–10%). may also be modifi ed locally due to the presence of these functional For the case of CNT fi llers, a substantial reduction in the fatigue groups. Heterogeneous cross-linking results in localized pockets of suppression with increasing ΔK (Figure Ae) is observed. This is enhanced molecular mobility;14 the correlated evolution of such because in the case of CNTs the dominant toughening and fatigue contiguous mobile regions under mechanical loading leads to craz- ing which substantially boosts the toughness and the resistance to 10−3 fatigue crack propagation of the baseline epoxy. Epoxy −3 10-20 μm 10 1-2 μm Epoxy 5-8 nm 10-20 nm (long) 20-30 nm 50-70 nm 10−4 10-20nm (short) 10−4 da/dN(mm/cycle) da/dN(mm/cycle)

10−5 10−5

0.25 0.3 0.35 0.4 0.45 0.5 0.55 0.6 0.65 0.25 0.3 0.35 0.4 0.45 0.5 0.55 0.6 0.65 ΔK(MPa√m) Δ √ abK(MPa m) Figure B. (a) Effect of nanotube diameter on fatigue suppression performance; (b) effect of nanotube length on fatigue suppression perfor- mance.30

Vol. 62 No. 2 • JOM www.tms.org/jom.html 55 searched for a wide variety of second- tigue crack by crack defl ection, crack CONCLUSION phase particles. Different kinds of fi ll- pinning, debonding, and plastic void Fatigue life improvements in epoxy ers have different ways of obstructing growth whereas short fi bers undergo composites have been studied and re- FCP. Particulate fi bers obstruct the fa- crack bridging and crack trapping. Rubber composites degrade the original stiffness of the epoxy matrix whereas inorganic particles are not as effi cient as rubber in enhancing the Figure 6. A schematic toughness of the composite. High-as- of crack front under- pect-ratio fi bers such as carbon nano- going (a) tilt and (b) twist under mixed tubes are able to signifi cantly enhance mode on encounter- the epoxy’s fatigue crack propagation ing second phase resistance at relatively low weight particles,84 (c) and (d) scanning electron fractions (below 0.5%), but they lose micrograph compar- their effectiveness at high stress in- ing the surface rough- tensity factor amplitudes. Graphene is ness of neat and Al2O3 modifi ed epoxy.85 another emerging nano-material and ab shows promise to enhance fracture and fatigue properties of polymers114 at low loading fractions due to its high specifi c surface area, two-dimensional sheet geometry, strong fi ller-matrix adhesion, and the outstanding me- chanical properties of the sp2 carbon bonding network in graphene.115 Due to the increasing interest of research- ers in graphene composites, in-depth study of the fracture and fatigue prop- c d erties of grapheme-based composite materials is clearly warranted. In addition, theoretical/computa- tional modeling of fracture and fatigue in nanocomposites is still in its infan- cy. Improved models for toughness in nanocomposites and validation with experimental data are essential for development of nanocomposite poly- mers with optimized strength, stiff- ness, fracture, and fatigue properties.

Crack ACKNOWLEDGEMENTS ­ Propagation Direction N.K. acknowledges the fund- ing support from the U.S. Offi ce of a μ b 10 m Naval Research (Award Number: Figure 7. (a) A schematic of second phase particles pinning the crack front, leaving behind tail-like structure.41 (b) Scanning electron micrograph illustrating the fracture surface of N000140910928) and the U.S. Na- 75 epoxy reinforced with TiO2 nanoparticles, demonstrating the tail-like structures. tional Science Foundation (Award Number: 0900188).

References 1. S. Suresh, Fatigue of Materials, 2 (Cambridge Univer- sity Press, 1998). 2. L.A. Pilato and M.J. Michno, Advanced Composite Materials, 1 (Berlin, Germany: Springer, 1994). 3. P.M. Ajayan et al., Nanocomposite Science and Tech- nology, 1 (Weinheim, Germany: Wiley-VCH, 2003). 4. D. Hull and T. Clyne, An Introduction to Composite 50 μm 10 μm Materials, 2 (Cambridge, U.K.: Cambridge Univ. Press, a b 1996). 5. I.S. Lipatov et al., Polymer Reinforcement, 1 (Ontario, Figure 8. (a) and (b) Scanning electron micrographs illustrating the fracture surface of Canada: ChemTec Publishing, 1995). 16 glass-particle epoxy matrix after debonding. 6. V. Altstadt, “Fatigue of Polymeric Materials,” Encyclo- pedia of Materials: Science and Technology (Amster-

56 www.tms.org/jom.html JOM • February 2010 dam, NY: Elsevier Science Ltd, 2001), pp. 2972–2978. (1) (1983), pp. 161–174. 83. L.E. Nielsen, J. Composite Materials, 9 (2) (1975), 7. P.C. Paris et al., The Trend in Engineering, 13 (1961), 44. M. Fujiyama and T. Wakino, J. Applied Polymer pp. 149–156. p. 6. Science, 42 (10) (1991), pp. 2739–2747. 84. K.T. Faber and A.G. Evans, Acta Metallurgica, 31 8. Y.W. Mai and J.G. Williams, J. Materials Science, 14 45. R.N. Rothon, Particulate Fillers for Polymers, 1 (4) (1983), pp. 565–576. (8) (1979), pp. 1933–1940. (Shawbury, U.K.: Rapra Technology Ltd., 2001). 85. B.B. Johnsen et al., Polymer, 48 (2) (2007), pp. 9. E. Urbaczewski-Espuche et al., J. Applied Polymer 46. J.M. Felix and P. Gatenholm, J. Applied Polymer 530–541. Science, 47 (6) (2003), pp. 991–1002. Science, 42 (3) (1991), pp. 609–620. 86. D. Hull, J. Materials Science, 31 (17) (1996), pp. 10. B. Blackman et al., J. Materials Science, 42 (16) 47. S. Zhao et al., Composites Science and Technology, 4483–4492. (2007), pp. 7049–7051. 68 (14) (2008), pp. 2965–2975. 87. K.T. Faber and A.G. Evans, Acta Metallurgica, 31 11. W.J. Plumbridge, J. Materials Science, 7 (8) (1972), 48. S. Zhao et al., Composites Science and Technology, (4) (1983), pp. 577–584. pp. 939–962. 68 (14) (2008), pp. 2976–2982. 88. F.F. Lange, Philosophical Magazine, 22 (179) 12. M.J. Salkind, Composite Materials Testing and 49. Y. Wang et al., Carbon, 41 (15) (2003), pp. 2939– (1970), pp. 983–992. Design (2nd Conf). ASTM STP 497 (Baltimore, MD: 2948. 89. X. Fu and S. Qutubuddin, Polymer, 42 (2) (2001), Amer. Soc. for Testing and Materials, 1972). 50. D.R. Paul and L.M. Robeson, Polymer, 49 (15) pp. 807–813. 13. W. Bascom et al., J. Materials Science, 16 (10) (2008), pp. 3187–3204. 90. S.C. Tjong et al., Chemistry of Materials, 14 (1) (1981), pp. 2657–2664. 51. H.S. Katz and J.V. Milewski, Handbook of Fillers for (2002), pp. 44–51. 14. W. Zhang et al., Small, 5 (12) (2009), pp. 1403– Plastics (New York: Van Nostrand Reinhold, 1987). 91. Z. Wang and T.J. Pinnavaia, Chemistry of Materials, 1407. 52. K. Tamura et al., Chemistry of Materials, 20 (6) 10 (7) (1998), pp. 1820–1826. 15. T. Johannesson et al., J. Materials Science, 19 (4) (2008), pp. 2242–2246. 92. K.E. Strawhecker and E. Manias, Chemistry of (1984), pp. 1171–1177. 53. C. Henrist et al., J. European Ceramic Society, 27 Materials, 12 (10) (2000), pp. 2943–2949. 16. T. Kawaguchi and R.A. Pearson, Polymer, 44 (15) (2-3) (2007), pp. 1023–1027. 93. P.C. LeBaron and T.J. Pinnavaia, Chemistry of (2003), pp. 4239–4247. 54. S. Iijima, Nature, 354 (6348) (1991), pp. 56–58. Materials, 13 (10) (2001), pp. 3760–3765. 17. J. Scheirs, Compositional and Failure Analysis of 55. J.N. Sultan et al., Applied Polymer Symposium, 16 94. Y. Zhou et al., Materials Science and Engineering A, Polymers (Chichester, NY: Wiley, 2000). (1971), pp. 127–136. 402 (1-2) (2005), pp. 109–117. 18. A.C. Garg and Y.-W. Mai, Composites Science and 56. J.N. Sultan and F.J. McGarry, Polymer and 95. O. Becker et al., Polymer, 43 (16) (2002), pp. Technology, 31 (3) (1988), pp. 179–223. Engineering Science, 13 (1) (1973), pp. 29–34. 4365–4373. 19. N.J. Johnston, “Synthesis and Toughness 57. M.E. Roylance et al., Am. Chem. Soc. Div. Org. 96. S.C. Tjong, Materials Science and Engineering: R: Properties of Resins and Composites” (Presentation Coatings Plast. Chem., 34 (2) (1974), pp. 294–299. Reports, 53 (3-4) (2006), pp. 73–197. at the ACEE Composite Struct. Technol. Program, 58. F.J. McGarry, Polymeric Materials Science and 97. S.R. Ha et al., J. Nanoscience and Nanotechnology, Seattle, WA, 1984). Engineering, Proceedings of the ACS Division of 7 (11) (2007), pp. 4210–4213. 20. A.S. Argon and R.E. Cohen, Polymer, 44 (19) Polymeric Material, 49 (1983), p. 223. 98. M. Battistella et al., Composites Part A: Applied (2003), pp. 6013–6032. 59. A.J. Kinloch et al., Polymer, 24 (10) (1983), pp. Science and Manufacturing, 39 (12) (2008), pp. 1851– 21. S.C. Kunz and P.W.R. Beaumont, J. Materials 1341–1354. 1858. Science, 16 (11) (1981), pp. 3141–3152. 60. R.A. Pearson and A.F. Yee, J. Materials Science, 26 99. V.Z. Mordkovich, Theoretical Foundations of 22. A.J. Kinloch et al., Polymer, 24 (10) (1983), pp. (14) (1991), pp. 3828–3844. Chemical Engineering, 37 (5) (2003), pp. 429–438. 1355–1363. 61. C.B. Bucknall and I.K. Partridge, Polymer 100. O.M. Teyssier et al., Macromolecular Materials 23. C.B. Bucknall and I.K. Partridge, British Polymer Engineering & Science, 26 (1) (1986), pp. 54–62. and Engineering, 291 (12) (2006), pp. 1547–1555. Journal, 15 (1) (1983), pp. 71–75. 62. S. Sankaran and M. Chanda, J. Applied Polymer 101. M.L. Shofner et al., Composites Part A: Applied 24. C.B. Bucknall and A.H. Gilbert, Polymer, 30 (2) Science, 39 (8) (1990), pp. 1635–1647. Science and Manufacturing, 34 (12) (2003), pp. 1207– (1989), pp. 213–217. 63. S.C. Kunz et al., Polymer, 23 (13) (1982), pp. 1217. 25. R.A. Pearson and A.F. Yee, Polymer, 34 (17) (1993), 1897–1906. 102. T. Takahashi et al., Chemical Physics Letters, 436 pp. 3658–3670. 64. G. Levita et al., Polymer, 26 (7) (1985), pp. 1110– (4-6) (2007), pp. 378–382. 26. T. Kyu and J.-H. Lee, Physical Review Letters, 76 1116. 103. H. Miyagawa and L.T. Drzal, Composites Part A: (20) (1996), pp. 3746–3749. 65. O. Gryshchuk et al., Polymer, 43 (17) (2002), pp. Applied Science and Manufacturing, 36 (10) (2005), 27. J. Spanoudakis and R.J. Young, J. Materials 4763–4768. pp. 1440–1448. Science, 19 (2) (1984), pp. 473–486. 66. E.J. Robinette et al., Polymer, 45 (18) (2004), pp. 104. Y. Zhou et al., J. Materials Science, 42 (17) (2007), 28. E. Urbaczewski et al., Die Makromolekulare 6143–6154. pp. 7544–7553. Chemie, 191 (4) (1990), pp. 943–953. 67. J.-F. Hwang et al., Polymer Engineering and 105. L. Tong et al., 3D Fibre Reinforced Polymer 29. W. Zhang et al., Applied Physics Letters, 91 (19) Science, 29 (20) (1989), pp. 1477–1487. Composites, 1 (Oxford, U.K.: Elsevier Science, 2002). (2007), pp. 193109–193113. 68. E. Dreerman et al., J. Applied Polymer Science, 72 106. A.P. Mouritz, Proceedings of the Institution of 30. W. Zhang et al., Nanotechnology, 19 (28) (2008), (5) (1999), pp. 647–657. Mechanical Engineers, Part L: Journal of Materials: pp. 1–5. 69. G.-T. Wang et al., Engineering Failure Analysis, 16 Design and Applications, 218 (2) (2004), pp. 87–93. 31. N.L. Harrison, Fibre Science and Technology, 4 (2) (8) (2009), pp. 2635–2645. 107. A.P. Mouritz, Composites Science and Technology, (1971), pp. 101–114. 70. T. Okamatsu et al., J. Adhesion Science and 68 (12) (2008), pp. 2503–2510. 32. H.J. Konish, Jr. et al., AIAA Journal, 11 (1) (1973), Technology, 12 (8) (1998), pp. 813–830. 108. V.P. Veedu et al., Nat. Mater., 5 (6) (2006), pp. pp. 40–43. 71. J. Wang et al., J. Applied Polymer Science, 60 (9) 457–462. 33. S.S. Ray and M. Okamoto, Progress in Polymer (1996), pp. 1425–1437. 109. M.F. De Riccardis et al., Carbon, 44 (4) (2006), Science, 28 (3) (1995), pp. 1539–1641. 72. E. Sipahi-Saglam et al., Polymer Engineering and pp. 671–674. 34. M. Alexandre and P. Dubois, Materials Science and Science, 41 (3) (2001), pp. 514–521. 110. S. Zhu et al., Diamond and Related Materials, 12 Engineering R: Reports, 28 (1) (2000), pp. 1–63. 73. D. Peretz and A.T. DiBenedetto, Engineering (10-11) (2003), pp. 1825–1828. 35. Y.P. Wu et al., Composites Science and Technology, Fracture Mechanics, 4 (4) (1972), pp. 979–984. 111. I. Srivastava et al., J. Nanoscience and 65 (7-8) (2005), pp. 1195–1202. 74. H.R. Azimi et al., J. Materials Science, 31 (14) Nanotechnology, 10 (2) (2010), pp. 1025–1029. 36. S.T. Kim et al., J. Materials Science, 31 (13) (1996), (1996), pp. 3777–3789. 112. J. Qiu et al., Nanotechnology, 18 (27) (2007), pp. pp. 3523–3533. 75. B. Wetzel et al., Engineering Fracture Mechanics, 1–11. 37. L.E. Nielsen and R.F. Landel, Mechanical 73 (16) (2006), pp. 2375–2398. 113. R. Andrews and M.C. Weisenberger, Current Properties of Polymers and Composites, 2 (New York: 76. T.A. Morelli and M.T. Takemori, J. Materials Science, Opinion in Solid State and Materials Science, 8 (1) CRC Press, 1994). 18 (6) (1983), pp. 1836–1844. (2004), pp. 31–37. 38. S.Y. Fu and B. Lauke, J. Materials Science and 77. S. Kunz-Douglass et al., J. Materials Science, 15 114. M.A. Rafi ee et al., Small (2009), doi 10.1002/ Technology, 13 (5) (1997), pp. 389–396. (5) (1980), pp. 1109–1123. smll.200901480. 39. N. Amdouni et al., J. Applied Polymer Science, 46 78. J. Spanoudakis and R.J. Young, J. Materials 115. A.K. Geim, Science, 324 (5934) (2009), pp. (10) (1992), pp. 1723–1735. Science, 19 (2) (1984), pp. 487–496. 1530–1534. 40. J. Moczo and B. Pukánszky, J. Industrial and 79. A. Kinloch et al., J. Materials Science, 40 (18) Engineering Chemistry, 14 (5) (2008), pp. 535–563. (2005), pp. 5083–5086. I. Srivastava, research assistant with the Depart- 41. A.J. Kinloch and R.J. Young, Fracture Behaviour of 80. Y. Nakamura et al., Polymer, 32 (12) (1991), pp. ment of Materials Science and Engineering, and Polymers (London: Elsevier Applied Science, 1983). 2221–2229. N. Koratkar, professor with the Department of 42. C. Becker et al., Organic-Inorganic Hybrid Materials 81. R.J. Young and P.W.R. Beaumont, J. Materials Mechanical, Aerospace and Nuclear Engineering, for Photonics, ed. L.G. Hubert-Pfalzgraf and S. Iraj Science, 10 (8) (1975), pp. 1343–1350. are with Rensselaer Polytechnic Institute, 110 Najafi (Bellingham, WA: SPIE, 1998), pp. 88–98. 82. F.F. Lange and K.C. Radford, J. Materials Science, 8th Street, Troy, NY 12180. Ms. Srivastava can be 43. J. Menczel and J. Varga, J. Thermal Analysis, 28 6 (9) (1971), pp. 1197–1203. reached at [email protected].

Vol. 62 No. 2 • JOM www.tms.org/jom.html 57 Nanocomposite Materials Research Summary The Current Activated Pressure Assisted Densifi cation Technique for Producing Nanocrystalline Materials

J.E. Alaniz, J.R. Morales, and J.E. Garay

The harnessing of length-scale-de- larger role in the overall properties of the sidebar for a description of CAPAD pendent material properties has been an the bulk materials. Several oxide ma- processing. area of intense research in the past two terials have been produced using this MAGNETIC PROPERTIES decades. Retention of nanocrystalline unique process with grain sizes <100 OF IRON OXIDE features in fully dense bulk materials nm and even <20 nm.3–6 NANOCOMPOSITES has been particularly elusive. Recently This paper will highlight some of our the current activated pressure assisted group’s ongoing work with the over- Recently we have been exploring densifi cation (CAPAD) technique has arching theme of tailoring properties CAPAD as an effi cient method for pro- emerged as one of the most successful using the nanocrystalline structure. See cessing bulk magnetic nanocomposites methods for the production of function- that display length-scale dependent 8,9 al and structural nanocrystalline mate- How would you… magnetic properties. Because mag- rials with large sizes. In this article we …describe the overall signifi cance netic domain behavior and interactions review some ongoing efforts in using of this paper? of magnetic moments occur in true na- the CAPAD technique for producing This article outlines the benefi ts of noscale dimensions (<100 nm), materi- single component and nanocomposite the Current Activated Pressured als for both hard and soft magnetic ap- systems from nanocrystalline powders. Assisted Densifi cation (CAPAD) plications can be signifi cantly affected technique for producing bulk The properties of these nanocrystal- nanocrystalline materials from by careful control of both composition line materials include improved light nanopowders. The CAPAD technique and nanoscale grain structure. Conven- transmittance and magnetic properties is not only effi cient, but it also offers tional reaction sintering can achieve caused by interfacial coupling of anti- a higher degree of control than precise control of magnetic phase com- the better known and widely used ferromagnetic/ferrimagnetic phases. methods of powder densifi cation. position in bulk materials, but it re- quires long processing times that come INTRODUCTION …describe this work to a materials science and engineering at the expense of nano-structure. The Current activated pressure assisted professional with no experience in strategy for making bulk nanocompos- densifi cation (CAPAD), often referred your technical specialty? ites is two-fold: We rely on phase trans- to as spark plasma sintering (SPS) or Retention of nanocrystallinity formations so that the materials “self- is diffi cult since densifi cation at fi eld assisted sintering (FAST), uses high temperature is accompanied assemble” into nanocomposites, and high heating rates (via Joule heating by thermally activated grain we densify materials quickly to prevent from an applied current) combined growth. However, the CAPAD exaggerated grain growth. CAPAD is with simultaneous pressure application technique’s unique combination excellent for taking advantage of both of simultaneously applied electric to densify materials. The densifi cation current and pressure allow the metastable nature of iron oxide is believed to be aided by high current the production of non-porous phases and the high densifi cation rates fl ux which infl uences diffusion con- bulk materials with nanometric necessary to maintain the nano-scale trolled mass transport as well as phase microstructure. At nano-length grain sizes of the starting powder. scales, interactions between grains 1,2 γ growth in the reaction process. Often, are different than their micro- Commercial single-phase -Fe2O3 as in the case of bulk nanocrystalline structured counterparts. Two nanopowders (Alfa Aesar, USA) with materials, CAPAD is capable of produc- examples of this nanoscale related average crystallite diameters of 8 nm phenomenon are shown here. ing materials unobtainable through tra- and 40 nm were used. Figure 1 shows ditional powder densifi cation methods. …describe this work to a a typical sample after CAPAD process- layperson? High heating rates and short processing ing, next to a coin for size comparison. Large sized nanocrystalline materi- times on the order of minutes allow a als are exciting because they behave SEM images of samples derived from high degree of densifi cation while min- in very different ways allowing for 8 nm and 40 nm powders are shown in imizing grain growth. This allows high new applications. In this paper we Figure 2a and b. These samples were grain and phase boundary densities to present a method that can produce densifi ed at 650C. For all processing bulk nano-materials for optical and be maintained, allowing the properties magnetic applications. temperatures, the average fi nal grain of these boundaries/interfaces to play a size of material from the 8 nm powder

58 www.tms.org/jom.html JOM • February 2010 is smaller (~3 times) than that from the 40 nm-derived powder. XRD analysis reveals that the samples are composed of multiple iron oxide phases (i.e., they are composites). Figure 3 shows promi- nent peaks representing a hexagonal lat- α tice structure ( -Fe2O3) which gradually diminish as the processing temperature is increased to 900C. Inversely, peaks consistent with a cubic lattice structure Figure 1. A typical sample next to a quar- ter. These bulk samples are 19 mm in (Fe O ) increase in amplitude as the 3 4 diameter and about 1 mm thick. a 500 nm temperature is increased until they are the only peaks present at 900C. creases at ~20 K, ~110 K, and 265 K. Table I summarizes density, average The sample processed at 900C demon- grain size, and compositional informa- strates the highest step height at ~110 tion for both 8 nm and 40 nm samples K as well as the highest net magneti- at all of the processing temperatures. It zation. Figure 5 shows data for the 40 is seen that 40 nm samples reach >0.90 nm samples ZFC’ed at 100 Oe, and density by 650C, whereas 8 nm sam- processed between 300 and 800C. ples reach this density between 700C Clear step-like increases in magnetiza- and 800C. Quantitative analysis of tion appear at two temperatures, 120 XRD data shown in the third column of K and 260 K. In samples processed b Table I refl ects the total phase transfor- below 650C the step heights at both 500 nm mation of α-Fe O (hexagonal) to Fe O temperatures are similar in proportion Figure 2. SEM micrographs of (a) 40 2 3 3 4 nm and (b) 8 nm densifi ed nanopowder. (cubic) mentioned earlier. as is the magnitude of magnetization. In Each sample was processed at 650C. The magnetic data for the 8 nm sam- samples processed above 600C, the in- ples zero-fi eld cooled (ZFC’ed) at 100 crease at 120 K is distinctly prominent Coercivity measurements (Figure 6) Oe are shown in Figure 4. In samples over the diminutive step at 260 K. The show different behavior between the 8 processed below 750C, the most sig- largest change in step height as well as nm and 40 nm samples. 8 nm samples nifi cant increase in magnetization oc- net magnetization occurs in the sample show a steep decrease in coercivity curs at ~265 K. Samples processed processed at the highest temperature, around 20 K and a small change at ~110 above 650C show three separate in- 800C. K with the exception of the 900C sam- ple. In contrast, the change in coercivity is unmistakable at 120 K for the 40 nm CAPAD PROCESSING samples. In the 900C sample, both the Important components of a CAPAD system include a vacuum chamber and electrodes coercive value as well as the change in for delivering large electric currents and mechanical loads. We use a custom built CAPAD apparatus. Within the chamber, two water cooled copper electrodes are used to apply a coercivity is less than it is for samples load of up to 150 kN from a Universal test frame (Instron 5584, Instron Inc., Canada) to a processed at lower temperatures such as graphite die and plunger system that contains the powder to be densifi ed. These electrodes 550C and 675C. γ are capable of supplying a 4800 A current from programmable power supplies (Xantrex Ferrimagnetic -Fe2O3(fM) undergoes Inc., Canada) while the temperature of the sample is measured and controlled using a type- a phase transformation to antiferromag- k thermocouple placed in the α netic -Fe2O3(AFM) at temperatures high- die. Figure A shows a simple er than 300C. A second phase trans- schematic of this system. formation occurs between α-Fe O Commercial nanocrys- 2 3(AFM) and Fe3O4(fM) due to oxygen reduction talline powders were used during processing. Both α-Fe O in all of the work discussed 2 3(AFM) and Fe O are subject to changes here. Typical processing in- 3 4(fM) volves placing 1–2 g of the in magnetic behavior due to tempera- nanopowder between graph- ture dependent magnetic ordering. For α ite dies of 19 mm diameter. example, in -Fe2O3(AFM), the Morin We use a pressure of 100– transition coincides with a change of 140 MPa. For a detailed re- antiparallel magnetic moment align- view of the CAPAD process, ment parallel to the c axis below TM please see Reference 7. = 265 K. Above 265 K, the magnetic moment alignment fl ips to point par- Figure A. Schematic of the allel to the basal plane. The reorienta- device used to perform the CAPAD technique in the stud- tion is associated with a small canting ies presented in this paper. of the moments which results in a bulk weak ferromagnetic property. Evidence

Vol. 62 No. 2 • JOM www.tms.org/jom.html 59 and is higher in composites with small- er grain sizes.9 OPTICAL PROPERTIES OF NANOCRYSTALLINE CUBIC ZIRCONIA There has been increasing interest in coupling the optical properties of single Figure 3. XRD data for samples crystal ceramics with polycrystalline prepared over a range of tem- materials in order to make them more peratures. A gradual temperature 13,14 dependent phase transformation versatile. The CAPAD technique between 600C and 900C is has been used to make several transpar- seen as hexagonal peaks vis- ent bulk polycrystalline ceramics. This ible at 600C eventually disap- pear and cubic peaks gradually article will discuss how it was used to increase until they are the only make polycrystalline 8 mol.% yttria peaks present in the 900C stabilized zirconia (8YSZ) transparent15 sample. and how these optical properties were of the reorientation can be detected ity. Similar to the correlation between characterized.16 through small increases in magnetiza- temperature and coercivity, there is a Aside from having excellent mechan- 11,12 tion values. In the case of Fe3O4(fM), strong correlation between temperature ical properties, polycrystalline 8YSZ is the change in the easy axes of magne- and magnetization in the 40 nm samples widely used as a thermal barrier coating tization coincides with a structural lat- (i.e., the changes in magnetization at TV (such as for gas turbine components, 17 tice change. Below the Verwey transi- =120 K and TM = 265 K). Interestingly, etc.) and as an ionic conductor (such tion (TV = 125 K), the monoclinic easy our nanocomposites display exchange as used in oxygen sensors). However, by direction is <100>. Above the TV, the bias caused by magnetic coupling at capturing the optical properties found crystal structure changes to FCC with interfaces. The magnitude of the ex- in single-crystal zirconia, this mate- easy directions of magnetization in the change fi eld is composition dependent rial could potentially be used in a much <111> directions.11,12 This change pro- duces a decrease in coercivity at the TV γ which is not seen in -Fe2O3(fM). In the 8 nm samples, the 265 K in- crease in ZFC magnetization is due to α the Morin transition of -Fe2O3(AFM). The strongest increase in magnetization at ~110 K for the 8 nm 900C sample can be linked to the lower temperature (~110 K) decrease of coercivity. This suggests that a low-temperature Ver- Figure 4. Mass moment wey transition is responsible for both (measured in ZFC (100 changes in magnetization and coerciv- Oe)) data for the sam- ples prepared from the 8 nm powder. Table I. Comparison of Experimental Parameters (Temperature and Avg. Crystallite Size of Nanopowder) and the Resulting Bulk Material Properties (Density, Avg. Grain Size, and Composition) Powder Avg. GS % Comp. Temp. (nm) Density (nm) Hex:Cubic

600C 8 0.69 — 95:5 40 0.83 — 100:0 650C 8 0.71 58 80:20 40 0.99 180 76:24 Figure 5. Mass mo- 700C 8 0.78 77 62:38 ment (measured in 40 0.97 191 32:68 ZFC (100 Oe)) data for the samples pre- 800C 8 0.93 91 1:99 pared from the 40 nm 40 0.97 372 15:85 powder. 900C 8 0.94 173 0:100 40 0.97 558 0:100

60 www.tms.org/jom.html JOM • February 2010 wider variety of applications such as furnace and cryogenic devices. Each of the zirconia samples discussed in this paper were processed under vac- uum (<4 × 10–4 torr) and contained 1.5 g of cold pressed (71 MPa for 1 min.) 8YSZ powder (Tosoh Corporation, To- kyo, Japan) with a reported crystal size of 50 nm. Two different processes were Figure 7. Photographs of densifi ed 8YSZ ceramics on a light table on top of text showing used to create these samples. The two- varying colors of samples held for various times during the 2-step process: (a) 10 min. hold; (b) 11 min. hold; (c) 12 min. hold. The text is legible through all samples and the color step load process consisted of an ini- of the samples gets increasingly darker with time held at temperature. Reprinted with –1 tial load ramp at 10 kN min. up to a permission from S.R. Casolco et al.15 pressure of 106 MPa (30 kN load). The samples were then heated to 1,200C at approximately 1 mm thick. The two-step refl ection, and absorption data were tak- a rate of 200C min.–1 after which the samples showed different characteristics en for this sample using light of 633 nm pressure was then raised to 141 MPa (40 as the fi nal hold time was changed (Fig- wavelength. As seen in Figure 10, the kN load). These samples were then held ure 7). As each of these samples were absorption decreases dramatically with at temperature and pressure for varying created with the same temperatures and increased annealing time, indicating that amounts of time (10, 11, and 12 min.). pressures, it is reasonable to expect a dif- free electrons (due to oxygen vacancies The one-step process simply consisted ferent light interaction mechanism than and their associated defect structures) of the initial load increase to either 30 porosity is causing these differences. are indeed responsible for light absorp- or 40 kN, the heating process, and the Scanning electron microscopy results tion in the zirconia samples. Detailed fi nal hold. For comparison, the hold in show that the fi nal grain sizes of these data and characterization explanations the one-step process was adjusted to samples are 55 nm on average (Figure are available in Reference 16. match the overall experiment time of a 8a,b). This is a clear example of how the When comparing the absorption data two-step process with a 10 min. hold benefi cial characteristics of CAPAD can of the reduced polycrystalline sample (17 min. from start to fi nish). For an in- lead to fully dense samples with little to (as processed) with that of a reduced depth discussion of these processes see no grain growth, based on a reported single crystal, the data is strikingly Reference 15. starting crystal size of 50 nm. similar (Figure 11). Likewise, when the Each of the two-step samples as well To quantitatively characterize the as the one-step 40 kN sample were two-step samples, the total transmission found to be >99.9% dense through the and refl ection were measured. Figure 9 Archimedes method. These samples shows how the absorption coeffi cients showed transparency and translucency, increase dramatically as the fi nal hold respectively. The one step 30 kN sample time is increased and decrease as wave- was found to be 99.5% dense and was length. To determine whether oxygen observed to be opaque. This difference vacancies or their associated defects— in optical properties is most likely due caused by the reducing atmosphere of to porosity as this has been shown to be the CAPAD—are responsible for the the dominant mechanism in light inter- observed change in absorption a sample a 500 μm action with ceramics.18,19 Each sample was annealed in air at 750C for 2, 4, 8, was a cylinder 19 mm in diameter and 16, and 24 hour periods. Transmission,

b 500 μm Figure 8. SEM micrographs of (a) pow- der and (b) fracture surface. (a) Micro- graph showing the commercial powder morphology of 8YSZ; (b) Micrograph of fracture surface of fully dense mate- Figure 6. Plot of the rial densifi ed using 2-step process. The coercivity for 8 nm sample has an average grain size of 55 and 40 nm derived nm as is characteristic of all the 8YSZ samples. samples processed in this manner. Re- printed with permission from S.R. Ca- solco et al.15

Vol. 62 No. 2 • JOM www.tms.org/jom.html 61 non-reduced polycrystalline sample is materials like transparent polycrystal- this process are still not fully under- compared to the non-reduced single line 8YSZ and alumina21 suggest that stood, which means the full versatility crystal sample,20 the data is still in good many other wide band gap materials of the CAPAD process has yet to be agreement. that have functional polycrystalline realized. Property studies like the ones characteristics may have the potential presented here are excellent indicators CONCLUSION to be made transparent. The strategies of the potential of this process. The CAPAD technique is capable shown here can potentially be used to ACKNOWLEDGEMENTS of consolidating powders to very high tailor other material properties. Ther- densities through simultaneous ap- mal22 and electrical behavior23 can be Support of this work by the Army Re- plication of both pressure and current dramatically affected by either grain search Offi ce and the Air Force Offi ce and imparting new properties onto tra- size or grain boundary stoichiometry. of Scientifi c Research is gratefully ac- ditional materials by controlling the In addition, the electrical environment knowledged. nanostructure of the bulk material. The of grain boundaries can affect impor- control of both nanostructure and com- tant characteristics like ionic mobility References 24 position in iron oxide nanocomposites and stoichiometric homogeneity of 1. U. Anselmi-Tamburini, J.E. Garay, and Z.A. Munir, opens the door to further advances in the bulk materials. The effects of the Mater. Sci. Eng. A, 407 (2005), p. 24. magnetic materials. The studies of fundamental components involved in 2. J.E. Garay, U. Anselmi-Tamburini, and Z.A. Munir, Acta Mater., 51 (2003), p. 4487. 3. U. Anselmi-Tamburini et al., J. Mater. Res., 19 (2004), p. 3255. 4. G. Zhan et al., J. Am. Ceram. Soc., 86 (2003), p. 200. 5. L. Mitoseriu et al., Appl. Phys. Lett., 84 (2004), p. 84. 6. U. Anselmi-Tamburini, J.E. Garay, and Z.A. Munir, Scripta Mater., 54 (2006), p. 823. 7. J.E. Garay, Annual Reviews of Material Research (to Figure 9. Wavelength dependence be published 2010). of the Absorption Coeffi cient for 8. J.R. Morales et al., Appl. Phys. Lett., 93 (2008), samples processed using the 2- 02254. step procedure with varying hold 9. J.R. Morales et al., Appl. Phys. Lett., 96 (2010) times (10, 11, and 12 minute). 013102. Reprinted with permission from 10. U. Schwertmann and R.M. Cornell, The Iron J.E. Alaniz et al.16 Oxides: Structure, Properties, Occurences, and Uses. 2nd ed. (Weinheim, Germany and New York: VCH, 1991), p. 26. 11. D.J. Dunlop and Ö. Özdemir, Rock Magnetism: Fundamentals and Frontiers. Cambridge Studies in Magnetism (Cambridge, U.K. and New York: Cambridge University Press, 1997). 12. F. Walz, J. Physics-Condensed Matter, 14 (12) (2002), pp. R285–R340. 13. L. Ji-Guang et al., J. Am. Ceram. Soc., 83 (2000), Figure 10. Absorption coeffi cient p. 961. vs. annealing time for a sample 14. X. Su et al., J. Mater. Sci., 39 (2004), p. 6257. produced using the 2-step 15. S.R. Casolco, J. Xu, and J.E. Garay, Scripta Mater., process and a 10 min. hold. All 58 (2008), pp. 516–519. data was taken using 633 nm 16. J.E. Alaniz et al., Opt. Mater. (2009), doi:10.1016/ light. The annealing was done j.optmat.2009.06.004 17. C.G. Levi, Curr. Opin. Solid Stat. Mater. Sci., 8 at 750C in air. Reprinted with permission from J.E. Alaniz et (2004), p. 77. al.16 18. R. Apetz and M.P.B. van Bruggen, J. American Ceramic Society, 86 (2003), p. 480. 19. U. Anselmi-Tamburini, J.N. Woolman, and Z.A. Munir, Advanced Functional Materials, 17 (2007), p. 3267. 20. B. Savioni et al., Physical Review B, 57 (1998), p. 13439. 21. R.L. Coble, U.S. patent, 3,026,210 (20 March 1962). 22. R. Berman, Proceedings of the Physical Society, A Figure 11. Comparison of the 65 (1952), p. 1029. absorption coeffi cient of as-pro- 23. V.I. Barbashov et al., Physics of the Solid State, 50 duced and reoxidized samples. (2008), pp. 2261–2262. Both samples were produced us- 24. Gianguido Baldinozzi et al., Mater. Res. Soc. Symp. ing the 2-step procedure (10 min. Proc. Vol. 1122 (Warrendale, PA: Materials Research hold time). The re-oxidized sample Society, 2009). was annealed at 750C for over 24 hours. Reprinted with permission J.E. Alaniz, J.R. Morales, and J.E. Garay are with the from J.E. Alaniz et al.16 Department of Mechanical Engineering, Materials Science and Engineering Program, University of California, Riverside. Prof. Garay can be reached at [email protected].

62 www.tms.org/jom.html JOM • February 2010 Nanocomposite Materials Research Summary Nanostructured Metal Composites Reinforced with Fullerenes

Francisco C. Robles-Hernández and H.A. Calderon

This work presents the results of the means of: reaction milling, thixocast- sial;1,13–18 some authors argue that when characterization of nanostructured Al ing, rheocasting, metal infi ltration, and using Al the carbon nanoparticles tend or Fe matrix composites reinforced electromagnetic stirring.9–12 The use of to form carbides, but when CNT are with fullerenes. The fullerene used is a metallic matrices and carbon nanoparti- added in small amounts (<2%) they act mix of 15 wt.%C60, 5 wt.%C70, and 80 cles to produce composites is controver- as effective reinforcements with limited wt.% soot that is the product of the pri- to no interaction with the metallic ma- mary synthesis of C60. The composites How would you… trix. For instance, it has been observed were produced by mechanical alloying …describe the overall signifi cance that mechanical properties improve with and sintered by spark plasma sintering of this paper? the formation of carbides (e.g., Al4C3) (SPS). It was found that in both compos- Mechanical alloying and spark in metallic matrix composites; unfortu- plasma sintering (SPS) are ideal ites, C60 withstands mechanical alloying, nately this decreases electrical conduc- and acts as a control agent, reducing the methods to produce nanostructured tivity. Al and Fe matrix composites agglomeration of the particles. In both reinforced with carbon nanoparticles. Mechanical alloying is a method that composite systems the as-mechanically Al and Fe metallic matrices was originally proposed by Gilman and alloyed powders as well as the SPS have different effects on carbon Benjamin in the 1960s to manufacture sintered products are nanostructured. nanoparticles. For instance, Al high-temperature nanostructured ma- sponsors the formation of carbides 11,12 During the SPS process the effect of the while Fe promotes an in-situ synthesis terials at room temperature. Using metal (Al or Fe) matrix with the fuller- mechanical alloying it is possible to of C60. The synthesis of the C60 is ene is different for each composite. For presumably assisted by the catalytic prevent or promote chemical reactions instance, Al reacts with all the carbon in nature of Fe and the electric fi eld by controlling temperature, intensity, generated during the SPS sintering the fullerene mix and forms Al C ; on the and time and in all cases the fi nal prod- 4 3 process. contrary, in the Fe-fullerene composite, uct is usually highly homogeneous and …describe this work to a Fe sponsors the synthesis of C during nanostructured.19 The ideal conditions 60 materials science and engineering the SPS process. The synthesis of the C60 professional with no experience in to preserve the nanostructured nature of is presumably assisted by the catalytic your technical specialty? mechanically alloyed powders include nature of Fe and the electric fi eld gener- This work describes conventional low temperature and short times, which ated during the SPS sintering process. methodologies to produce advanced are the typical operating conditions of metal matrix composites for spark plasma sintering (SPS).8 INTRODUCTION various applications. For instance, mechanical alloying is a well known In the present work mechanical alloy- The successful synthesis of fullerenes method and widely used and SPS ing and SPS are combined to produce and carbon nanotubes (CNT) is possible offers the potential to sinterize nanostructured composites of Al, Fe, by different methods and has resulted in powders at low temperatures and and fullerene. The characterization is pressures in relatively short times, 1–4 diverse structures. The most common preserving the nanostructured nature conducted by means of scanning elec- structures (buckyballs) are C60 and C70 of the composites. The main novelty tron microscopy (SEM), x-ray diffrac- that are composed of atomic arrange- in our work is the in-situ synthesis of tion (XRD), and transmission electron C during the SPS process. ments of carbon with pentagons and 60 microscopy (TEM). In addition, hard- hexagons and heptagons.2,5 Convention- …describe this work to a ness and density measurements have ally, the fullerenes are synthesized and layperson? been performed on the sintered compos- enriched by the Krätschmer method,6 In this paper we present a new ites. Results are reported and discussed method to sinterize fullerene (C60) and fullerenes have been widely used in-situ during sintering (SPS). This accordingly. as reinforcements for composites.7,8 is a novel approach to produce See the sidebar for experimental pro- The dispersion of carbides (e.g., Al C ) nanostructured composites with cedures. 4 3 nanoparticles of C embedded as in Al matrices is challenging, some at- 60 reinforcements. Furthermore, the use RESULTS tempts to increase the homogeneity and of Al of Fe has different effects with cohesion among the matrix and the re- Figure 1 shows SEM micrographs C60 during the SPS process. inforcements has been conducted by of the original and mechanically al-

Vol. 62 No. 2 • JOM www.tms.org/jom.html 63 EXPERIMENTAL PROCEDURES ene phase that has been previously reported as tetragonal C .13 In Figure μ 60 Commercial powders of iron (Fe; 99.9% purity and a particle size <100 m), aluminum 2b a dark fi eld of a fullerene particle (Al, 99.9 at.% purity and a particle size <180 μm), and fullerene mix (15.5at.%C + 60 is presented which shows that the C 4.5at.%C + 80at.%C ) are used as starting materials. The nominal composition of the 60 70 Soot in the fullerene mix has a face-centered composite is Al or Fe and 15.7at.%C. The C present in the fullerene mix is approximately 60 cubic structure and the nanostructured 2.35 at.%C that is equivalent to or 0.04 mol.%C . 60 nature of C . The calculated lattice pa- Mechanical alloying has been conducted at different times up to and including 100 h, 60 but only the results of 100 h of mechanical alloying are presented here. The powder-to- rameters for C60 are: 1.44 Å and 1.43 ball ratio used in the mechanical alloying process is 100:1 without additions of any control Å as per the results of XRD and TEM, agents. The manipulation of the powders is conducted under controlled atmosphere (Ar). respectively. The XRD calculated lat-

The milled powders are sintered by SPS at 773 K under 100 MPa for 600 s at a heating rate tice parameter for C70 is 1.92 Å. These of 3.3 K/s. The SPS sintered samples are 13 mm in diameter and 5 mm in thickness. values are within 1% different from the X-ray diffraction (XRD) is conducted on a SIEMENS D5000 apparatus using a Cu tube theoretical lattice parameters.

with a characteristic Kα wavelength of 0.15406 nm. The scanning electron microscopy Figure 3a and b presents the XRD (SEM) observations are carried out on a JEOL JSFM35CF microscope operated at 20 patterns for the mixes (before mechani- kV. The SEM samples in powder form are prepared by dispersing the powders onto cal milling) of fullerene with the re- a graphite tape. Transmission electron microscopy (TEM) is conducted on a JEOL spective metallic (Al or Fe). Figure 3c JEM200FXII operated at 200 kV. For TEM, the samples are prepared by dispersing the and d shows the XRD matrix spectra powders on ethanol to form a dilute suspension from which an aliquot is taken and is after mechanical alloying. The limited deposited on a Cu-graphite 300 mesh. The as sintered products are surface polished for amount of C present in the original SEM, and electropolished on a double jet electropolisher for TEM. The electropolishing 60 fullerene mix makes its identifi cation, is conducted at 230 K using an electrolyte solution of 25 vol.% HNO3 + 75 vol.%. The density measurements are conducted by the Archimedes method and the rule of mixes. by XRD, challenging; for this reason, The density of fullerene is as per recommended by the manufacturer. The Vickers micro- the intensity of the Al-fullerene sys- hardness is conducted following the ASTM E92-82 standard procedures for a load of tems (before and after mechanical mill- 25 g and a time of 10 s. ing) is presented in logarithmic scale. The logarithmic scale helps reduce the contrast among the refl ections between loyed powders. Figure 1a and b shows particles with a normal particle size Al and C60 and enhances the C60 refl ec- the typical spray cast structure of the distribution and irregular shapes. The tions, allowing its identifi cation. Un- original powders, and Figure 1c shows limited agglomeration of the alloyed fortunately, this process increases the the fullerene powders synthesized by powders is attributed to the control level of noise in the XRD pattern. The evaporation methods with its typical agent effect of fullerene. logarithmic scale treatment is not pos- “fl uffy-like” appearance. Figure 1d and Figure 2a shows the XRD pattern for sible for the Fe-fullerene system due to e presents the mechanically alloyed the as-synthesized fullerene showing the relatively high amount of noise in powders for the Al-fullerene and Fe- the presence of C60 and C70. As per the the background complicating the iden- fullerene composites, respectively. The relative intensity of the refl ections it tifi cation of the C60 refl ections in this mechanical alloying of the Al-fullerene is clear that C60 is more abundant than system (Figure 3b and d). and Fe-fullerene original mixes results C70. The arrow in the XRD pattern in- Figure 4 shows XRD results for the in the production of nanostructured dicates the presence of another fuller- mechanically alloyed powders and SPS

Figure 1. (a–c) SEM micrographs of the origi- nal and (d, e) mechanically alloyed powders of (a) aluminum, (b) iron, (c) fullerene, (d) Al- fullerene composite, and (d) Fe-fullerene composite, respectively.

64 www.tms.org/jom.html JOM • February 2010 of the aluminum matrix. On the other hand, the iron matrix in the Fe-fuller- ene composite does not seem to be af- fected by mechanical alloying (Figure 3). Figure 4 shows the different effects of the metal (Al or Fe) matrices react- ing with the fullerene mix during the SPS sintering. For example, in the Al Relative Intensity (a.u.) Relative matrix composite all the carbon pres- ent in the fullerene mix reacts with Al

to form Al4C3. On the other hand, the 10 20 30 intensity of the (111)C60 refl ection is 2θ degrees intensifi ed considerably during SPS of Figure 2. (a) X-ray diffraction pattern of the as synthesized fullerene (arrow indentifying a the Fe-fullerene composite. This result refl ection of tetragonal fullerene), and (b) TEM dark fi eld and respective micro-diffraction indicates that C is synthesized in-situ selected area diffraction pattern (SADP), and simulated pattern. Location, plane and 60 direction (g) for the SADP are indicated. during the SPS sintering process. Us- ing the XRD results from Figures 3b sintered samples. Comparing Figure 3 by mechanical alloying. Nonetheless, and 4b it is determined that the amount and Figure 4 it can be concluded that in the case of the Al-fullerene com- of carbon in the form of C60 is approxi- the XRD results show no evidence of posite the Al refl ections become wider mately 8.2 at.%C that is equivalent to chemical reactions among the metallic and proportionally smaller indicating a 0.14 mol.%C60. The above indicates matrices and the reinforcement induced potential reduction in the crystal size that the amount of C60 present in the

Figure 3. X-ray diffraction patterns for the (a) Al-fullerene and (b) Fe-fullerene systems, the mechanically alloyed (c) Al-fullerene and

(d) Fe-fullerene composites. Note the absence of the C60 refl ections in the Fe-fullerene system that is attributed to the low amount of C60 in the original mixes as well as the low amount of C60 present in the fullerene mix.

Vol. 62 No. 2 • JOM www.tms.org/jom.html 65 Figure 4. XRD difractograms of the sintered products for the (a) Al-fullerene and (b) Fe-fullerene systems. sintered Fe-fullerene sample is ap- present in the Fe-fullerene composite is in Figure 6a and b. The composition of proximately 3.5 times larger than in the smaller than the ones present in the Al- the Al4C3 particles has been verifi ed by original samples (2.35 at.%C or 0.04 fullerene composite. means of SEM-EDS and the results are mol.%C60) before SPS. The complete reaction of the carbon presented in Figure 7.

Figure 5 shows TEM dark fi eld im- and Al forming the Al4C3 carbide in- Measurements of Vickers micro- ages and selected area diffraction pat- dicates that the high affi nity of carbon hardness show that the hardness of the terns (SADP) of the mechanically al- for Al during SPS process is not lim- Al-fullerene composite increases to 291 μ loyed powders of Al-fullerene and Fe- ited to C60 or C70, but it also includes HV25 that is approximately 6 times the fullerene composites. The “g” identi- the carbon in the soot (graphene-like micro-hardness of pure aluminum (50 μ fi es the corresponding refl ecting planes arrangements). Tentatively, the affi nity HV25). The Fe-fullerene composite μ utilized to produce the dark fi elds. In of carbon in the soot for Al is due to the micro-hardness increases to 722 HV25 the TEM micrographs, it can be ob- large surface area characteristics of the which represents an increase of more served a high level of homogeneity for graphene-like structures present in the than 700% when compared to pure Fe μ both metal (Al or Fe) matrix compos- carbon soot allowing a rapid reaction (100 HV25). The increments in Vick- ites. This demonstrates that mechani- between carbon and Al and leading to ers micro-hardness for the Al-fullerene cal alloying has the ability to produce the complete transformation of carbon composite are attributed to the pres- homogeneous and nanostructured Al into Al4C3 at high homologous tem- ence of Al4C3. The results of densifi ca- ≈ and Fe matrix composites reinforced perature (T/Tm 0.8). Figure 6 shows tion conducted on the as-SPS sintered with fullerene. The bright zones on the TEM images of the sintered products samples indicate that both composites TEM dark fi elds (Figure 5) are crystals for both composites. It is important to reach an approximate density of 82% of C60 identifi ed in either the Al-fuller- notice that for the Al-fullerene com- with respect to their corresponding ene or the Fe-fullerene composites. It posite some particles are pure Al and theoretical values of 2.58 g/cm3 and 3 is of interest that the grain size of C60 other particles are pure Al4C3 as seen 6.91 g/cm for the Al-fullerene and Fe-

Figure 5. TEM dark fi eld micro- graphs and their respective SADP after mechanical alloying for 100 h for the (a) Al-fullerene and (b) Fe-fullerene composites, “g” indi- cates the crystallographic plane used to produce the respective dark fi elds; the bright zones in the

dark fi elds are crystallites of C60.

66 www.tms.org/jom.html JOM • February 2010 fullerene composites, respectively. The fullerene composites, it is evident that further confi rmed by TEM means. This calculated density for the Al-fuller- the crystallites of C60 are smaller in the type of synthesis is attributed to the ene composite is 2.12 g/cm3 and 5.66 Fe-fullerene (Figure 5a and b). In addi- presence of the electric fi eld generated g/cm3 for the Fe-fullerene compos- tion, this clearly demonstrates that me- by the SPS and the catalytic nature of ite. This densifi cation seems smaller chanical alloying is a promising method Fe to synthesize carbon nanoparticles 21–24 when compared to values reported by to produce a nanostructured composite such as C60 or nanotubes. The sin- other authors for composites sintered for both Al-fullerene and Fe-fullerene tering temperature, 773 K, is relatively by SPS;14–16 relatively short densifi ca- systems. low for the Fe-fullerene composite (T/ ≈ tion times were chosen for the current The most important result in the Tm 0.42) which results in a lower af- work, since the intention was to prevent present work is the interaction of the fi nity for carbon when compared to Al coarsening, and preserve the nano- different metallic matrices when used at the same temperature. The amount scale microstructural features. Further in their respective composites with the of Fe3C formed during the SPS process work will be necessary to ascertain the fullerene mix. It is evident that in the is relatively small which makes XRD optimum processing conditions (tem- Al-fullerene composite during the SPS phase identifi cation diffi cult, but the perature, pressure, time) for achieving process the chemical reaction between Fe3C is clearly identifi ed by TEM. maximum density whilst preserving Al and the carbon present in the fuller- The measured densifi cation, with re- the benefi cial nanoscale features. ene mix gives rise to the formation of spect to the theoretical is relatively low; Al C . On the other hand, in the Fe- however, this is due to the temperature, DISCUSSION 4 3 fullerene composite we have identifi ed pressure, and time used during the

Mechanical alloying effects on the an increase in the intensity of the C60 re- SPS sintering process. The conditions constituents of the Al-fullerene and Fe- fl ection. The intensity of the peak is ap- used during the SPS were intended to fullerene composites are almost neg- proximately 3.5 times higher, and rep- prevent grain coarsening. In addition, ligible. No conclusions can be drawn resents an increase from 0.06 mol.%C60 these conditions slow down reactions from the XRD results regarding the ef- to 0.14 mol.%C60 in the amount of C60 between Fe and the carbon present in fects of mechanical alloying on fuller- present in the sintered product. the fullerene mix, thereby promoting enes for the Fe-fullerene composite The increase in the intensity of the C60 the nucleation and growth of C60 dur-

(i.e., compare Figure 3b and d). How- refl ection (111)C60 can be an indication ing the SPS process without excessive ever, by observing the respective TEM of the in-situ synthesis of C60 during the coarsening. The low homologous tem- dark fi eld for the Al-fullerene and Fe- SPS process. The in-situ synthesis is perature may also be contributing to the

Figure 6. Dark fi elds and respective selected area diffraction patterns for the (a, b) Al- fullerene and (c–e) Fe-fullerene composites.

(a) Al matrix, (b) Al4C3, (c) Fe matrix, (d) fullerene, and (e) iron carbide (Fe3C).

Vol. 62 No. 2 • JOM www.tms.org/jom.html 67 11. J.S. Benjamin, New Materials by Mechanical Alloy- Al ing Techniques, ed. E. Arzt and L. Schultz (Oberursel, Germany: DGM Informationgesellschaft, 1989), pp. C 3–18. 12. C. Suryanarayana, Prog. in Mater. Sci., 46 (2001), pp. 1–184. 13. M. Umemoto et al., Mater. Sci. Forum, 312-314 0 246810 (1999), pp. 93–102. Energy keV 14. I. Wei Chen and X.H. Wang, Nature, 404 (2000), pp. 168–171. Figure 7. SEM micro- 15. R.L. Coble, J. Appl. Phys., 32 (1961), pp. 787–799. graph of the Al-fullerene 16. R.L. Coble and J.E. Burke, Progress in Ceramic composite after the SPS Science, 3rd edition, ed. J.E. Burke (New York: Wiley- sintering and respective American Ceramic Society, 1963), pp. 197–251. EDX spectrum from the 17. R. Pérez-Bustamante et al., Mater. Sci. and Eng. A, 502 (2009), pp. 159–163. Al4C3 particle indicated with the arrow. 18. M. Oda et al., “Infl uence of Dispersed Carbon Nano-Fibers/Carbon Nano-Tubes in Al Matric Com- posite” (Paper presented at the TMS 2009 Annual Meeting and Exhibition, San Francisco, CA, 15–19 References February 2009). lower reactivity between fullerene and 19. J. Guerrero-Paz et al., Mater. Sci. Forum, 360-362 the Fe matrix. Therefore, using higher 1. P.J.F. Harris, Carbon Nanotubes and Related (2001), pp. 317–322. temperatures and/or longer times may Structures (Cambridge; Cambridge University Press. 20. T. Kuzumaki et al., J. Mater. Res., 13 (1998), pp. 1999). 2445–2449. lead to higher densifi cation, but will 2. H.W. Kroto et al., Nature, 318 (1985), p. 162. 21. J.H. Hafner, Chem. Phys. Lett., 296 (1998), pp. likely result in coarsening of the grain 3. S. Iijima, Nature, 354 (1991), pp. 56–58. 195–202. and for the Fe-fullerene composite this 4. D. Ugarte, Nature, 359 (1992), pp. 707–709. 22. H. Dai et al., Chem. Phys. Lett., 260 (1996), p. 5. H. Terrones and M. Terrones, J. Phys. Chem. Solids, 471. may even preclude the desirable in-situ 58 (1997), pp. 1789–1796. 23. T. Gou et al., Chem. Phys. Lett., 243 (1995), p. 49. 6. W. Krätschmer et al., Nature, 347 (1990), pp. 354– synthesis of C60. 24. W. Zhou et al., Chem. Phys. Lett., 350 (2001), p. 6. 358. CONCLUSIONS 7. F.C. Robles Hernández (M.Sc. thesis, Instituto Po- Francisco C. Robles-Hernández is assistant pro- litecnico Nacional, Mexico, 1999). fessor at the University of Houston, 304 Technol- In the Al-fullerene composite, the 8. V. Garibay-Febles et al., Mater. and Manufac. Pro- ogy Building, Houston, TX 77057; H.A. Calderon, high affi nity of Al for carbon togeth- cesses, 15 (2000), p. 547. professor, is with the Departamento de Ciencia 9. F. Adams and C. Vivès, Netherlands Society for Ma- de Materiales, ESFM-IPN, Mexico DF 07738. Prof. er with the SPS sintering temperature terials Science, 1 (1997), p. 337. Robles-Hernández can be reached at fcrobles@ promoted their complete reaction and 10. D.M. Hulbert et al., Mater. Sci. and Eng. A, 488 uh.edu, and Prof. Calderon can be reached at (2008), pp. 333–338. [email protected]. their transformation into Al4C3. Al- ternatively, the combined effect of the electric fi eld, relatively low SPS sinter- ing temperature, and catalytic nature of Fe resulted in the in-situ synthesis of C60 which is a quite unique fi nding. The level of reactivity of the various Visit the Knowledge Resource Center to reserve your copy today! constituents of the fullerene mix used Modeling of Casting, Welding, and Advanced in this work is uncertain, as is the type Solidifi cation Processes XII (MCWASP XII) of carbon (C60, C70, or soot) that reacts by Steve L. Cockcroft and Daan M. Maijer with Fe and forms Fe C. However, it is 3 These proceedings encapsulate the progress made in mod- conjectured that the C that has been 60 eling casting processes and liquid-solid phase transforma- synthesized in-situ during the SPS sin- tions over the years. The papers presented provide a clear tering process is obtained from the soot picture of the modeling activities for processes involving (i.e., graphene-like structures) present solidifi cation, covering continuous casting, welding, remelt- in the fullerene mix. ing, semi-solid, and shape and ingot casting processes. ACKNOWLEDGEMENTS Other topics related to casting technologies, such as liquid metal treatment and casting tools (e.g., mold manufacturing FCRH would like to express his grat- and life time), are also examined. These proceedings bring itude to the University of Houston and together leading academic research that brings up to date the government of Texas for the start- information from the industry. Classical studies of structure up package funding. The authors would and segregation formed upon solidifi cation as well as new in-situ observation techniques are addressed. like to thank Drs. M. Umemoto and V. Garibay Febles for their outstand- A collection of papers from the 12th International Conference ing support with mechanical alloying held in Vancouver, British Columbia June 7th-14th, 2009. and spark plasma sintering methods. TMS Member price $169; Non-member price $244; CONACYT and SIP-COFAA-IPN are TMS Student Member price $135 acknowledged for fi nancial support through grants 45887-Y, 28952U, 58133 To order these or related publications, contact TMS: (both authors), and SIP-20080337. E-mail [email protected] • Phone (724) 776-9000, ext. 256 • Fax (724) 776-3770 Bulk Metallic Glasses Commentary Bulk Metallic Glasses: Overcoming the Challenges to Widespread Applications

Peter K. Liaw, Gongyao Wang, and Judy Schneider

Since an amorphous solid was fi rst Second, size effects in BMGs are in- the atomic level and the microscopic synthesized by rapidly quenching with teresting.9 The mechanical behaviors of mechanisms that control the properties cooling rates up to 106 K/s,1 the for- metallic glasses during bending, com- of metallic glasses. mation, structure, and property stud- pression, tensile, impact, and fatigue In “Mechanical Response of Metallic ies of metallic glasses have attracted exhibit variations with the specimen Glasses: Insights from In-situ High En- increasing attention because of their geometry.10,11 Thus, the constraint of ergy X-ray Diffraction,” Mihai Stoica fundamental scientifi c importance and shear-band formation and deformation et al. review the use of the synchrotron engineering application potentials. processes is critical to elucidate these radiation to evaluate the elastic-plastic New multicomponent glassy-alloy sys- phenomena. Third, understanding of the response of BMGs. tems with lower critical cooling rates, fatigue behavior of BMGs is limited.12Ð14 “Amorphous Metals for Hard-tissue such as Zr-Cu-Ni, Zr-Cu-Ni-Al, and BMGs can exhibit a wide range of fa- Prosthesis” by Marios D. Demetriou et Zr-Ti-Cu-Ni-Be were discovered in the tigue properties. Moreover, the charac- al. reviews the performance of amor- 1990s.2Ð4 Now amorphous alloys can teristic for the formation of shear bands phous alloys relevant to hard-tissue be fabricated into bulk forms ranging is unknown during cyclic deformation prosthesis. from several millimeters to several cen- of metallic glasses, which raises addi- “Metallic Glasses: Gaining Plasticity timeters in thickness. Therefore, “bulk tional questions: When and how will for Microsystems” by Yong Yang et al. metallic glasses” (BMGs) is the name shear bands form? How does the fatigue summarizes the experimental fi ndings commonly used for these glassy met- crack initiate and propagate in metallic related to size effects on plasticity in als. In addition, BMG composites were glasses? Another challenge is the fab- metallic glasses. developed by introducing specifi c crys- rication of high-quality and large-size References talline phases in an amorphous matrix samples. Some good glass-forming al- 1. W. Klement, R. H. Willens, and P. Duwez, Nature, 187 to improve plasticity and toughness, loys involve the use of elements that are (1960), p. 869. which could broaden BMG applica- either expensive or diffi cult to handle 2. A. Inoue, Acta Mater., 48 (2000), p. 279. tions.5,6 safely (such as beryllium3). However, 3. A. Peker and W.L. Johnson, Applied Physics Letters, 63 (1993), p. 2342. In general, BMGs exhibit many de- with the development of alloys that are 4. M.K. Miller and P.K. Liaw, editors, Bulk Metallic sirable properties: high strength, high based on less expensive elements and Glasses (New York: Springer, 2007). hardness, excellent elasticity, low coef- more tolerant of impurities, the cost of 5. D.C. Hofmann et al., Nature, 451 (2008), p. 1085. 6. J.W. Qiao et al., Applied Physics Letters, 94 (2009), fi cients of friction, near-net-shape cast- BMGs will continue to decline relative 151905. ing, and almost perfect as-cast surfaces.4 to other high-performance alloys,4 and 7. W.J. Wright, R. Saha, and W.D. Nix, Mater. Trans., 42 As excellent candidates for structural high-quality and large-size samples (2001), p. 642. 8. J. Schroers and W.L. Johnson, Phys. Rev. Lett., 93 materials, BMGs have been applied to will become available. Finally, applica- (2004), 255506. new industrial products requiring high tions of BMGs are very few. Thus, how 9. R.D. Conner et al., Acta Materialia, 52 (2004), p. performance, such as microgears, cas- to fi nd new applications for BMGs is 2429. 10. F.X. Liu et al., Intermetallics, 14 (2006), p. 1014. ings for electronic products, and golf- pertinent for continued research activi- 11. H. Guo et al., Nature Materials, 6 (2007), p. 735. club heads.4 To apply BMGs extensive- ties of BMGs. 12. C.J. Gilbert, V. Schroeder, and R.O. Ritchie, ly, critical issues need to be solved. This JOM topic on BMGs presents Metallurgical and Materials Transactions A, 30A (1999), p. 1739. First, BMGs usually lack ductility. discussions and developments related 13. B.C. Menzel and R.H. Dauskardt, Acta Mater., 54 The deformation and fracture behav- to the above questions. Due to the ex- (2006), p. 935. iors of BMGs at room temperature are tent of the information available, four 14. G.Y. Wang et al., Mater. Sci. and Eng., 494 (2008), controlled by the initiation and propa- articles will be presented in this issue; p. 314. gation of shear bands, which result in another three will be published in April. Peter K. Liaw and Gongyao Wang are with the De- partment of Materials Science and Engineering, the limited global deformation of about The following papers are in this issue. The University of Tennessee, Knoxville, TN 37996; 2%, particularly under tensile defor- “Understanding the Properties and and Judy Schneider is with the Mechanical Engi- mations.7 It is important to understand Structure of Metallic Glasses at the neering Department, Mississippi State University, MS. Dr. Schneider is the JOM advisor from the Me- the relationship between the ductility Atomic Level” by T. Egami reviews chanical Behavior of Materials Committee of the and atomic microstructure in BMGs. the behavior of glasses and liquids at TMS Structural Materials Division.

Vol. 62 No. 2 • JOM www.tms.org/jom.html 69 Bulk Metallic Glasses Research Summary Understanding the Properties and Structure of Metallic Glasses at the Atomic Level

T. Egami

Liquids and glasses have been well fully appreciated by the materials com- known to human kind for millennia. How would you… munity. This article discusses why this And yet major mysteries remain in the …describe the overall signifi cance problem is so diffi cult, and introduces of this paper? behavior of glasses and liquids at the alternative approaches that may hope- This paper provides a “big picture” atomic level, and identifying the mi- of the science of metallic glasses at fully lead to better understanding. croscopic mechanisms that control the the atomic level. It points out that the DIFFICULTIES properties of glasses is one of the most structure and properties of metallic challenging unsolved problems in phys- glasses are discussed often implicitly Computer Simulation assuming a hard-sphere model, even ical sciences. For this reason, applying though real atoms are far from hard- Today, fi rst-principles calculations simplistic approaches to explain the spheres. It then discusses serious based upon quantum mechanics and the behavior of metallic glasses can lead problems associated with such an approach, and suggests a more density functional theory (DFT) are so to serious errors. On the other hand balanced approach based on local advanced that many properties of crys- because metallic glasses are atomic structural fl uctuations. talline materials are well understood at glasses with relatively simple structure, …describe this work to a a most fundamental level, and it is even they may offer better opportunities to materials science and engineering possible to design new materials us- advance our fundamental understand- professional with no experience in ing theoretical calculations.3 When the ing on the nature of the glass. The dif- your technical specialty? DFT calculation is combined with the fi culties inherent to the problem and Today the structure and properties molecular dynamics (MD) technique it some recent advances are reviewed of crystalline solids can be understood from the fi rst-principles is possible to study liquids. However, here. calculations with impressive there is a major problem of time-scale accuracy. In comparison the science INTRODUCTION of glasses and liquids is in a much when the fi rst-principles MD method is more primitive stage, because even applied to supercooled liquid or glass- Elucidating the properties of materi- describing the atomic structure es. As is widely known, MD simulation als in terms of the atomic, mesoscopic, accurately is a challenging task. uses fi nite time steps to describe the and microscopic structure is one of the However, because it is so challenging temporal evolution. Because the typical primary purposes of materials science. it is a very attractive fi eld for creative and imaginative researchers. frequency of atomic vibration in solids But if we try to proceed with such an ef- Unfortunately at present, simple, and liquids is 1012–13 1/sec, to faithfully fort for glasses and liquids we run into intuitive approaches are all too describe such dynamics the time step a major roadblock right away. Because popular among the experimentalists in the fi eld. This paper discusses the has to be of the order of a femto-second glasses and liquids do not have the lat- problems with such approaches and (10–15 sec). So even when one runs the tice periodicity it is not easy to char- suggests alternative ways to make MD simulation for a million steps, the acterize their structure and dynamics at progress. time scale is only a nano-second (10–9 the atomic level with suffi cient accura- …describe this work to a sec). On the other hand the glass transi- cy. Consequently it is diffi cult to con- layperson? tion is defi ned by the viscosity becom- struct a realistic and meaningful model. Major mysteries remain in the ing 1013 poise, which corresponds to In the absence of effective approaches behavior of glasses and liquids at the atomic level. In general the science of the time scale of 103 seconds. Thus we many researchers resort to either com- non-crystalline materials, including have a gap in the time scale by a factor puter simulations or phenomenological macromolecules such as proteins, of 1012. Even when we use the model models such as the free-volume theory. is less advanced than the science interactomic potential instead of the However, both of these approaches suf- of crystalline materials. The main reason is that we have not developed fi rst-principles calculation, the gap in fer from severe drawbacks. In fact the effective theoretical methods to deal the time scale can be improved only by nature of glass and glass transition is with extensively disordered systems two orders of magnitude or so, leaving always listed among the most diffi cult with strong correlation. This paper a gap of 1010. unsolved problems in theoretical phys- discusses the pitfalls of simple and intuitive approaches, and suggests The consequence of this huge gap ics.1,2 Unfortunately the seriousness alternative approaches. in time scale is that the models created of the problem appears to be less than by MD simulation, either by the fi rst-

70 www.tms.org/jom.html JOM • February 2010 paper, even though they may use this Table I. Factors that Indicate the Hardness Table II. Ratio of the Heat Released (ΔH) of the Potential, for the Hard-sphere model. The most important parameter over the Change in Volume (ΔV) during (HS) Potential, the Lennard-Jones (L-J) in the free-volume theory is the ratio Structural Relaxation of the critical value of the free-volume Potential, and the Johnson Potential Δ Δ Δ Δ (see text for defi nitions) v* H/ V H/ V to accept an atom, , to the atomic (eV/Å) (eV/atom) Ref. volume, v , v*/v . This value is about γ R R′ a a 0.8 for van der Waals liquids, and the Zr55Cu30Al10Ni5 0.416 7.61 22 HS — — 0 theory assumes this value to be close to Zr44Ti11Cu10Ni10Be25 0.461 7.70 23 L-J 2.5 36 0.109 unity. For metallic liquids this value is Pt60Ni15P25 0.460 8.48 24 Johnson 0.874 27.8 0.153 much less than unity, and is close to 0.1 (see Table II of Reference 4). ρ principles calculation or with model The validity of the free-volume packed (DRP) HS system is DRP-HS = potentials, are seriously under-relaxed concept is obvious for systems of 0.63.7 On the other hand the volume because they are effectively quenched hard-spheres (HS). For a hard sphere density of the crystalline close packed with extremely high cooling rates from to move in a close-packed structure system, face-centered cubic (f.c.c.) or ρ high temperatures. Thus the short- there must be extra space between the hexagonal close packed (h.c.p.) is CP range order in the model refl ects that of spheres, otherwise they are trapped in = 0.74. Thus the volume of the HS a high-temperature liquid, not necessar- the cage made by neighbors and cannot system increases by as much as 17% ily that of a glass. The glass transition move. The Lennard-Jones (L-J) poten- at melting. Indeed the volume expan- observed by simulation corresponds to tial that describes van der Waals solids sion at melting is about 15% for argon viscosity reaching only 102 to 103 poise is a hard potential as discussed below. that can be described well by the L-J rather than 1013 poise. Consequently For this reason the free-volume theory potential. However, as shown in Figure the glass transition temperature deter- is successful for organic molecular 1 the volume expansion upon melting, mined by MD simulation is consider- liquids and glasses for which the inter- ΔV/V, is much smaller for metals and ably higher than the real glass transi- molecular interaction can be modeled semimetals, ranging from 0 to 6%, with tion temperature. We still can learn well by the L-J potential. some of them being negative.8 The den- from such simulations, but we should However, if atoms are soft and sity of metallic liquids is much higher ρ remember to use many grains of salt in squeezable they may be able to move than DRP-HS = 0.63, because atoms can interpreting such results. It is simply without free space, simply by squeez- be closer than the average distance that naïve to believe that the MD models ing in. Thus the relevance of the free- defi nes the atomic size. For instance describe the true state of the glasses. volume concept is less obvious for the density of the DRP model with the systems with softer interatomic poten- Johnson potential for iron9 is 0.74, vir- Free-volume Theory ρ 10 tials, such as metals. That metals are tually identical to CP . These results Another approach frequently adopt- far from the HS systems can be clearly show that the HS model is unrealistic ed by researchers in the fi eld of metal- demonstrated by the density change for metals. lic glasses is the free-volume theory.4–6 upon melting. The volume fraction oc- Hardness of the Potential One of the great mysteries of glasses cupied by spheres in the dense-random- and liquids is that viscosity changes as In order to make this point even much as 15 orders of magnitude over a clearer let us discuss how to evaluate Equations relatively small temperature range just the hardness of the interatomic poten- above the glass transition temperature, 23 tial. Any pair-wise interatomic poten- ÈØÈØrr T . The free-volume theory explains Vr Vr A1 B 1 .... (1) tial V(r) can be expanded around the g 0 ÉÙÉÙ   ÊÚÊÚrr00 this mystery beautifully, by assuming potential minimum at r = r0 (see Equa- that the excess free-volume in the liq- tion 1 in the table). ˜Qln B uid increases linearly with temperature J   (2) For most of the potentials A > 0, B ˜lnV 2A above Tg. However the free-volume < 0. Hardness of the potential may be theory itself does not explain why the described by the ratio, –B/A. This ratio DE 1 DE free-volume increases linearly with V if ijr ij (3) is related to the Grüneisen constant of : Ç temperature. It is a phenomenological i j anharmonicity, as given in Equation 2, theory based upon the experimental ob- where ν is the phonon frequency, which 2A ÈØr 11 servations. DE i1ij rrDE (4) is proportional to A. When the value V ÇÉÙij ij Another, more serious problem for :i j ÊÚU of γ is high the potential is strongly the metallic glass community is that the asymmetric at the bottom of the po- 2 free-volume theory in its original form p kT tential, and repulsion is much stronger E (5) does not work well for metals. This is p 24B than attraction. The value of γ for the clearly stated in the original paper by HS potential is infi nite. The values of 4 γ Cohen and Turnbull. However, unfor- 2 BV T 2 31 Q for the L-J potential and the Johnson kT H , K (6) tunately many young researchers to- gv D potential are shown in Table I. Another KD 21  2Q day apparently do not read such an old indicator of the hardness of the poten-

Vol. 62 No. 2 • JOM www.tms.org/jom.html 71 tial is the curvature to the depth ratio,

R = A/V(r0), and the ratio between the distance at which the potential crosses over zero, V(rc) = 0 and r0, relative to ′ unity, R = 1 – rc/r0. These two factors are closely related. When R is large or R′ is small the potential energy quickly Figure 2. The ratio of the melting tempera- rises as r decreases below r , so that the 0 ture, Tm, to the boiling atom cannot be squeezed much. These temperature, Tb, for rare values for the HS potential, the L-J po- gases, and for metals and semimetals. For tential and the Johnson potential are rare gases other than also shown in Table I. radon the triple-point Now the depth of the potential is re- temperature is used in- stead of T .12 lated to the cohesive energy and the boil- m ing temperature TB, and the curvature A is related to the phonon frequency and the melting temperature, Tm. Therefore the parameter R2 is related to the ratio ed it should exert a major effect on dif- characterize “free-volume,” or negative of Tm/TB. As shown in Figure 2, this ra- fusivity, whereas the observed effect (n-type) density fl uctuations, but then tio is close to unity for van der Waals is much less dramatic.14 A more re- short bonds could describe “anti-free- systems such as radon (other rare gases cent measurement by Klugkist et al.15 volume,” or positive (p-type) density do not show melting under 1 atm.), but underscores this point. This, again, is fl uctuations, regions of high density, much lower for metals and semimetals, because the free-volume in metals is as shown in Figure 4.19–21 Recombina- ranging from 0.3 to 0.6.12 These results much smaller than atoms. tion of these two types of local density again show that van der Waals liquids When the metallic glass is annealed fl uctuations can explain the structural and solids are more diffi cult to squeeze at a temperature below Tg the structure relaxation. and behave like hard-spheres, whereas relaxes toward a more stable state. This During the structural relaxation heat metals are unlike hard-spheres and are phenomenon is commonly observed in is released as the structure relaxes to a much more squeezable and harmonic. all kinds of glasses, and is known as more stable state. The amount of heat structural relaxation.16 It is possible to released, ΔH, and the volume change, Experimental Evidence against observe what is happening to the struc- ΔV, during the relaxation were mea- the Simple Free-volume Concept ture by diffraction measurement.17 Fig- sured by various researchers.22–24 The Δ Δ One of the examples of experimen- ure 3 shows recent data on the change values of Ea = H/ V are shown in tal data Cohen and Turnbull used in in the x-ray atomic pair-density func- Table II, in the units of eV/Å3, and cautioning against the use of the free- tion (PDF).18 It is clear that both long eV/atom. Usually the vacancy forma- volume theory for metals is the obser- and short bonds are eliminated as a tion energy is about 2 eV/atom, but vation by Nachtrieb and Petit13 that result of relaxation. Because the inter- the values shown here are signifi cantly the pressure dependence of diffusiv- atomic potentials of metals are more larger, by a factor of four. The value of ity is much less than expected for the harmonic, when the structure is rapidly Ea should also be comparable to the ac- free-volume theory. Because pressure cooled some bonds remain too short or tivation enthalpy for diffusion, which would squeeze out free-volume and too long, and these bonds are relaxed is also 2–3 eV.14 This disagreement can make the atomic motion more restrict- out by annealing. Long bonds may be resolved if we assume that during the structural relaxation not only free- volumes (n-type defects) are elimi- nated, decreasing the volume, but also “anti-free-volumes” (p-type defects) are eliminated, increasing the volume. Thus the total change in volume repre- sents the small difference after cancel- lation of the bulk of the positive and Figure 1. Relative negative changes.20,24 Thus the use of change in volume at melting for rare gases the total volume change in evaluating and for metal and ΔH/ΔV results in overestimate. semimetal elements Another very interesting observa- plotted against the atomic number, Z.8 tion is that irradiating a metallic glass The dashed line is for with high energy particles does not a hard-sphere model. seem to result in void formation eas- ily.25,26 It is well known that irradiat- ing crystalline materials with electrons

72 www.tms.org/jom.html JOM • February 2010 is completely frozen. We now have a better understanding of the behavior of the system in the vicinity of this critical density, for instance in terms of various scaling laws.28,29 Thus the properties of the system can now be semi-quantita- tively predicted. Theories of Local Density Figure 3. (Top) The change in the fi rst peak of the x-ray Fluctuation atomic pair-density function However, the HS system is a very (PDF) of Zr52.5Cu17.9Ni14.6Al10 Ti5 due to annealing at 630 artifi cial model, and is pathological in K (40 K below the glass many ways. First, because there is no transition temperature) for 20 min. Bars indicate errors. attraction, the system has to be kept to- (Bottom) The fi rst peak gether by an external pressure, or the of the PDF of the as-cast boundary condition. Its potential en- sample.18 ergy is indeterminate because during the collision of hard spheres the infi - nite potential is experienced for zero time. Thus we cannot even talk about or neutrons produces voids and leads materials. Furthermore, it is the right the virial theorem. As is well known to swelling. If the defects in metal- potential for granular materials, such the lattice dynamics of crystalline sol- lic glasses were free-volumes that are as sands. For these reasons the models ids is very well described in terms of similar to vacancies, after irradiation with the HS potential have been stud- phonons. But in the HS solids there they may accumulate and coagulate as ied extensively. are no phonons unless the system is voids. Apparently that is not the case. A notable recent achievement is the slightly expanded, and even in such a Irradiation may produce both positive development of the jamming theory, case phonons are strongly attenuated and negative local density fl uctuations, which is closely related to the free-vol- and do not propagate well. We have but they will soon recombine without ume theory. At low densities the sys- already discussed how poor this model leaving much trace. For this reason tem of hard spheres are fl uid, but fl uid- is to describe the metallic systems real- metallic glasses are likely to be quite ity is reduced as the density is reduced istically. Perhaps the HS model is too radiation resistant. and free-volume is removed from the far away from the reality of the metal- These results demonstrate that in system. At a critical density the system lic systems, and to use this model as a metals the critical value of free-vol- becomes jammed, as hard spheres are starting point to develop the theory of ume, v*, is much smaller than the interlocked with each other, just as in metallic glasses may be ill-advised af- atomic volume, and free-volumes are the traffi c jam caused by gridlock. As ter all. There may be a better way to accompanied by their mirror state, noted above the density of the HS- start the effort. ρ anti-free-volumes. A more balanced DRP system is DRP = 0.63. Jamming The lattice dynamics theory of a description of the state of liquids and occurs, however, already at a slightly solid starts with the harmonic approxi- ρ 27 30 glasses requires both negative as well lower density of J = 0.53, because mation. Then it should be possible to as positive local density fl uctuations. the movement of spheres becomes very start the theory of liquids in a similar much restricted even before the system way. For instance, in the HS system the ALTERNATIVE APPROACHES

Jamming Theory In spite of the problems with using the HS models for metals, the HS po- tential has many attractive features. It is the simplest potential, with only one parameter. Because there is no energy scale in this potential, all the problems Figure 4. Distribution of the atomic volume strain. Strain larger than 0.11 are reduced to those of geometry and defi nes the “free-volume,” or the n- mathematics rather than physics. Thus type local density defects, whereas defi nite answers for various well-de- strain less than –0.11 defi nes the “anti-free-volume”, or the p-type fi ned problems can be obtained, and local density defects.21 these answers can serve as the yardstick for real materials, even though the HS potential may not be realistic for many

Vol. 62 No. 2 • JOM www.tms.org/jom.html 73 virial, or equipartition, theorem.37,38 This relationship thus forms the basis for the statistical mechanical theory of liquids. Equation 5 extrapolates to zero at T = 0. However, such a state, the ideal state, Figure 5. kTg/2BV for vari- ous bulk metallic glasses, cannot be achieved, because the sys-

where Tg is the glass tran- tem becomes jammed and immobile at sition temperature, B bulk some temperature where becomes modulus and V atomic volume, plotted against small enough. The minimum value of the Poisson’s ratio. The ε the volume strain, v = p/B, was found solid line is the theoretical to be ± 0.06 for homogeneous strain Equation 6.40 and ± 0.11 for local strain (Figure 4), from the topological instability condi- tion.21,40 The derivation of these critical values is too lengthy for this article, and local volume cannot be contracted, but vidual is too often not consistent with the readers are referred to References can only be expanded. In an harmonic reality, which results in mental stresses. 21, 39, and 40. From this condition system, however, the local volume can Stresses can either destroy a weak per- Equation 6 describes the glass transi- be contracted or expanded at the equal son or let a strong person grow, but in tion temperature, where ν is Poisson’s cost in energy. Thus the local density turn, we may be able to characterize the ratio, was deduced. This equation fi ts fl uctuations are equally negative and society by the level of the stress. In a the experimental data with impressive positive. Then the local density fl uctua- similar way the structure of a glass or accuracy as shown in Figure 5.40 This tion may be a good starting point for the a liquid may be characterized by the is probably the fi rst time that the glass theory of liquids. Indeed a large body atomic level stresses. transition temperature was calculated of literature on hydrodynamic theories At the atomic level the stress can be from the fi rst principles virtually with- exists that starts with the local density defi ned by Equation 3,35 where α and out a fi tting parameter. Currently the β α fl uctuations, both negative and posi- are Cartesian coordinates, fij is the theory is being extended to account for tive.31 The mode-coupling theory32,33 α component of the two-body force the mechanical deformation and the ef- β was developed by extending the linear between atoms i and j, and rij is the fects of stresses on the fl ow properties hydrodynamic theory by introducing β component of the distance vector in terms of the intrinsic local fl uctua- non-linear coupling. It has been quite between atoms i and j. In a system of tions in shear stresses, rather than de- successful in describing the dynamic atoms interacting by a pair-wise har- fects. behavior of liquids from high tempera- monic potential (B = 0 in Equation 1), CONCLUSION ture down to the supercooled liquid re- the force is proportional to the devia- gion. However, it is questionable if it tion from the ideal distance, r0. Then, Although the science of crystals is can be extended down to the tempera- from Equation 1 we can derive Equa- at a very advanced stage, the science ture range for the glassy state.34 tion 4. Thus the atomic level stress of glasses and liquids is still languish- describes the distortion of the environ- ing at the level of infancy. The primary Theory of Atomic Level Stresses ment of each atom away from the ideal reason for this is the lack of effective

In the local density fl uctuation theo- state, where all neighbors are at rij = methods of describing the structure and ries liquids are regarded as a continu- r0, which never occurs in reality. The dynamics at the atomic level. Many re- um body with locally varying density, stress is a 3 × 3 tensor, with six inde- searchers working on metallic glasses and atomic discreteness does not enter pendent components. The trace corre- still describe the structure-property the theory directly. It is only obliquely sponds to the pressure, p, whereas the relationship only in terms of free-vol- refl ected in the non-linear coupling of deviatory components describe the fi ve umes, even though the creators of the density-density correlations.33 On the components of shear stresses.35 theory cautioned against using this con- other hand the concept of the atomic- This theory was initially introduced cept for metallic liquids without due level stress was introduced in order to for the purpose of defi ning the elusive modifi cations. In metallic systems the describe the atomic environment re- defects in the glassy structure, but it critical value of free-volume is much alistically.35 The starting point is the was found to have much wider utility. It smaller in volume than atoms, and is realization that in liquids and glasses was successfully used in calculating the only about 10% of the atomic volume. the atoms are not ideally packed as in composition limit of glass formation in The state of a metallic glass is better the closed packed structures, and con- binary glasses.36 It was also found that described in terms of local density fl uc- sequently the atomic environment of the elastic self-energy of the pressure tuations, not only negative fl uctuations each atom is severely distorted. This is is proportional to temperature at high (free-volume), but also positive fl uc- similar to the situation in human soci- temperatures, see Equation 5, where tuations (anti-free-volume). In addition ety. We do not live in an ideal world. B is the bulk modulus, and other fi ve the local shear distortion plays a major The aspiration and desire of each indi- shear components also follow the same role, for instance, in deformation. It is

74 www.tms.org/jom.html JOM • February 2010 about time for us to evolve gradually 910–911. Ball, Phys. Rev. E, 55 (1997), pp. 7203–7211. out of indiscriminate use of the free- 8. D. Wallace, Proc.: Math. Phys. Sci., 433 (1991), pp. 28. C.S. O’Hern et al., Phys. Rev. Lett., 86 (2000), pp. 631–661. 111–114. volume concept, and develop more re- 9. A.R. Johnson, Phys. Rev., 134 (1964), pp. A1329– 29. P. Olsson and S. Teitel, Phys. Rev. Lett., 99 (2007), alistic and accurate microscopic and at- 1336. 178001. omistic description of metallic glasses 10. K. Maeda and S. Takeuchi, J. Phys. F: Met. Phys., 8 30. M. Born and K. Huang, Dynamical Theory of (1978), pp. L283–L288. Crystal Lattices (Oxford: Clarendon Press, 1954). based on local structural fl uctuations. 11. For example, L. A. Girifalco, Statistical Physics of 31. For example, J.P. Hansen and J.R. McDonald, Materials (New York: John Wiley & Sons, 1973). Theory of Simple Liquids (London: Academic Press, ACKNOWLEDGEMENT 12. D.R. Lide and W.M. Haynes, editors, CRC 1986). Handbook of Chemistry and Physics, 90th Edition 32. S.P. Das, Rev. Mod. Phys., 76 (2004), pp. 785– The author is grateful to his collabo- (Oxford, U.K.: Taylor and Francis, 2009). 851. rators for valuable input, T. Nagase in 13. N.H. Nachtrieb and J. Petit, J. Chem. Phys., 24 33. W. Götze, Complex Dynamics of Glass-Forming particular for his comments and infor- (1956), pp. 746–750. Liquids: A Mode-Coupling Theory (Oxford: Oxford 14. F. Faupel et al., Rev. Mod. Phys., 75 (2003), pp. University Press, 2009). mation regarding radiation damage in 237–280. 34. For example, V.N. Novikov and A.P. Sokolov, Phys. metallic glasses. This work has been 15. P. Klugkist et al., Phys. Rev. Lett., 80 (1998), pp. Rev. E, 67 (2003), 031507. sponsored by the Division of Materi- 3288–3291. 35. T. Egami, K. Maeda, and V. Vitek, Phil. Mag. A, 41 16. T. Egami, Ann. N.Y. Acad. Sci., 37 (1981), pp. (1980), pp. 883–901. als Sciences and Engineering, Offi ce 238–251. 36. T. Egami and Y. Waseda, J. Non-Cryst. Solids, 64 of Basic Energy Sciences, U.S. Depart- 17. T. Egami, J. Mat. Sci., 13 (1978), pp. 2587–2599. (1984), pp. 113–134. ment of Energy. 18. W. Dmowski et al., Mater. Sci. Eng. A, 471 (2007), 37. S.-P. Chen, T. Egami and V. Vitek, Phys. Rev. B, 37 pp. 125–129. (1988), pp. 2440–2449. References 19. D. Srolovitz, T. Egami, and V. Vitek, Phys. Rev. B, 24 38. V.A. Levashov et al., Phys. Rev. B, 78 (2008), (1981), pp. 6936–6944. 064205. 1. P.W. Anderson, Science, 267 (1995), p. 1615. 20. T. Egami et al., J. de Phys., 41 (1980), pp. C8-272– 39. T. Egami, Mater. Sci. Eng., A226-228 (1997), pp. 2. J.S. Langer, Physics Today (February 2007), pp. 274. 261–267. 8–9. 21. T. Egami, Bulk Metallic Glasses, ed. M. Miller and 40. T. Egami et al., Phys. Rev. B, 76 (2007), 024203. 3. For example, R.M. Martin, Electronic Structure: P. Liaw (Berlin: Springer-Verlag, 2008), pp. 27–56. Basic Theory and Practical Methods (Cambridge, 22. A. Slipenyuk and J. Eckert, Scripta Mater., 50 T. Egami, University of Tennessee-Oak Ridge U.K.: Cambridge University Press, 2004). (2004), pp. 39–44. National Laboratory distinguished scientist and 4. M.H. Cohen and D. Turnbull, J. Chem. Phys., 31 23. M.E. Launey et al., Appl. Phys. Lett., 91 (2007), professor, is with the Joint Institute for Neutron (1959), pp. 1164–1169. 051913. Sciences, Department of Materials Science and 5. D. Turnbull and M.H. Cohen, J. Chem. Phys., 34 24. M. Kohda et al., unpublished. Engineering, and Department of Physics and (1961), pp. 120–125. 25. Y. Petrusenko et al., Intermetallics, 17 (2009), pp. Astronomy, University of Tennessee, Knoxville, TN 6. D. Turnbull and M.H. Cohen, J. Chem. Phys., 52 246–248. 37996, and Oak Ridge National Laboratory, Oak (197), pp. 3038–3041. 26. T. Nagase, private communication. Ridge, TN 37831. Prof. Egami can be reached at 7. J.D. Bernal and J. Mason, Nature, 188 (1960), pp. 27. For example, R.S. Farr, J.R. Melrose, and R.C. [email protected].

For the latest advances in all aspects of MOVPE Technology… It’s where you belong…

May 23 – 28, 2010 Hyatt Regency Lake Tahoe Incline Village, Nevada, USA Registration is Now Available!

http://www.tms.org/Meetings/Specialty/icmovpe-xv/home.aspx/

15th International Conference on Metal Organic Vapor Phase Epitaxy Bulk Metallic Glasses Research Summary Mechanical Response of Metallic Glasses: Insights from In-situ High Energy X-ray Diffraction

Mihai Stoica, Jayanta Das, Jozef Bednarčik, Gang Wang, Gavin Vaughan, Wei Hua Wang, Jürgen Eckert

The term “metallic glass” usually re- load-bearing conditions and restricts formation mechanisms characteristic fers to a metallic alloy rapidly quenched their application. This also hinders the to BMGs.11 Recently, characterization in order to “freeze” its structure from precise study of some fundamental is- of amorphous materials by diffraction the liquid state. A metallic glass is a sues in glasses, such as the deformation methods for the purpose of strain scan- metastable alloy, which lacks the sym- mechanism and the dynamics of plastic ning was established.12 Several glasses metry typical for crystalline materi- deformation, in which large plasticity were investigated since then, using dif- als and at room temperature shows an is needed for detailed analysis.9 Plastic ferent synchrotron sources: HASYLAB amorphous liquid-like structure. Bulk deformation of metallic glasses at room at Deutsches Elektronen-Synchrotron metallic glasses (BMGs) represent a temperature occurs through the forma- (DESY) Hamburg, Germany; Euro- class of amorphous alloys. The most tion and evolution of shear bands and is pean Synchrotron Radiation Facilities notable property of BMGs is their ul- localized in thin shear bands.10 There- (ESRF) Grenoble, France; or Advanced trahigh (near theoretical) strength and fore, brittleness is regarded as an intrin- Photon Source (APS) at Argonne Na- hardness. Because the known BMGs sic defect of metallic glasses. tional Laboratory, USA. Monochromat- usually miss tensile plasticity and thus Many methods have been devel- ic hard x-rays with energies at 80–100 exhibit catastrophic failure upon ten- oped and employed to rule out the de- keV were used for these experiments. sion it is important to understand defor- Ex-situ compression tests are usually mation mechanisms involved and thus How would you… performed to study the mechanical be- improve their performance. This aricle …describe the overall signifi cance havior of BMGs. This method is rela- analyzes the use of synchrotron radia- of this paper? tively simple and suitable for small sam- tion for evaluating the elastic-plastic This article analyzes the use of the ples. Tensile tests have technical limita- synchrotron radiation for evaluating response of such materials. the elastic-plastic response of bulk tions. First, for such tests a dog-bone metallic glasses (BMGs). BMGs are shaped plate or rod sample is necessary, INTRODUCTION a new class of materials and their with a length of a few centimeters. This properties make them very attractive Due to the lack of crystalline struc- for applications. Here we show some requires a BMG sample with quite large ture, bulk metallic glasses (BMGs) may results obtained upon in-situ x-ray geometrical dimensions, which cannot achieve interesting properties, including diffraction using synchrotron beam. be achieved by a poor glass former. The high strength and high hardness, excel- …describe this work to a sample should be homogeneous, but in lent corrosion resistance, high wear re- materials science and engineering practice some small voids (as pores or professional with no experience in sistance, very good soft magnetic prop- your technical specialty? oxides inclusions) may be present upon erties, and, depending on composition, Due to the absence of a crystalline casting. Another limitation comes from biocompatibility.1,2 The high strength of network, the BMGs may achieve the device used for tests—it is quite dif- BMGs is sometimes accompanied by high strength and elasticity, together fi cult to create a proper clamping sys- with good wear and corrosion plastic deformation and their deforma- resistance. It is of great importance to tem. Due to the difference in hardness tion and fracture mechanisms are quite understand deformation mechanisms between BMGs and the hardened steel different from crystalline materials.3–7 involved and thus to improve their used for tools, the BMG sample tends to Bulk metallic glasses have strengths ap- performance. Time resolved in-situ x- slide from the grips. 8 ray diffraction experiments may give proaching the theoretical limit, but their crucial insights about the mechanical See the sidebar for experimental de- plasticity at room temperature is typi- behavior up to atomic level. tails. cally very low. In uniaxial tension, the …describe this work to a DATA TREATMENT AND plastic strain is almost zero.9 For most of layperson? THEORETICAL the known BMGs, plastic strain at room Bulk metallic glasses, novel materials BACKGROUND temperature is limited, less than 2%, with amorphous structure, have outstanding mechanical properties. even under compression, resulting from The emergence of such properties The elastic scattering intensity I(Q) is pronounced shear localization and work can be studied by analyzing the x- measured as a function of the scattering softening. The lack of plasticity makes ray diffraction patterns upon in-situ vector (or wave vector) Q, which is de- experiments. BMGs prone to catastrophic failure in fi ned as 4 π sin θ / λ, where θ is half of

76 www.tms.org/jom.html JOM • February 2010 the scattering angle (see Figure B) and grated with the help of the FIT2D soft- intensity I as a function of wave vector λ the wave length of the radiation. The ware19 and after integration the intensity Q can then be transformed into the struc- structure factor can be written as: curves were corrected for background, ture factor S(Q). The index i in Equation I(Q) polarization, and inelastic Compton 2 takes the discrete values from 1 to 18 (1) S(Q) 2 scattering. A typical diffraction pattern (due to the 36 section used for integra- N f(Q) after integrating the diffraction image is tion). Once the tensile load is applied, where N is the number of atoms, f(Q) is presented in Figure 1b. The diffracted the round concentric halos from Figure the atomic scattering factor for x-rays, and the angular brackets indicate aver- EXPERIMENTAL SET-UP AND MEASURED BMGs aging over the composition of the mate- The requirements for in-situ tensile tests under synchrotron radiation are: proper sample, rial.18 The real-space structural informa- proper testing device, access at the hard x-ray source, and proper geometric set-up. In our tion available from S(Q) is the pair dis- case, the samples used for testing were cast as amorphous plates which further were ma- chined by the spark erosion method in order to obtain a dog-bone shaped specimen with 10 tribution function g(r), (PDF), in which mm × 2 × 1 mm2 reduced section (Figure A). The dog-bone shaped specimen was strained r is the distance from an average atom using a tensile module from Kammrath and Weiss GmbH, which can achieve a maximum located at the origin. Without entering load of 5 kN. The room temperature in-situ x-ray diffraction experiments were performed too much in details—the entire math- on the wiggler beamline BW5 at the DORIS positron storage ring (HASYLAB at DESY, ematic background can be found in sev- Hamburg, Germany) using monochromatic synchrotron radiation of 103.8 keV (λ = 0.0119 eral other works12–18—one should men- nm). The layout of the experimental setup is shown in Figure B. The measured samples tion that the pair distribution function is were exposed for 10 s to the well collimated incident beam having a cross section of 1 × 1 related to S(Q) by a Fourier transform. mm2. Two dimensional (2-D) (2,300 × 2,300 pixels, 150 × 150 μm2 pixel size) x-ray dif- It is also common to write the real- fraction (XRD) patterns were collected using a MAR 345 2-D image plate detector care- fully mounted orthogonally to the x-ray beam. The diffraction pattern from LaB was used space structural information in terms of 6 to calibrate the sample-to-detector distance D and tilting of the image plate detector with the radial distribution function (RDF), respect to the beam axis. 2 which is defi ned as 4 π r g(r). With this 13,14 The BMGs studied using this set-up were of the composition Zr64.13Cu15.75Ni10.12Al10. defi nition, the coordination number of a At the same location, BW5 beamline, and with the same geometrical set-up, Wang et al.15 particular atomic shell of interest can be measured Zr62Al8Ni13Cu17 and La62Al14(Cu5/6Ag1/6)14Co5Ni5 BMGs using x-ray radiation obtained by integrating the RDF over a with λ = 0.012389 nm. In-situ tensile tests at BW5 were also reported by Mattern et al.16 suitably chosen range of r. Their samples were amorphous ribbons with the compositions Cu50Zr50 and Cu65Zr35 and When an amorphous material is sub- the radiation had λ = 0.01265 nm. Prior to these experiments, few others were performed in compression. At ESRF Grenoble, Poulsen et al.12 investigated the Mg Cu Y BMGs jected to forces that create a macroscopic 60 30 10 and Das et al.17 Cu Zr Al and Zr Cu Ni Al Ti BMGs. The wave length of the radia- stress, both S(Q) and g(r) will be affect- 47.5 47.5 5 55 20 10 10 5 tion used for these experiments was λ = 0.01412 nm for Mg-based glasses and λ = 0.0155 ed. For uniaxial loading, the changes in nm for Zr-based glasses. Other BMGs, Zr57Ti5Cu20Ni8Al10, were tested in uniaxial com- real space can be easy to anticipate. The pressive load using a 0.0154 nm radiation at Advanced Photon Source at Argonne National tensile stress will tend to move atoms Laboratory by Hufnagel, Ott, and Almer.18 For further comparisons between measured apart in the loading direction, and thus a data it is important to point out that for almost all mentioned experiments, the diffraction peak in g(r) for that direction will move images were recorded with the same to larger values of r. For a compressive kind of 2-D detector, the MAR 345 stress, the opposite should happen. In image plate. The only exception came 17 the reciprocal space, Q, it is expected to from Das et al. They used as detec- tor a fast response low noise charge shift toward lower values in the case of coupled device (FReLoN CCD). tensile stress and higher values when the However, the primary diffraction im- sample is compressed. By analogy with ages were integrated with the help of the simple defi nition of engineering Figure A. A dog-bone shaped sample used for the FIT2D software.19 strain, the tensile strain for an applied tensile tests. stress σ can be defi ned as:

QQ(,0)(,)Mii MV HIVii(), (2) Q(,)MVi which is angular dependent. In the transverse direction, one expects a strain of the opposite sign due to the Poisson effect. A typical diffraction image of an amorphous sample is illustrated in Fig- ure 1a. The amorphicity is proved by the absence of any clear ring. Then the image is integrated upon the polar co- ordinates (s, ϕ). The integration is done Figure B. A sketch of a typical in-situ tensile ex- by dividing the entire circle into 36 sec- periment. tions of 10 each. The data were inte-

Vol. 62 No. 2 • JOM www.tms.org/jom.html 77 2,250 ative change in the position of the fi rst peak using Equation 2. The fi t of the 2,000 experimental data to Equation 3 yields ε 1,750 two components of the strain tensor, 11 and ε (the axial and tangential compo- 1,500 22 nents, respectively). The stress-strain 1,250 curves as observed for different strain tensor components are plotted in Figure 1,000 4. Within the experimental error all of 750 them show a linear behavior, indicating the elastic regime of the tensile deforma- 500 tion for the investigated specimens. The 250 /( Q ) (a.u.) Intensity, samples fractured at a stress of about 1,500 MPa with no sign of yielding, de- 0 140120100806040200 a 0 250 500 750 1,000 1,250 1,500 1,750 2,000 2,250 b Wave Vector, Q (nm−1) spite the fact that the compressive yield Figure 1. (a) Diffraction image as recorded by the MAR 345 2D image plate detector. The strength of this BMG was reported to be polar coordinates (s, ϕ) and the axis of tensile deformation are depicted. (b) X-ray diffraction 1,690–1,851 MPa,21 indicating a signifi - ϕ pattern resulting from integration among the polar coordinates (s, ) of the image presented cant strength asymmetry for this BMG. in (a). Both (a) and (b) are typical for metallic glasses and here they were recorded for The maximum axial strain (ε ) is 1.50 ± Zr64.13Cu15.75Ni10.12Al10 BMG. 11 0.01%. The elastic modulus determined

1a become elliptical and the asymmetry tallic glasses and confi rms the presence in tensile mode is E11 = 94 ± 1 GPa and is higher as the load increases. Then the of glassy structure without any hint of the experimentally determined Pois- ν ε ε angular variation of the strain can be fi t- crystalline inclusions (see Figure 1). son’s ratio = – 22 / 11 is 0.325 ± 0.01. ted with:13 The symmetric circular diffraction pat- Using the diffraction data by in- ε ϕ σ ε 2 ϕ ( , ) = 11 sin tern is characteristic for the samples situ high-energy XRD, the tensile prior to applying tensile stress. With in- modulus and Poisson’s ratio can be + ε sinϕ cosϕ + ε cos2ϕ (3) 12 22 creasing tensile load it becomes ellipti- accurately evaluated in the case of

As a result, the strain tensor can be de- cal. The changes are most pronounced Zr62Al8Ni13Cu17 and La62Al14(Cu5/6 ε 15 termined, and the axial 11, tangential for the fi rst and strongest diffuse ring Ag1/6)14Co5Ni5 BMGs. As in the case of ε ε 22, and in-plane shear component 12 (halo) appearing in the 2-D XRD pat- Zr64.13Cu15.75Ni10.12Al10, no tensile plas- can be derived. Taking into account the tern. To describe such changes more ticity was observed, despite the fact that set-up presented in Figure B, ϕ = 90 quantitatively one has to construct the both Zr-glasses are rather deformable in corresponds to the axial stress and ϕ = set of symmetrized intensity distribu- compression.15,21 The strains determined 0 to the tangential stress. Components tions as described previously and trace from the diffraction data of tensile/trans- not in the plane perpendicular to the in- the change in the fi rst peak position as a verse directions for the Zr62Al8Ni13Cu17 ϕ coming beam can be determined by ro- function of azimuth angle, , and tensile and La62Al14(Cu5/6Ag1/6)14Co5Ni5 BMGs tating the specimen around an axis per- stress, σ. It should be noted here that no are presented in Figure 5. There one can pendicular to the incoming beam. In the changes were observed between seven see the good linear behavior and basi- more general case of a multiaxial stress diffraction patterns, independently ac- cally no sign of yielding. By linearly fi t- state, the complete strain tensor can be quired along the length of the sample ting the points and calculating the ratio determined from measurement of the (in 1 mm steps), when holding the load of strains between the transverse and strain in various directions.14,18 at a fi xed value of external stress. The tensile directions for each alloy, the ten- The same analysis can be done in the experimental scatter of the measured sile elastic modulus and Poisson’s ratio real space. However, experimentally a strain values at each different stress were obtained, about 83 GPa and 0.37 good correlation was observed between level for seven independent locations for Zr62Al8Ni13Cu17 BMG and 34 GPa the two sets of data. Dmowski and Ega- along the gauge length is shown as error and 0.36 for La62Al14(Cu5/6Ag1/6)14Co5Ni5 mi20 pointed out that in the presence of bars in Figures 2 and 3. From Figure 2 BMG, respectively. structural anisotropy, it is necessary to it is evident that the asymmetry of the Using the same method, Mattern expand the PDF into spherical harmon- fi rst diffuse maximum increases with in- et al.16 measured upon tensile loading ics, otherwise systematic errors may oc- creasing load. The decrease in the peak amorphous ribbons with the composi- cur especially in the fi rst neighborhood. position with increasing tensile stress tions Cu50Zr50 and Cu65Zr35. The corre- (the curve corresponding to ϕ = 90 Fig- sponding data for Cu Zr are presented STRAIN ANALYSIS 50 50 ure 1) refl ects the fact that atoms move in Figure 6. There one can see again a apart along the tensile direction. An good linear behavior up to the highest Tensile Tests and Reciprocal opposite behavior is seen in transver- value before fracture of the strain with Space sal direction (the curve corresponding applied stress and no sign of plastic de- The 2-D diffraction pattern of as-cast to ϕ = 0 in Figure 1). Figure 2 shows formation. The Young’s modulus E = ν Zr64.13Cu15.75Ni10.12Al10 exhibits the dif- the angular variation of the strain at a 63 GPa and the Poisson’s ratio = 0.31 fuse scattering pattern typical for me- given stress σ as calculated from the rel- were calculated directly from the slopes

78 www.tms.org/jom.html JOM • February 2010 γ of the axial and tangential components ring after loading compared to a circular However, the shear component 12 val- ε ε 11 and 22 vs. stress. The other glass, feature of the unloaded state, indicating ues remain close to zero. The increment ε ε Cu65Zr35, behaves similarly. a decrease of the atomic spacing of the of 11 and 22 strains slightly deviates nearest neighbors along the loading axis. from linearity after 1,400 MPa (a dot- Compression Tests and The data analysis has been performed by ted line has been drawn to show the Reciprocal Space the Q-space method in reciprocal space, linearity of the elastic stress-strain rela- The evaluation of the elastic tensor as described earlier. In the case of each tionship), and the sample broke at 1,740 ε by synchrotron radian was established integrated intensity I(Q), the shift of the MPa with an axial strain 11 = −0.0174 12 ε by Poulsen et al. At ESRF, they mea- fi rst halo was determined with respect to and a transverse strain 22 of +0.00675. sured several Mg60Cu30Y10 BMGs under the unloaded condition. This strength value is similar to the compressive load. A good linear depen- Figure 7a and b, taken from Refer- macroscopic yielding (MY) of this al- dence of the strain as a function of the ence 17, shows the evolution of the loy at 1,727 MPa, as observed earlier.24 12 applied stress was also found. This different atomic-scale strain compo- Cu47.5Zr47.5Al5 shows (Figure 7b) a very 22 ε ε γ ε BMG was known to be brittle, and, nents ( 11, 22 and 12 = 12) with ap- similar stress-strain relationship. How- as a consequence, the samples failed plied stress for Zr55Cu20Ni10Al10Ti5 and ever, the nonlinear stress-strain behav- at the end of the elastic regime. Huf- Cu47.5Zr47.5Al5, respectively. In the case ior starts at around 1,200 MPa and, 18 nagel, Ott, and Almer measured, at of Zr55Cu20Ni10Al10Ti5, the strain com- fi nally, the elastic strain saturates at a ε APS, Zr57Ti5Cu20Ni8Al10 BMGs loaded ponents in the loading direction ( 11) stress of 1,506 MPa with an axial strain ε ε in uniaxial compression. They used the and transverse direction ( 22) increase. 11 = −0.0150 and a transverse strain structure factor S(Q) recorded for the loading direction from many x-ray scat- tering patterns taken at various stresses during incremental loading from 0 MPa to 1,080 MPa (approximately 60% of the yield stress for this alloy18) and back to zero. As the compressive stress in- creases, the largest peak in S(Q) shifts Figure 2. Shifting of the fi rst broad to larger Q in the loading direction. The peak as a function of applied tensile opposite trend was observed for S(Q) load, measured in tensile (ϕ = 90¡) and transversal direction (ϕ = 0¡), in the transverse direction. As a result, for Zr Cu Ni Al BMG. ε 64.13 15.75 10.12 10 the axial component 11 of the strain ten- sor became negative and the tangential ε component 22 became positive. The results indicated that the strain increases linearly with increasing com- pressive stress. A straight-line fi t to the data yields an elastic modulus of E = 87 ± 2 GPa, in good agreement with values Figure 3. Angular dependence for E determined by macroscopic mea- of the strain determined at various stages of tensile de- surements on closely related amorphous formation of Zr Cu Ni 18 64.13 15.75 10.12 alloys. Data for the transverse direction Al10 BMG, as calculated from are also in a linear dependence with the the relative change in the po- sition of the fi rst peak using applied stress and from both directions Equation 2. one can obtain the value for Poisson’s ratio of ν = 0.34 ± 0.01, also in reason- able agreement with the macroscopic measurements.18 Strain scanning by x-ray diffraction beyond the Hookean limit, in order to in- vestigate the plastic yielding phenomena of two different BMGs at higher resolu- Figure 4. Stress-strain curves tion, was done by Das et al.17 The alloys for different strain tensor are “plastic” Cu Zr Al 23 and macro- components measured for 47.5 47.5 5 Zr Cu Ni Al BMG. The 24 64.13 15.75 10.12 10 scopically “brittle” Zr55Cu20Ni10Al10Ti5 straight lines represent a linear BMGs. The experiments have been per- fi tting of the experimental data, formed at ESRF Grenoble, France, and starting from the origin of the coordinate system. the detailed experimental setup is de- scribed elsewhere.17 The diffraction pat- terns showed the elliptical nature of the

Vol. 62 No. 2 • JOM www.tms.org/jom.html 79 ε 22 = +0.0056 without alteration of the roscopic yielding under compression the data from the real space. For that, γ shear component 12, which is close to and the saturation of the elastic strain at one has to calculate the structure factor 0. The test was stopped at 1,700 MPa the atomic scale is consistent for both S(Q) or pair correlation function g(r) as at a plastic strain of about 0.6%–0.7%. the investigated alloys. described earlier. An example is given Note that the microscopic yield stress in Figure 8, which represents g(r) of Strain Analysis from the Real has been measured to be 1,547 MPa Zr64.13Cu15.75Ni10.12Al10 BMG at different 23 Space for Cu47.5Zr47.5Al5, as reported earlier. stages of deformation. The fi rst peak in Therefore, the stress required for mac- The strain analysis can be done using g(r) (see the inset) shifts to larger r with increasing load, as expected, and the transverse data (not shown) show the opposite trend. According to different authors,18,20 the peak positions in g(r) are diffi cult to determine accurately be- Figure 5. The strains deter- mined from the diffraction data cause the peaks at low r are asymmetric of tensile/transverse directions while those at larger r are rather broad,

for Zr62Al8Ni13Cu17 and which leads to signifi cant scatter in the La62Al14(Cu5/6Ag1/6) 14Co5Ni5 BMGs. As for Figure 4, the measured strain. A more robust tech- straight lines represent the lin- nique proposed by Hufnagel, Ott, and ear fi tting of the experimental Almer18 is to focus not on the tops of the data. peaks, but on the places where g(r) = 1. These crossing points are less sensitive to the effects of asymmetry and can be accurately determined even for peaks at large r. Although no dependence of strain on r was observed, it is interest- ing that the strain determined from the lowest value of r at which g(r) = 1 is consistently smaller in magnitude than Figure 6. Strain vs. applied stress the strains determined at larger values of r.18 This is related to the asymmetry of Cu50Zr50 glass measured by x-ray diffraction. Here also the of the peaks, due by the changes in the linear fi tting is presented by the straight lines. interatomic bonding lengths. More, the asymmetry becomes more pronounced when the applied stress is increasing.To investigate this further, one has to move to RDF and try to deconvolute the fi rst peak which corresponds to the fi rst co- ordination shell. We cannot unambigu- ously identify the atomic pairs contrib- uting to the fi rst peak in the RDF, but we can make some reasonable approxi-

mations for this Zr64.13Cu15.75Ni10.12Al10 alloy. First, because the contribution of each atomic pair to the RDF is weight- ed by the atomic scattering factors of a the elements and by their concentra- Figure 7. Evolu- tion, we can neglect the infl uence of tion of elastic strain Al, because it has a low atomic number components of (a) and is present at relatively low concen- Zr Cu Ni Al Ti and 55 20 10 10 5 tration. Second, the separation of the (b) Cu47.5Zr47.5Al5 during compressive loading. atoms in each pair is related to the sum ε The increment of 11 of their atomic radii; since Cu and Ni and ε strains deviates 22 are nearly the same size (1.28 and 1.25 from linearity on atomic 25 scale before the onset Å, respectively), their contributions of macroscopic yielding are indistinguishable. It is almost clear (MY). The fi gure is taken that Zr–Zr, Zr–Cu, and Zr–Ni are the from Reference 17. dominant atomic pairs which constitute b the fi rst coordination shell of the PDFs, so only two partials [Zr–(Cu,Ni) and

80 www.tms.org/jom.html JOM • February 2010 (Zr–Zr)] need to be resolved. Figure 9a shows the result of the deconvolu- tion for the fi rst shell of undeformed

Zr64.13Cu15.75Ni10.12Al10. The center of the fi rst peak was estimated to be at 2.68 Å corresponding to Zr–(Cu,Ni) atomic pairs. The major component centered at 3.14 Å originates from Figure 8. PDFs g(r) of Zr64.13Cu15.75Ni10.12Al10 Zr–Zr atomic pairs. As can be seen BMG at different stages from Figure 9b, an increase in tensile of deformation. The inset stress shifts both peaks toward higher shows a zoomed view of the r values. This proves that tensile stress fi rst coordination shell. increases the average atomic distances along the loading direction. Similar re- sults were obtained in the case of the other metallic glasses discussed here; in the case of compressive loadings, the average atomic distances along the loading direction became smaller.18 DATA SUMMARY, COMPARISON, AND DISCUSSION All of the above discussion assumes that the amorphous material is isotropic. This is often the case for amorphous al- a loys, but counterexamples can be found in thin fi lms and in bulk alloys subjected to processing that renders them aniso- tropic. Even in these cases the devia- Figure 9. Deconvolution of the fi rst coordination shell of Zr Cu tion from isotropy is usually small and 64.13 15.75 Ni10.12Al10 BMG into two Gaussians the scattering data are analyzed using for (a) unloaded sample and (b) the isotropic assumption. For the data sample at different stages of de- presented here it was assumed that one formation. The experimental data curve, the fi tting curves, the stress can neglect the anisotropy induced by level, and the contribution of pos- the small uniaxial elastic strain. Hufna- sible atomic pairs are correspond- gel, Ott and Almer mention18 that a full ingly denoted in the plot. treatment would involve the application b of cylindrical distribution functions to properly handle the symmetry. increased linearly with uniaxial stress. smaller than that at longer scales, the ef- Table I summarizes the values of The strain measured from the posi- fect being smaller than that reported by Young’s modulus E and Poisson’s ratio tion of the fi rst peak in I(Q) showed Poulsen et al.12 They propose that “the ν for the metallic glasses analyzed here. good agreement with strain calculated difference between the stiffness of the Two types of data are presented: mea- based on macroscopic measurements nearest-neighbor atomic environment sured by XRD using synchrotron ra- of Young’s modulus E. However, the and that over longer length scales can diation and measured using ultrasound strain calculated from g(r) showed a be attributed to the effect of topological methods or macroscopic tensile/com- pronounced dependence on r, being the rearrangements in the nearest-neighbor pression tests. Basically, there always smaller for the fi rst near neighbor peak environments of a relatively small frac- are differences in the constants mea- and increasing with peaks at higher r tion of the atoms, without the need to sured using the two methods. to asymptotically approach the strain invoke signifi cant structural rearrange- For compressive loads, as mentioned calculated from the I(Q) peak position. ments over longer length scales.”18 also by Hufnagel, Ott, and Almer,18 The strain calculated from the third g(r) The differences in mechanical con- Poulsen and co-workers12 demonstrated peak was even 2.7 times as large as stants measured using different tech- for the fi rst time that the theoretical con- that from the fi rst peak. Poulsen et al. niques are even higher when the sam- siderations used in this work are essen- attributed this observation to unspeci- ples were subjected to tensile loads tially correct, by measuring elastic strain fi ed “structural rearrangements on the (see Table I). Here the stiffness of the 12 and strain distributions in Mg60Cu30Y10 length scale of 4–10 Å.” Hufnagel and neighbor cells becomes more important. BMG. They showed that the strain mea- co-workers18 observed as well that the Basically, upon tensile tests no sign of sured in both reciprocal and real space strain in the nearest-neighbor shell is plastic deformation was observed. In

Vol. 62 No. 2 • JOM www.tms.org/jom.html 81 the fi rst, second, third, and consecutive ν Table I. The Values for Young’s Modulus E and Poisson’s Ratio for the Metallic Glasses atomic shells. Presented in this Paper ACKNOWLEDGEMENTS Composition Experiment E (GPa) ν

13 The authors thank N. Mattern and S. Zr64.13Cu15.75Ni10.12Al10 Tensile, synchrotron 94 0.325 Ultrasound13 78 0.377 Pauly for stimulating discussions. We 15 thank also the beamline staff from BW5, Zr62Al8Ni13Cu17 Tensile, synchrotron 83 0.37 Ultrasound15 80 0.38 Hasylab Hamburg, Germany and from 15 La62Al14(Cu5/6Ag1/6)14Co5Ni5 Tensile, synchrotron 34 0.36 ID11, ESRF Grenoble, France, for their Ultrasound15 35 0.36 help during diffraction measurements. 16 Cu50Zr50 Tensile, synchrotron 63 0.31 The authors are grateful for the fi nan- Ultrasound26 83 0.384 cial support provided by the European 16 Cu65Zr35 Tensile, synchrotron 97 0.33 Union within the framework of the Re- Ultrasound26 92 0.352 search Training Network on ‘‘Ductile Mg Cu Y Compression, synchrotron12 64.1 0.373 60 30 10 BMG composites” (MRTN-CT-2003- 22 Ultrasound 51.5 N/A 504692). The Alexander von Humboldt Zr Ti Cu Ni Al Compression, synchrotron18 87 0.34 57 5 20 8 10 Foundation is acknowledged for fi nan- Macroscopic compression27 82 0.362 cial support of G. Wang. 17 Cu47.5Zr47.5Al5 Compression, synchrotron 99.2 0.34 Ultrasound17 90.1 0.365 References 17 Zr55Cu20Ni10Al10Ti5 Compression, synchrotron 91.1 0.38 1. W.L. Johnson, MRS Bull., 24 (1999), p. 42. Ultrasound17 85.5 0.378 2. A. Inoue, Acta Mater., 48 (2000), p. 279. 3. H. Chen et al., Nature, 367 (1994), p. 541. 4. S. Venkataraman et al., Scripta Mat., 54 (2006), p. 13 835. fact we considered that here one can response of the nearest neighborhood 5. M. Stoica et al., Intermetallics, 13 (2005), p. 764. deal with two types of glasses: intrinsi- upon loading leads toward the direc- 6. U. Kühn et al., Appl. Phys. Lett., 77 (2000), p. 3176. cally brittle and plastic deformable in tional changes in the chemical short- 7. J. Eckert, Mat. Sci. Eng. A, 226 (1997), p. 364. 8. A.L. Greer, Science, 267 (1995), p. 1947. compression. Judging from the trend of range order. All together may explain 9. L.Q. Xing et al., Phys. Rev. B, 64 (2001), 180201. the values of the elastic constants sum- the differences in mechanical behavior 10. J.J. Lewandowski and A.L. Greer, Nat. Mater., 5 marized in Table I, as measured by ul- observed at the microscopic level (x-ray (2006), p. 15. 11. D.M. Dimiduk et al., Science, 312 (2006), p. 1188. trasonic methods and as calculated from diffraction) when compared with the av- 12. H.F. Poulsen et al., Nat. Mater., 4 (2005), p. 33. the strain tensor measured by diffraction eraged macroscopical behavior. 13. M. Stoica et al., J. Appl. Phys., 104 (2008), experiments, it is very clear that in the 013522. CONCLUSIONS 14. J. Bednarčik and H. Franz, J. Phys.: Conf. Series, case of “nondeformable” or intrinsically 144 (2009), 012058. brittle BMGs, the bulk elastic constants In-situ x-ray synchrotron diffraction 15. X.D. Wang et al., Appl. Phys. Lett., 91 (2007), as derived from ultrasound measure- enables the atomic level elastic strain of 081913. 16. N. Mattern et al., Acta Mat., 57 (2009), p. 4133. ments and the ones obtained from the metallic glasses under uniaxial tensile 17. J. Das et al., Phys. Rev. B, 76 (2007), 092203. strain tensor analysis are almost similar, or compressive stress to be character- 18. T.C. Hufnagel, R.T, Ott, and J. Almer, Phys. Rev. B, indicating a similar elastic behavior of ized. The elastic moduli can be estimat- 73 (2006), 064204. 19. A.P. Hammersley et al., High Press. Res., 14 each atomic shell. In the case of plas- ed not only considering the shift of the (1996), p. 235. tically deformable BMGs each atomic fi rst maximum of the scattering curve in 20. W. Dmowski and T. Egami, J. Mater. Res., 22 shell has a different stiffness, as revealed reciprocal space but also from the shift (2007), p. 412. 21. Y.H. Liu et al., Science, 315 (2007), p. 1385. from the large differences of the elastic of the larger inter-atomic distances in 22. A. Castellero et al., J. Alloys and Comp., 434-435 constants from ultrasonic and tensor the PDF in real space. The analysis of (2007), p. 48. analysis. Most likely, such local fl uc- the short-range order of several metallic 23. J. Das et al., Phys. Rev. Lett., 94 (2005), 205501. 24. K. Hajlaoui et al., J. Non-Cryst. Solids, 353 (2007), tuations of the elastic properties in the glasses vs. stress confi rms the structural p. 327. glassy structure can rather easily induce changes in the elastic regime. The an- 25. ASM Handbook: Volume 1.2 Metals Handbook local shear transformation and, thus, the elastic deformations are accompanied (Materials Park, OH: ASM International, 1992). 13 26. G. Duan et al., Scripta Mater., 58 (2008), p. 159. BMG exhibits macroscopic plasticity. by bond reorientation leading to direc- 27. J.J. Lewandowski, W.H. Wang, and A.L. Greer, Phi- However, in a glassy material a range tion dependent changes in chemical los. Mag. Lett., 85 (2005), p. 77. of local atomic environments of the short-range order. The elastic constants atoms exists. As a consequence of the calculated from the strain tensor are dif- Mihai Stoica and Gang Wang, researchers, and Jürgen Eckert, director, are with the Institute for disorder, fl uctuation of inter-atomic dis- ferent from those measured by macro- Complex Materials, Helmholtzstr. 20, Dresden, tances may occur, which leads to varia- scopic techniques. These differences 01069, Germany; Jayanta Das, senior research fel- tions of the atomic-level stress.16 The are supposed to arise because the ul- low, is with IIT Kharagpur, Department of Metallur- gical and Materials Engineering, Kharagpur, India; analysis of the fi rst neighborhood, done trasound techniques average the elastic Jozef Bednarčik, scientist, is with DESY, Hamburg, by Mattern et al.16 for their CuZr binary constants of different atomic shells and Germany; Gavin Vaughan, scientist, is with ESRF, glasses, confi rms the anelastic changes measure the bulk properties of the mate- Grenoble, France; and Wei Hua Wang, professor, is with the Institute of Physics, Chinese Academy of of the short-range order under tensile rial, while the diffraction measurements Sciences, Beijing, China. Dr. Stoica can be reached stress well below the yield strength. The may detect differences in stiffness of at [email protected].

82 www.tms.org/jom.html JOM • February 2010 Bulk Metallic Glasses Overview Amorphous Metals for Hard-tissue Prosthesis

Marios D. Demetriou, Aaron Wiest, Douglas C. Hofmann, William L. Johnson, Bo Han, Nikolaj Wolfson, Gongyao Wang, and Peter K. Liaw

Owing to a unique atomic structure durability, and longevity of orthopedic lacking microstructural defects, glassy How would you… implants as well as optimization of or- metals demonstrate certain universal …describe the overall signifi cance thopedic hardware from the perspective properties that are attractive for load- of this paper? of patient comfort, mobility, and func- bearing biomedical-implant applica- This article surveys a range of tionality will be of widespread benefi t features of the performance of tions. These include a superb strength, amorphous metals relevant to hard- to implant recipients. which gives rise to very high hardness tissue prosthesis, compares these Manufacturers of orthopedic hard- and a potential for minimizing wear features to those of the current state ware have active research programs of the art, and proposes guidelines for and associated adverse biological re- development of new biocompatible aimed at developing materials with im- actions, and a relatively low modulus, amorphous metal alloys suitable for proved biocompatibility, which entails which enables high elasticity and holds load-bearing implant applications. environmental stability in living tissue, a promise for mitigating stress shield- …describe this work to a mechanical performance, and abil- ing. There are, however, other non-uni- materials science and engineering ity to match, functionally replace, and versal properties specifi c to particular professional with no experience in integrate with various types of living your technical specialty? amorphous metal alloys that are infe- tissue. Amorphous metals with unique Amorphous metals differ rior to presently used biometals and fundamentally from conventional mechanical and chemical properties may be below acceptable limits for crystalline metals in their atomic and attractive processing capabili- hard-tissue prosthesis. In this article, structure and fundamental ties are promising candidate materials features of the performance of amor- thermodynamic state. Specifi cally, for achieving these goals. This article an amorphous metal is a phous metals relevant to hard-tissue confi gurationally-frozen non- surveys a range of features of the per- prosthesis are surveyed and contrasted equilibrium metallic liquid that fails formance of amorphous metals and as- to those of the current state of the art, to form a stable crystalline state, sesses their suitability for hard-tissue and consequently reveals no long- and guidelines for development of new range atomic order. The distinctive prosthesis. biocompatible amorphous metal alloys structural state of amorphous metals See the sidebar on page 86 for back- suitable for hard-tissue prosthesis are gives rise to unique corrosion ground on biocompatibility of current- proposed. characteristics, mechanical ly-used materials. performance, and processing INTRODUCTION capabilities that render amorphous AMORPHOUS METALS metals attractive for load-bearing AS PROSPECTIVE Medical use of metals as prosthetic biomedical implant applications. BIOMATERIALS replacements or fracture fi xation de- …describe this work to a vices for human bone has become layperson? Bulk amorphous metals are a rela- widespread over the past two decades Tens of thousands of hard-tissue tively new class of engineering materi- with millions of hard-tissue replace- replacement surgeries are carried als developed over the past two decades. out annually in the United States. ment and repair surgeries carried out Despite medical statistics showing Unlike traditional engineering met- annually in the United States. Despite high rates of success, there remain als which are structurally crystalline, medical statistics showing high rates serious problems, particularly with amorphous metals are confi gurationally of success, there remain serious prob- respect to tissue compatibility as well frozen liquids which fail to crystallize as from the perspective of patient lems. For example, the rate of revi- comfort, mobility, and functionality. during solidifi cation from the molten sion surgeries for orthopedic implants Improvements in the biological state. While they are electronically and is approximately 7% after 10 years compatibility, durability, and optically metallic like ordinary metals, 1 functionality of orthopedic implants of service. With an aging population will be of widespread benefi t to the absence of crystals and associated and surgeries on younger patients due implant recipients. Amorphous metals extended defects such as vacancies, dis- to trauma or sports injuries, the rate of with unique mechanical and chemical locations, or grain boundaries gives rise revision procedures is growing at an ac- characteristics are highly promising to physical, chemical, and mechanical 1 candidate materials for achieving celerated rate of approximately 60%. these goals. properties that differ fundamentally Improvements in the biocompatibility, from conventional metals. For example,

Vol. 62 No. 2 • JOM www.tms.org/jom.html 83 250 amorphous metals typically exhibit very high strength, hardness, and elastic- ity compared with conventional metals. 200 Unlike silicate glasses, they tend to be tough and have fracture toughness that is often comparable to ordinary metals. 150 The absence of microstructural defects infl uences their chemical behavior as well, often resulting in improved resis- 100 tance to corrosion and chemical attack. Lastly, their ability to soften and fl ow Young’s Modulus (GPa) Young’s upon relaxation at the glass transition 50 gives rise to a viscoplastic-fl ow behav- ior which opens the possibility of “ther- moplastic” forming processes similar 0 0 500 1,000 1,500 2,000 2,500 3,000 3,500 4,000 4,500 to those employed for processing plas- Yield Strength (MPa) tics. The above features have stimulated Figure 1. Young’s modulus vs. yield strength data for amorphous metals [×]11 and ductile- broad interest in metallic-glass engi- phase reinforced amorphous metals [+],12 shown together with data for stainless steels neering applications. Q 2 Q 2 Q 2 [ ], Co-Cr-based [ ], and Ti-based alloys [ ]. The aspects of the performance of 1,400 amorphous metals that are most attrac- tive for hard-tissue prosthesis are their high strength and hardness, which sug- 1,200 gest good load-bearing capability and ) 2 high wear resistance; their low modu- 1,000 lus, which implies better load transfer to the surrounding bone and a potential 800 for mitigating stress-shielding; and their good corrosion resistance. Their unique 600 processing capabilities represent anoth- er attractive feature of these materials, 400

Vickers Hardness (kgf/mm Vickers as they can potentially simplify the cur- rently complex fabrication procedures 200 and bring down fabrication costs. The attractive advantages presented by the 0 0 500 1,000 1,500 2,000 2,500 3,000 3,500 4,000 4,500 thermoplastic and micro-forming capa- Yield Strength (MPa) bilities of these materials as they relate Figure 2. Vickers hardness vs. yield strength data for amorphous metals [×],16 shown to- to biomedical applications are empha- gether with data for stainless steels [Q],2 Co-Cr-based [Q],2 and Ti-based alloys [Q].2 sized in a recent article by Schroers and 10 180 co-workers. 160 MECHANICAL PERFORMANCE OF ) 140

1/2 AMORPHOUS METALS 120 Elasticity 100 Owing to a liquid-like atomic struc- 80 ture associated with structural length scales on the order of few hundred at- 60 oms, amorphous metals offer an un- usual combination of very high strength Fracture Toughness (MPa-m Toughness Fracture 40 and low elastic modulus.11 Low elastic 20 modulus is highly desirable in hard-tis- sue prosthesis, as it would minimize the 0 0 500 1,000 1,500 2,000 modulus mismatch between bone and Yield Strength (MPa) biomaterial, facilitate load transfer to the Figure 3. Fracture toughness vs. yield strength data for amorphous metals [×]18 and ductile- phase reinforced amorphous metals [+],12 shown together with data for stainless steels surrounding bone, and potentially miti- [Q],19 Co-Cr-based [Q],19 and Ti-based alloys [Q].19 gate stress shielding. As demonstrated in Figure 1, for a given strength require-

84 www.tms.org/jom.html JOM • February 2010 ment an amorphous metal exhibits a sig- 1,200 nifi cantly lower modulus than any of the conventional crystalline metallic bioma- 1,000 terials. The recently developed ductile- phase-reinforced amorphous metals also 12,13 800 Cycles) (MPa)

present this advantage, as their over- 7 all elastic behavior appears to resemble that of the monolithic amorphous metal. 600 Consider for example a 1-GPa strength requirement. An amorphous metal com- 400 bines such strength with a modulus that is 75% the modulus of a titanium-based 200 alloy, 25% the modulus of a stainless steel alloy, or 20% the modulus of a Limit (10 Endurance Fatigue Co-Cr-based alloy. Essentially, a low 0 0 500 1,000 1,500 2,000 modulus for a given strength implies a Yield Strength (MPa) high elastic limit. Hard tissue (cortical Figure 4. Fatigue endurance limit (stress-range based) vs. yield strength data for amorphous bone) exhibits an elastic strain limit that metals [×]21 and ductile-phase reinforced amorphous metals [+],23 shown together with data ranges from 1 to 1.5%, depending on for stainless steels [Q],23 Co-Cr-based [Q],23 and Ti-based alloys [Q].23 the degree of mineral content.14 Crystal- line metals exhibit elastic strain limits that typically range between 0.1% and 0.5%.15 Amorphous metals, on the other hand, exhibit elastic strain limits that range from 1.5 to 2%,15 and thus more closely match the elasticity of natural tissue. This unique combination of high strength and elasticity renders amor- phous metals attractive load-bearing biomaterials that could potentially miti- gate stress shielding. Hardness Hardness is the mechanical property that mostly infl uences the wear resis- tance capabilities of a material. Since 10−1 100 101 102 103 104 105 106 hardness is understood to be a measure Corrosion Rate (mm/year) of fl ow stress, it correlates linearly with Figure 5. Corrosion rates of various amorphous Zr-based alloys compared with surgical- the material yield strength. Specifi - grade Co-Cr-based, Ti-based, and stainless steel in 12 M HCl solution.29 cally, the Vickers hardness is expected 1 to scale linearly with the material yield σ 0.9 strength (HV ~ 3 y). Owing to their su- perb yield strength, amorphous metals 0.8 demonstrate an advantage over crystal- line metals in terms of hardness. The 0.7 hardness of amorphous metals as a 0.6 function of yield strength is plotted in Figure 2,16 along with data for conven- 0.5 tional crystalline metallic biomateri- 0.4 als.2 As expected, the hardness of all 0.3

metallic materials, whether amorphous Strength Stress Range/Yield or crystalline, is shown to obey a tight 0.2 universal relation with strength. By cor- 0.1 relating the abrasive and sliding wear resistance of various amorphous metal 0 4 5 6 7 8 alloys to their microhardness values, 10 10 10 10 10 Cycles to Failure Greer and Myung17 demonstrated a one- Figure 6. Stress-life curves for cyclic uniaxial loading of glassy Zr-Ti-Co-Be in air (•) and 0.6 to-one correspondence between wear M NaCl (|)33 and for 316 stainless steel in air (•) and 0.5 M NaCl (|).34 resistance and hardness. Moreover, the

Vol. 62 No. 2 • JOM www.tms.org/jom.html 85 wear resistance of amorphous metals rial before fracture, and is a critical propagating plastic-fl ow bands (shear was found to exceed that of ceramics at property determining the overall me- bands) prior to turning into opening the same hardness, which suggests that chanical performance of a load-bear- cracks, and consequently their frac- the mechanism of wear deformation in ing implant. Despite their excellent ture toughness is generally lower than amorphous metals does not involve brit- biological compatibility and high wear values typical of structural crystalline tle fracture. The high wear resistance resistance, ceramics are not the materi- metals. Some amorphous metal alloys, of amorphous metals implied by their als of choice in hard-tissue prosthesis however, tend to generate highly dense high hardness could be of great benefi t because of their low fracture toughness shear band networks upon yielding that 1/2 in load-bearing biomedical implant ap- (KIC < 10 MPa m ). Instead, the pre- are able to self-arrest in a manner that plications, as it presents a potential for ferred materials for such applications leads to a higher toughness, approach- reducing wear and mitigating the asso- are metals, because despite an inferior ing values characteristic of structural ciated adverse biological effects. biological compatibility, metallic bio- crystalline metals. Amorphous metals materials exhibit very high toughness therefore exhibit toughness values that Toughness 1/2 (KIC ~ 100 MPa m ). The atomic struc- vary from as low as the values charac- Fracture toughness is a measure of ture of an amorphous metal lacks mi- teristic of brittle ceramics to as high as the load-bearing capacity of a mate- crostructural mechanisms for arresting the values characteristic of engineer- ing metals.15 The fracture toughness of BIOCOMPATIBILITY AND CURRENT STATE OF THE ART certain amorphous metals is plotted in Figure 3 against their respective yield Biocompatibility is regarded as the ability of a biomaterial to perform with an appropri- 18 ate biological response.2 The aspects that defi ne biocompatibility are the “host response” strengths, together with data for con- and the “material response.” The “host response” is defi ned as the local and systemic re- ventional crystalline metallic biomate- sponse (other than the intended therapeutic response) of the living systems to the material, rials.19 It should be noted that many of while the “material response” is the response of the material to the living system.2 The most the toughness data for amorphous met- basic material responses encountered in load-bearing implant applications include corro- als shown in Figure 3 do not represent sion or dissolution due to chemical attack, friction and wear, and mechanical failure due to actual KIC values (because of sample plastic deformation, fracture, or fatigue. Typical host responses include tissue adaptation, sizes insuffi cient to warrant plane- which can be either positive (e.g., osseointegration) or negative (e.g., stress shielding), strain conditions or lack of pre-existing infl ammation, allergic response, and carcinogenesis. fl aws), and are thus not directly com- Owing to their high strength and hardness and superior fracture and fatigue resistance, metallic materials demonstrate a much more favorable mechanical response to biological parable to those of crystalline bioma- systems than polymeric or ceramic materials, and are presently regarded as the materials terials. Nevertheless, some amorphous of choice for load-bearing implant applications. The most common metallic biomaterials metal systems appear to exhibit frac- used traditionally in load-bearing implant applications are stainless steels, Co-Cr-based ture toughness that can be adequate for alloys, and Ti-based alloys, while recently certain Zr-based and Ta-based alloys have been load-bearing biomedical implant ap- introduced.1 Despite demonstrating an overall favorable mechanical performance, metallic plications. Interestingly, reinforcement orthopedic biomaterials have been associated with certain adverse local and remote tissue of the glassy structure with a ductile responses. For example, the stiffness of these materials, particularly of Co-Cr alloys and dendritic crystalline phase provides a stainless steels, is very high compared to that of hard tissue (Young’s modulus of Co-Cr- unique microstructural mechanism for based alloys is 200Ð250 GPa vs. 10Ð20 GPa for cortical bone). Consequently, the modulus mismatch and uneven load sharing between implant and bone, known as stress shielding, arresting propagating shear bands that 12,13 can lead to bone resorption and eventual loosening of the implant.3 Another issue critical to leads to superb fracture toughness. the long-term performance and longevity of the joint replacement prosthesis is the degrada- As presented in Figure 3, the fracture- tion products of these biomaterials and the subsequent tissue reactions to these products. toughness values of the highly tough- The degradation processes most frequently associated with adverse biological reactions ened ductile-phase-reinforced amor- and implant failures are wear and corrosion. Particulate debris generated by metal-on- phous metals,12 which were determined metal wear is known to induce infl ammatory responses that often lead to clinical failures. 20 to roughly represent actual KIC values, For example, foreign-body granulation tissue response to wear debris has the ability to clearly surpass those of crystalline met- invade the bone-implant interface and initiate a progressive local bone loss (wear-debris- als. The very high strength and tough- induced osteolysis) that can threaten the fi xation of the devise and limit the longevity of the total joint prosthesis.4 Electrochemical corrosion occurring at metal-implant surfaces can ness of these materials is an attractive have two main adverse effects: the degradation process may reduce the structural integrity combination that could lead to superior of the device, and the degradation products (e.g., metal ions) may remotely or systemi- biomechanical performance. cally affect the biological host. Structural failure mechanisms related to corrosion include Fatigue stress-corrosion cracking, corrosion fatigue, and fretting corrosion.5 Metal ions that have been associated with adverse biological reactions include Ni (contained in relatively high Fatigue endurance is perhaps the fractions in certain stainless steels), Al (common in Ti-based alloys), Co (the base metal of single most important property that Co-Cr-based alloys), and Cr (contained in relatively high fractions in stainless steels and dictates the overall service life of a Co-Cr-based alloys). Hypersensitivity to Ni is the most common (affects approximately load-bearing implant. It is estimated 14% of the population), followed by Co and Cr.6 Nickel, Co, and Cr are known carcinogens in pure form,7 while Ni and Cr form certain compounds that are found to be carcinogenic.8 that the average non-active person Aluminum has been linked to several neurodegenerative diseases such as Parkinsonism imposes several millions of cycles of dementia and Alzheimer’s disease.9 stress on the hip joint per year.19 Over a period of 20Ð30 years, this rate cor-

86 www.tms.org/jom.html JOM • February 2010 responds to approximately 108 load- rosion rate of surgical-grade 316L stain- CORROSION BEHAVIOR OF ing cycles. Therefore, the resistance of less steel, Ti-6Al-4V, and Co-Cr-Mo in AMORPHOUS METALS a material to cyclic loading is critical the same solution, the corrosion rate of in ensuring longevity of the device. Be-bearing Zr-based glasses varies from Static Corrosion The fatigue performance of amor- a much higher value in the case of Zr- phous metals has not been broadly The chemically homogeneous struc- Ti-Ni-Cu-Be to a low comparable value investigated. Most investigations were ture of amorphous metals lacking mi- in the case of Zr-Ti-Be.28 In phosphate- performed on Zr-based glasses. The crostructural defects such as disloca- buffered saline, Al-bearing Zr-based fatigue endurance limit for this class tions, grain boundaries, and precipi- glasses were also found to exhibit cor- of amorphous metals at 107 cycles is tates/segregates is thought to promote rosion rates comparable to conventional plotted in Figure 4 against the corre- the formation of uniform passive layers biometals.29 sponding yield strengths.21 The data in- without weak points that can provide en- Several amorphous Fe-based alloys cludes results from uniaxial, bending, hanced protection against corrosion or are known to exhibit superb resistance and rotating fatigue experiments. Data chemical attack. From an electrochemi- to corrosion.30 For example, Fe-Cr-P-C for conventional crystalline biometals cal standpoint, amorphous metal alloy glasses are found to exhibit corrosion are also presented on the same plot.22 families can be grouped into early-tran- rates in acidifi ed chloride solutions The ratio of the fatigue endurance limit sition-metal based (e.g., Zr-based), fer- many orders of magnitude lower than to the material yield strength is known rous-metal based (e.g., Fe-based), and conventional (crystalline) stainless-steel as the fatigue ratio. Unlike crystalline noble-metal based (e.g., Pd-based). Zir- alloys.31 These Fe-based glasses appear biometals which exhibit rather consis- conium-based and Fe-based amorphous to resist pitting when anodically polar- tent endurance limits and high fatigue metals are the most extensively studied ized. Their high corrosion resistance is ratios, the endurance limits of Zr-based in corrosion. The nature of corrosion re- attributed to their ability to passivate glasses are shown to be highly scattered sistance of these glasses is directly relat- spontaneously upon immersion, form- and the fatigue ratios ranging from very ed to the formation of a passive surface ing thick, uniform, and chemically ho- small (~0.1) to rather high (~0.5). The layer.24 Noble-metal based amorphous mogeneous passive layers enriched with origin of such high scatter in the endur- metals have not been extensively stud- corrosion-resistant elements. Owing to ance limit of Zr-based glasses is not yet ied, and the key mechanism controlling their exceptionally high corrosion resis- clear. Extrinsic factors such as varia- their corrosion behavior is unknown. tance, these amorphous Fe-based alloys tions in loading geometry and loading In chloride environments, Zr-based may be attractive alternatives to conven- frequency or intrinsic factors such as glasses (both Be-bearing and Al-bear- tional stainless-steel biomaterials. variations in alloy composition, defect ing) exhibit a rather low resistance to pit Stress/Fatigue Corrosion concentration, or glass temperature initiation and growth and a rather weak history are possible causes of the ap- repassivation ability.25,26 The corrosion Stress corrosion and fatigue corrosion parent inconsistency. Since fatigue data resistance of Zr-based glasses relies on failures occur as a consequence of the for families other than the Zr-based the formation of barrier-type protective combined interaction of electrochemi- family (e.g., Fe-based or Pd-based) are layers several nanometers thick, consist- cal reactions and mechanical damage. rather scarce, no defi nitive conclusions ing mainly of Zr oxides and often small Resistance to such failure is an im- can be drawn for the overall fatigue be- fractions of Al or Be oxides. Other late portant attribute of a load-bearing bio- havior of amorphous metals. It can be transition metals often included in the medical implant. Studies of stress and fi rmly stated, however, that unless the composition of these alloys (e.g., Ni and fatigue corrosion for amorphous metals amorphous metal alloy demonstrates a Cu) are enriched around sub-surface lay- are limited to mostly Zr-based and Fe- fatigue-endurance limit that is reliable ers, as a result of selective surface oxi- based. Zirconium-based glasses (both and reproducible and represents a sig- dation.27 Wiest28 recently demonstrated Be-bearing and Al-bearing) are found nifi cant fraction of its yield strength, its that by properly selecting these late to exhibit substantially lower fatigue ra- use as load-bearing implant material transition metals and optimizing their tios in saline solutions compared to their would be severely hindered. Interest- atomic fractions, the overall corrosion values in air, a consequence of consider- ingly, the highly toughened ductile- resistance of these alloys in chloride ably higher (2Ð3 orders of magnitude) phase-reinforced amorphous metals environments can be dramatically en- fatigue-crack growth rates realized in (data also plotted in Figure 4) appear hanced. As shown in Figure 5, the cor- saline compared to air.32,33 In Figure 6 to perform substantially better than rosion resistance of Zr-based Be-bearing the stress-life curves for glassy Zr-Ti- monolithic amorphous metals in terms glasses in an acidifi ed chloride solution Co-Be in air and saline solution34 are of fatigue endurance exhibiting a fa- is drastically enhanced when late-tran- contrasted to the respective curves for tigue ratio comparable to crystalline sition-metals Ni and Cu are replaced 316 stainless steel.35 The two materials biometals.23 The high and consistent by Co and the overall concentration is exhibit comparable fatigue ratios in air endurance limit of ductile-phase-re- reduced. With the aid of x-ray photo- (~0.55). In saline, however, glassy Zr- inforced metallic glasses is attributed electron spectroscopy, Wiest observed Ti-Co-Be exhibits a much lower fatigue to the same “microstructural arrest” more chemically-uniform passive layers ratio (~0.1) than stainless steel (~0.4). mechanism that controls the toughness on systems with low fractions of late- The mechanism governing the corrosion of these alloys.23 transition metals. Compared to the cor- fatigue behavior of these glasses was

Vol. 62 No. 2 • JOM www.tms.org/jom.html 87 BIOLOGICAL COMPATIBILITY OF AMORPHOUS METALS A set of testing procedures is required to study the biological be adequate and compatible (Figure A). In vitro cytotoxicity and compatibility of a biomaterial. National and international stan- cellular-adhesion investigations for other amorphous metal alloys dards organizations provide assistance in the design of biomateri- have been reported,10,40-44 most of which yielded mostly favorable als, including the American Society for Testing Materials (ASTM) results. and the International Standards Organization (ISO). Biocompat- In-vivo Investigations ibility tests can typically be grouped into in-vitro and in-vivo. In- vitro evaluations typically include cytotoxicity and cellular adhe- A standard in-vivo biocompatibility test (ISO protocol 10993 sion tests. In-vivo tests involve tissue compatibility assessments, part 6: Tests for local effects after implantation) was conducted and include evaluation of (i) sensitization, irritation, and reactivity, by NAMSA on Zr-Ti-Co-Be and Pd-Ag-P-Si amorphous alloys. (ii) systemic, subchronic, and chronic toxicity, (iii) genotoxicity, In the test, sterilized metallic-glass test samples were implanted in (iv) carcinogenicity, (v) hemocompatibility, and (vi) immune re- the muscle tissue of three male New Zealand White rabbits. Af- sponses.39 ter administering buprenorphine and a general anesthetic to each animal, one incision was made on each side of the back through In-vitro Investigations the skin and parallel to the lumbar region of the vertebral column. A standard in-vitro biocompatibility test (ISO protocol 10993 Four metallic-glass test article sections and four high-density poly- part 5: Test for cytotoxicity Ð in vitro methods) was conducted ethylene negative control sections were implanted in the right and by NAMSA on Zr-Ti-Co-Be and Pd-Ag-P-Si amorphous alloys. left paravertebral muscle of each rabbit. The skin incisions were In the test, a single extract of the metallic-glass test article was closed with tissue glue. The rabbits were euthanized two weeks prepared using single-strength Minimum Essential Medium sup- later, muscle tissues were excided and the implant sites were ex- plemented with 5 vol.% serum and 2 vol.% antibiotics. The test amined microscopically and macroscopically. No signifi cant mac- extract was placed onto three monolayers of L-929 mouse fi bro- roscopic reactions were detected for either material compared to

blast cells incubated at 37C in a 5% CO2 atmosphere for 48 hours. the negative control implant material. The tissue responses to the Following incubation, the cultures were examined microscopically Zr-Ti-Co-Be articles observed microscopically were similar to to evaluate the cellular characteristics and percent lysis. The test those for the negative control, while those for Pd-Ag-P-Si were extracts for both amorphous metal alloys showed no evidence of slightly more favorable. Both materials were classifi ed as non-ir- causing toxicity. Specifi cally, the performance of both materials ritant. A follow-up study involving amorphous Pd-Ag-P-Si rods received the grade of zero, which corresponds to no reactivity, no implanted subcutaneously in rats for 28 days verifi ed the superb cell lysis, and all cultures appear as discrete intracytoplasmic gran- biological performance of this material, as no tissue reaction was ules. In another 7-day in vitro cellular adhesion test where amor- detected over the 28-day period of the experiment (Figure B). In phous Zr-Ti-Co-Be discs were immersed into a fi broblast-enriched vivo tissue compatibility investigations for other amorphous metal environment, the material demonstrated a favorable biological alloys have been reported,10,44 most of which yielded mostly favor- performance as cell adherence and proliferation were observed to able results.

3.5

) 3 4 2.5 2 1.5 1 Cell Number (10 0.5

HDPE Zr-Ti-Co-Be Figure A. Fibroblast cell attachment on amorphous Zr-Ti-Co-Be and high-density polyethylene (HDPE) discs. (a) Micrograph of the amorphous metal surface after 7 days (arrows point to cell-layer Figure B. Micrographs of the tissue surrounding a Pd-Ag-P-Si rod buildup at the interface). (b) Cell proliferation on the amorphous (circular region in the center of the image) implanted intramuscularly metal and plastic surfaces. in rat for 28 days: (a) 40× magnifi cation, (b) 100× magnifi cation.

88 www.tms.org/jom.html JOM • February 2010 identifi ed to be stress-corrosion crack- ing involving stress-assisted anodic dis- solution at the crack tip.33,36 Overall, the poor corrosion-fatigue performance of Zr-based glasses in saline environments has been attributed primarily to their electrochemistry, which yields passive layers that lack the ability to rapidly repassivate in solution thereby provid- ing only limited protective effect at the crack tip against chloride attack.36 By contrast, certain Fe-based glasses known to be immune to pitting corro- sion demonstrate rapid repassivation ability in chloride solutions. In passive fi lm damage/reformation experiments, the repassivation of glassy Fe-Ni-Cr-P- Figure 7. (a) Amorphous Pd-Ni-Cu-P foam (88% porosity).51 (b) and (c) Micrographs of the C in an acidifi ed chloride solution was cellular structure at high magnifi cations. measured to be more rapid than con- 35 ventional (crystalline) stainless-steel alloys.37 Owing to their ability to repas- sivate rapidly at a crack tip, these alloys 30 exhibit good stress-corrosion character- istics. For example, amorphous Fe-Cr- 25 Ni-P-C is found to exhibit very limited stress-corrosion cracking and embrittle- 20 ment in tensile tests performed in neutral sodium chloride solutions or in acidic 15 solutions with low chlorine concentra- Stress (MPa) tion.38 Such features are highly desirable in load-bearing biomedical implant ap- 10 plications. See the sidebar on page 88 for details 5 on the biological compatibility of amor- phous metals. 0 0 0.005 0.01 0.015 0.02 0.025 0.03 0.035 0.04 POROUS AMORPHOUS Strain METALS Figure 8. Compressive loading responses of amorphous Pd-Ni-Cu-P foam51 and cancellous bone.52 Porosity is thought to be a form of 5 surface irregularity, and its importance 10 in bioprosthesis in guiding cells and aid- ing tissue repair is increasingly grow- ing. Interconnected-pore architectures with pore sizes in the range of 75Ð250 μm are thought to be optimal for tissue 4 ingrowth.45 In addition to eliciting cell 10 attachment and tissue ingrowth, poros- ity functions to reduce the structural properties of monolithic bulk materials (e.g., strength and stiffness) to values

closer to those of natural bone. Stiffness (MPa) Compressive Owing to their high strength com- 103 bined with relatively low moduli, amor- phous metals can be thought as attrac- tive base materials for developing strong 101 102 highly-elastic porous solids capable of Compressive Strength (MPa) matching the mechanical properties of Figure 9. Compressive stiffness vs. strength data for amorphous Pd-Ni-Cu-P foam51 shown cancellous bone. Methods for develop- together with data for porous Ti,53 porous Ta,54 and cancellous bone.52 ing porous amorphous metals with ei-

Vol. 62 No. 2 • JOM www.tms.org/jom.html 89 ther open- or closed-pore architectures tion. These include a low toughness for substitution may not be easy. A Zr-based and porosities exceeding 90% have been some Fe-based glasses that could limit glass should also be adequately tough reported over the past fi ve years (Figure their overall mechanical performance, and have high and consistent fatigue 7).46Ð49 At a specifi c porosity level, these an inconsistent fatigue-endurance limit endurance, a criterion that could be met cellular structures exhibit a modulus for Zr-based glasses that could render through careful alloying and proper test- that roughly refl ects the low modulus their fatigue performance unreliable, ing. Lastly, the resistance to pitting cor- of the amorphous metal. Their strength, and a low pitting-corrosion resistance rosion and stress-corrosion cracking of however, varies according to the mecha- and poor stress/fatigue corrosion perfor- these alloys should be high. The poor nism accommodating failure, and does mance for Zr-based glasses that could resistance to these reactions has been not always refl ect the high strength of limit their use in biological solutions. attributed to the overall electrochemical the amorphous metal.50 It has recently Another critical aspect is the chemical character of these glasses,36 therefore been demonstrated that by matching the composition of these alloys. Almost all overcoming this defi ciency may not be characteristic length scales controlling known amorphous metal alloy compo- easy. Nevertheless, the value in develop- structural failure, the porous amorphous sitions contain elements that are known ing a biocompatible Zr-based glass will metal can inherit the high strength of to cause adverse biological reactions. be enormous, as it would represent a the monolithic amorphous metal over a For example, most known compositions high-hardness low-modulus alternative broad range of porosity, and thus emerge contain Ni and/or Cu. Both elements to biometals of comparable cost such as as one of the strongest porous solids are highly allergenic and toxic while surgical Ti-based and Zr-based alloys. of any kind.51 As shown in Figure 8, a Ni is possibly carcinogenic, and should Interestingly, many of the defi ciencies strong highly-elastic porous amorphous be avoided in long-term implantations of Zr-based glasses can be overcome by metal is able to replicate the mechanical that involve wear and corrosion reac- reinforcing the glassy phase with a duc- response of cancellous bone, roughly tions. All bulk-glass-forming Zr-based tile dendritic crystalline phase. Ductile- matching its stiffness, strength, and compositions contain either Al or Be. phase reinforced Zr-based glasses were plastic deformability. More importantly, Aluminum is a known neurodegenera- found to exhibit superb toughness12,13,20 as shown in Figure 9, porous amor- tive toxin linked to Alzheimer’s disease and a high and consistent fatigue-en- phous metals of varying porosity match and is highly undesirable even in small durance limit.23 It is not obvious how the strength and modulus of cancellous quantities. Beryllium is a known re- the ductile-phase reinforcement affects bone over a broad range. Conventional spiratory toxin causing chronic lung their corrosion behavior and stress/fa- porous metals such as porous Ti and Ta, disease when inhaled, and despite the tigue corrosion performance, as no on the other hand, which exhibit moduli lack of conclusive evidence of Be tox- corrosion investigations have been per- and strengths that refl ect the high modu- icity in the bloodstream, Be remains formed to date. Also, the highly tough- lus and low strength of their base met- biologically controversial. Lastly, cost ened compositions reported to date con- als, are not able to match the range of is another critical issue. Several noble- tain Be (albeit in small concentrations), stiffness and strength of cancellous bone metal-based (Pd, Pt, and Au) glasses are which may be problematic in terms of by varying porosity (Figure 9). Porous known, which exhibit rather high tough- long-term biocompatibility. Nonethe- amorphous metals can hence be thought ness. Even though no extensive corro- less, a corrosion-resistant ductile-phase as attractive scaffold materials for bone- sion studies have been reported to date, reinforced Zr-based glass (preferably ingrowth applications. their corrosion resistance is expected to Be free) will represent a superior load- be rather high as well. Owing to their bearing implant material with structural DESIGN CRITERIA FOR high cost, however, their extensive use properties that cannot be matched by BIOCOMPATIBLE in hard-tissue prosthesis would be hin- any other known biomaterial. AMORPHOUS METALS dered and may be limited to dental ap- Iron-based glasses bearing Cr are Based on the discussion in the pre- plications. known to exhibit superb corrosion re- ceding sections, amorphous metals ex- Recognizing the above defi ciencies, sistance31 as well as good stress-corro- hibit certain universal properties that a set of criteria for designing new bio- sion-cracking resistance.38 Bulk-glass- could be attractive for load-bearing compatible amorphous alloys can be forming Fe-based compositions exhibit biomedical-implant applications. These laid out. The “structural” amorphous very high strength and hardness but include a very high strength, which metals can be grouped into three major unfortunately tend to be brittle. Re- gives rise to high hardness and a poten- alloy families: early-transition-metal cently a series of amorphous Fe-based tial for improving wear resistance over based (e.g., Zr-based), ferrous-metal compositions has been reported56 which existing hard-tissue prosthesis metals, based (e.g., Fe-based), and noble-metal exhibit a level of toughness that may be and a relatively low modulus, which based (e.g., Pd-based). adequate for load-bearing implant ap- gives rise to high elasticity and holds a Zirconium-based systems free of Ni plications. Some of these compositions promise for mitigating stress shielding. and Cu have been developed;55 however, are very low cost, are free of undesirable There are also non-universal properties no Al or Be free bulk-glass-forming elements such as Ni, and bear Cr which specifi c to particular amorphous metal compositions have been reported. Alu- is highly desirable for good corrosion alloy systems that are below acceptable minum and Be were proven to be vital and stress/fatigue corrosion behavior. limits for hard-tissue prosthesis and elements that critically facilitate glass These Fe-based glasses may be attrac- could hinder their use in such applica- formation in those systems, hence their tive ultra-hard low-modulus alternatives

90 www.tms.org/jom.html JOM • February 2010 to existing stainless-steel biomaterials. sistance to corrosion and stress/fatigue 26. M.L. Morrison et al., Intermetallics, 12 (2004), pp. 1177–1181. The ferromagnetic characteristics ex- corrosion, while most contain elements 27. A. Gebert et al., Materials Science and Engineering, hibited by many Fe-based compositions that have been associated with adverse A 267 (1999), pp. 294–300. however may render their biomedical biological reactions. It appears that 28. A. Wiest (Ph.D. thesis dissertation, California Institute of Technology, 2009). use problematic. none of the presently known composi- 29. M.L. Morrison et al., J. Biomedical Materials Lastly, a tough corrosion-resistant tions collectively exhibits the entire set Research, Part A, 74A (2005), pp. 430–438. bulk-glass-forming Pd-based alloy free of biocompatibility requirements vital 30. Y. Waseda and K.T. Aust, J. Materials Science, 16 (1981), pp. 2337–2359. of Ni and Cu would be desirable for den- for hard-tissue prosthesis. Nevertheless, 31. K. Hashimoto, Corrosion, 58 (2002), pp. 715–722. tal applications. Noble-metal glasses are the ability of various amorphous metal 32. V. Schroeder, C.J. Gilbert, and R.O. Ritchie, Scripta generally tough and are expected to be alloys to compete or rival the present Materialia, 40 (1999), pp. 1057–1061. 33. M.L. Morrison et al., Materials Science and highly corrosion resistant. Also, since state of the art in various biocompatibil- Engineering, A 467 (2007), pp. 198–206. their glass-forming ability is controlled ity areas suggests that development of a 34. A. Wiest et al., Scripta Materialia (in press). mostly by metal/metalloid interactions, highly biocompatible amorphous-metal 35. Y.R. Qian and J. R. Cahoon, Corrosion Science, 53 (1997), pp. 129–135. substitution of Ni and Cu (which are implant material may be within reach. 36. V. Schroeder and R.O. Ritchie, Acta Materialia, 54 present in all known compositions) with (2006), pp. 1785-1804. other less harmful late-transition-met- References 37. R.B. Diegle, Corrosion, 35 (1975), pp. 250–258. 38. A. Kawashima, K. Hashimoto, and T. Masumoto, 1. N.J. Hallab, J.J. Jacobs, and J.L. Katz, Biomaterials als may be easily attainable. Such Pd- Corrosion Science, 16 (1976), pp. 935. Science, An Introduction to Materials in Medicine, based glass will represent an attractive 39. J.M. Anderson and F.J. Schoen, Biomaterials 2nd ed., ed. B.D. Ratner et al. (Amsterdam: Elsevier Science, An Introduction to Materials in Medicine, 2nd high-hardness low-modulus alternative Academic Press, 2004), pp. 526–555. edition, ed. B.D. Ratner et al. (Amsterdam: Elsevier 2. J. Black, Biological Performance of Materials: to existing dental Pd-based alloys. Academic Press, 2004), pp. 360–367. Fundamentals of Biocompatibility (Boca Raton, FL: A notable recent advance in the de- 40. S. Buzzi et al., Intermetallics, 14 (2006), pp. 729– CRC Press, 2006). 734. sign of biocompatible amorphous metal 3. R. Huiskes, H. Weinans, and B. Vanrietbergen, Clinical 41. C.L. Qiu et al., Scripta Materialia, 55 (2006), pp. Orthopaedics and Related Research, 274 (1992), pp. alloys is the development of biodegrad- 605–608. 57 124–134. able Mg-Zn-Ca glass. Owing to the 42. L. Liu et al., Scripta Materialia, 58 (2008), pp. 231– 4. J.J. Jacobs et al., J. American Academy of Orthopedic 234. liquid-like atomic structure of the glassy Surgeons, 2 (1994), pp. 212–220. 43. L. Liu et al., J. Biomedical Materials Research Part 5. K.J. Bundy, C.J. Williams, and R.E. Luedemann, metal allowing broad solubility of alloy- A, 86A (2008), pp. 160–169. Biomaterials, 12 (1991), pp. 627–639. ing elements, a glass-forming composi- 44. L. Liu et al., Intermetallics, 17 (2009), pp. 235–240. 6. D.A. Basketter et al., Contact Dermatitis, 28 (1993), 45. R.M. Pillar, J. Biomedical Materials Research, 21 tion was designed with corrosion char- pp. 15–25. (1987), pp. 1–33. 7. D. Beyersmann, Toxicology Letters, 72 (1994), pp. acteristics in biological solutions (e.g. 46. J. Schroers et al., J. Applied Physics, 96 (2004), pp. 333–338. degradation rate and hydrogen evolu- 7723–7730. 8. A. Hartwig, Toxicology Letters, 102-103 (1998), pp. 47. A.H. Brothers, R. Scheunemann, and D.C. Dunand, tion rate) optimal for a biodegradable 235–239. Scripta Materialia, 52 (2005), pp. 335–339. 9. M. Kawahara, J. Alzheimers Disease, 8 (2) (2005), implant. Although glassy Mg-based al- 48. T. Wada and A. Inoue, Materials Transactions, 45 pp. 171–182. loys are generally not considered “struc- (2004), pp. 2761–2765. 10. J. Schroers et al., JOM, 61 (9) (2009). pp. 21–29. σ 49. M.D. Demetriou et al., Applied Physics Letters, 91 turally advanced” ( y < 1 GPa; KIC < 10 11. W.L. Johnson and K. Samwer, Physical Review (2007), 161903. 11,18 Letters, 95 (2006), 195501. MPa), their strength and elasticity 50. M.D. Demetriou et al., Advanced Materials, 19 12. D.C. Hofmann et al., Nature, 451 (2008), 1085-U3. far surpass those of conventional (crys- (2007), pp. 1957–1962. 13. D.C. Hofmann et al., Proceedings of the National 51. M.D. Demetriou et al., Physical Review Letters, 101 talline) biodegradable Mg-based alloys, Academy of Sciences of the United States of America, (2008), 145702. 105 (2008), pp. 20136–20140. and are thus thought of as attractive al- 52. T.M. Keaveny et al., J. Biomechanics, 27 (1994), pp. 14. S.W. Yu et al., J. Bone and Joint Surgery-American, ternatives for biodegradable implant ap- 1137–1146. 57 (1975), pp. 692–695. 53. I-H. Oh et al., Scripta Materialia, 49 (2003), pp. plications. 15. M.F. Ashby and A.L. Greer, Scripta Materialia, 54 1197–1202. (2006), pp. 321–326. 54. D.A. Shimko et al., J. Biomedical Materials Research CONCLUSION 16. W.L. Johnson, unpublished data collection (2005). Part B, 73B (2005), pp. 315–324. 17. A.L. Greer and W.N. Myung, Supercooled Liquid, 55. A. Wiest et al., Acta Materialia, 56 (2008), pp. The lack of long-range crystalline Bulk Glassy and Nanocrystalline States of Alloys, MRS 2625–2630. 2000 Fall Meeting Proceeding, ed. A. Inoue et al., 644 order in the atomic structure of amor- 56. M.D. Demetriou et al., Applied Physics Letters, 95 (2000), L10.4. phous metals and the absence of asso- (2009), 041907. 18. J.J. Lewandowski, W.H. Wang, and A.L. Greer, 57. B. Zberg, P. J. Uggowitzer, and J. F. Löffl er, Nature ciated microstructural defects give rise Philosophical Magazine Letters, 85 (2005), pp. 77–87. Materials, 8 (2009), pp. 887–891. to certain universal properties, such as 19. J.R. Davis, Handbook of Materials for Medical high strength, hardness, and elasticity Devices (Materials Park, OH: ASM International, 2003). Marios D. Demetriou, senior research fellow, Aaron 20. M.E. Launey et al., Applied Physics Letters, 94 Wiest, civilian scientist with the Department of the that are considered very attractive for (2009), 241910. Navy, Douglas C. Hofmann, visiting scientist, and load-bearing biomedical implant appli- 21. G.Y. Wang, P.K. Liaw, and M.L. Morrison, William L. Johnson, professor of engineering and cations. In addition to these desirable Intermetallics, 17 (2009), pp. 579–590. applied science, are with California Institute of 22. J.B. Brunski, Biomaterials Science, An Introduction Technology, 1200 E. California Blvd., Pasadena, universal properties, amorphous metal to Materials in Medicine, 2nd edition, ed. B.D. Ratner CA 91125; Bo Han, assistant professor of research, alloys demonstrate certain composition- et al. (Amsterdam: Elsevier Academic Press, 2004), pp. is with the University of Southern California, Los specifi c properties that make them less 137–153. Angeles, CA; Nikolaj Wolfson, orthopedic surgeon, 23. M.E. Launey et al., Proceedings of the National is with Mills-Peninsula Health Services, and the appealing for such applications. Some Academy of Sciences of the United States of America, California Pacifi c Medical Center, San Francisco, amorphous metal alloys have tough- 106 (2009), pp. 4986–4991. CA; and Gongyao Wang, research associate, and ness and fatigue-endurance limits that 24. J.R. Scully, A. Gebert, and J.H. Payer, J. Materials Peter K. Liaw, professor of materials science and Research, 22 (2007), pp. 302–313. engineering, are with the University of Tennessee, are low compared to ordinary crystal- 25. W. H. Peter et al., Intermetallics, 10 (2002), pp. Knoxville, TN. Dr. Demetriou can be reached at line metals, others demonstrate low re- 1157–1162. [email protected].

Vol. 62 No. 2 • JOM www.tms.org/jom.html 91 HSC Chemistry® 7 Outotec's new innovative process calculation software contains an updated flowsheet simulation module and a thermochemical database expanded to over 25,000 species. With 22 calculation modules and 12 databases at your fingertips, HSC 7 is an invaluable tool for any process engineer or scientist because one laboratory experiment may cost much more than a single HSC license. Once the compass gave us a competitive edge when navigating in foggy waters. Today modeling and simulation give us a similar advantage when we navigate oceans of data with hundreds of variables. This is the only way to utilize the current massive information overload. The new HSC 7 provides us with an easy simulation tool to steer process development and research. Ask for the 32 page “What’s new in HSC 7” paper from: Outotec Research Oy Email: [email protected], Tel: +358-20-529 211

www.outotec.com/hsc Bulk Metallic Glasses Research Summary Metallic Glasses: Gaining Plasticity for Microsystems

Yong Yang, Jianchao Ye, Jian Lu, Yanfei Gao, and Peter K. Liaw

Since the 1960s, metallic glasses search efforts over the past years, MGs sive experiments at room temperature (MGs) have attracted tremendous re- of bulk form usually show brittle-like (RT) due to the catastrophic shear-band search interest in materials science and fracture in uniaxial tensile or compres- propagation upon yielding. The lack engineering, given their unique combi- of RT ductility in MGs remains as the nation of mechanical properties. How- How would you… longstanding issue to be solved and the ever, the industrial applications of MGs major hurdle hindering the deployment …describe the overall signifi cance have been hindered due to their lack of of this paper? of MGs in various structural applica- 28 ductility in bulk form at room tempera- This article provides an overview of tions. ture. In contrast, it was observed that the recent studies on the size effect in Unlike metallic crystals, the atomic MGs could exhibit excellent plasticity metallic glasses at room temperature. structure of MGs lacks long-range at the small size scale. In this article, The experimental and theoretical order because of the overall random results indicate that metallic glasses, we summarize the related experimental although brittle in bulk forms, can packing of atoms. In some regions, fi ndings having been reported so far to- display extensive plastic fl ows at the the atoms are loosely packed, forming gether with the possible origins of such microscopic scale without sacrifi ce atomic clusters, called the ‘shear trans- of their mechanical strengths. The 7 a size effect in MGs. The enhanced combined superb strength and high formation zones’ (STZs), which are a 16 plasticity of MGs in small volumes, ductility indicate that metallic glasses few nanometers in size and prone to together with their high mechanical are promising for the fabrication of shear transformations upon mechani- strengths and remarkable thermoplas- microsystems of next generation. cal loading. In the others, the atoms tic formability, strongly implies that …describe this work to a are closely packed, forming the elastic MGs are the promising materials for materials science and engineering media that surround the STZs. Upon professional with no experience in fabricating the next generation of mi- your technical specialty? the application of external stresses, the cro- and nano-devices. Unlike crystalline metals, metallic STZs behave in a liquid-like manner on an individual basis to initiate plastic INTRODUCTION glasses deform plastically via shear banding at room temperature. In fl ows in MGs through shear transfor- Metallic glasses (MGs) emerged as a bulk forms, metallic glasses exhibit mations.26 On the other hand, the elastic brittle fracture, suffering from the newcomer to the family of amorphous catastrophic growth of shear bands. media, consisting of all sorts of solid- solids in the 1960s and were fi rst syn- In sharp contrast, metallic-glass like clusters in an amorphous structure, thesized through the rapid quenching of micro-specimens can withstand offer the essential resistance to such an supercooled liquids.1 Compared with extensive plastic deformation without inelastic deformation process.26 When shear fracture. The transition metallic crystals, MGs exhibit much from brittle to plastic deformation the applied shear strain reaches the higher mechanical strengths and better originates from the shear-banding critical value of about ~2% at room corrosion resistance for the absence of dynamics, which entails the shear- temperature,11,12 the shear transforma- crystalline defects, which renders MGs induced materials softening and tion events occurring on the individual 2Ð4 subsequent structural relaxation in promising for structural applications. metallic glasses. STZs percolate and result in the local Since the advent of the fi rst Au-based breakdown of the elastic media, which, …describe this work to a MG, tremendous research efforts have layperson? then, gives rise to the formation of a been stimulated with a gigantic amount Despite brittleness in bulk forms at shear band, namely, a 10-nm thick pla- of research works archived in the lit- room temperature, metallic glasses, nar defect on which severe strain con- erature.1Ð27 Different strategies are seen an emerging material for future centration occurs.20 to have been practiced in the synthe- microsystems, exhibit good plasticity If unhindered, the propagation of the and high strength when machined into sis and characterization of a variety small volumes. Such a size effect in shear band is catastrophic and detri- of MGs aiming to fi nd the recipe that metallic glasses implies that metallic mental to the overall ductility of bulk produces the metallic-glassy system glasses are a good candidate for metallic glasses (BMGs). During the with optimized mechanical properties structural applications in the area shear-band propagation, the viscos- of micro-electro-mechanical systems 6,7 and a critical dimension as large as (MEMS). ity of materials plunges within the possible.5,27 Despite the dedicated re- shear band and is accompanied by a

Vol. 62 No. 2 • JOM www.tms.org/jom.html 93 shape of the desired micro-part after removing the mold material via chem- ical etching (Figure 1b). Figure 1c presents the three-dimensional (3-D) array of the Pd-based MG micropillars fabricated using the micro-imprinting technique. As a comparison, Figure 1d shows the MG micropillar fabricated through focused-ion-beam (FIB) mill- ing, whose geometry has been widely ab adopted in the prior micro-compression experiments.29,31Ð35 It is evident that the present micro-imprinting technique results in signifi cant surface roughen- ing and tip rounding in the MG-based MEMS micropillars, which contrasts the smoothness and sharpness of the FIB-milled micropillars. Until now, re- search efforts have mainly focused on the characterization of the mechanical properties of the FIB-milled micropil- lars whereas reports on the mechanical properties of MEMS micropillars are cd still scarce. Figure 1. The schematics of the fabrication of the MG-based MEMS through the micro/nano-imprinting process: (a) fi lling the mold with the MG liquid at high process MICROMECHANICAL STUDY temperature and (b) etching the mold away to form the MG-based micro-parts after cooling down the process temperature; (c) the scanning-electron-microscopy (SEM) OF METALLIC GLASSES image of the 3-D array of the MG-based MEMS micropillars (by courtesy of Prof. J. Chu of the National Taiwan University of Science and Technology) and (d) the single MG Since the pioneering work of Uchic micropillar fabricated through FIB milling. et al.,42 who used the FIB-milled mi- cropillars to study the size effect in single crystals, the FIB-based micro- temperature surge,9,19 which accompa- allow for the free gliding of shear compression method has been widely nies severe materials softening therein bands across the lateral dimension of employed in the study of the mechani- and subsequent brittle-like fracture. a micro-sample. The stable shear-band cal behavior of a variety of materials in Through various numerical simula- propagation, as already witnessed in small volumes43 and, in the meantime, tions, it has been shown that the stress- the micro-compression/tension tests, the method itself has also received con- induced dilatation triggers the materi- thus implies an enhanced RT plasticity siderable investigations.44Ð46 Although als-softening process at the early stage in MGs at the microscopic scale. it is still debated whether or not the size of the shear-band propagation, whereas The experimental results reported effect witnessed in the micro-compres- the temperature rise is mainly respon- for the micromechanical behavior of sion of single crystals is caused by the sible for the fi nal fracture in BMGs.9,13 MGs are encouraging to the future ap- FIB-induced damage,44Ð47 it is gener- As a consequence, the coupled effect plications of MG-based micro-electro- ally believed that, for amorphous met- of these two materials softening mech- mechanical systems (MEMS), which als that are radiation tolerant, the FIB- anisms renders BMGs vulnerable to emerged recently as a new area in mi- induced surface damage is not severe, the shear-band propagation with little cro- and nanotechnology.39Ð41 By ex- only occupying a 3Ð4 nm thick region RT ductility shown at the macroscopic ploiting the thermoplastic formability in the vicinity of the sample surface if scale.17,21 of MGs, a variety of MG-based micro- an appropriate FIB milling procedure In contrast, the shear-banding pro- and nano-devices have been success- is followed.29,33 In such a case, the cess is much less threatening in MG fully fabricated through the micro- and major concern about the validity of micro-samples than in MG bulk sam- nano-imprinting technique. This fabri- the measured MG micropillars’ yield ples at room temperature. It was ob- cation process involves the compres- strengths is the micropillar’s tapering, served that shear bands propagated in sion of an amorphous MG material which arises due to the ion-beam di- a sluggish manner in the compression into a mold at the processing tempera- vergence and becomes unavoidable for tests of MG micropillars29Ð36 and also ture higher than the glass-transition the submicrometer-sized FIB-milled 31 in the tension tests of MG submicrom- point, Tg, but below the melting point, micropillars. 37 eter-sized strips. Unlike classic na- Tx, of the corresponding MG material To address the taper effect, we re- noindentation tests in which the propa- (Figure 1a). After cooling down the cently extended the shear-plane cri- gation of shear bands is constrained,38 whole system, the amorphous MG ma- terion proposed by Packard et al.10 to the micro-compression/tension tests terial settles in the mold and forms the the case of micro-compression. Pro-

94 www.tms.org/jom.html JOM • February 2010 vided that the MohrÐCoulomb law is strength of MGs can be regarded as a ORIGIN OF SIZE EFFECT IN satisfi ed on the whole shear plane at size-independent mechanical property METALLIC GLASSES the yielding point, the taper effect on on average (Figure 2b), the plastic- the micropillar’s yield strength can be ity of MGs shows the strong size de- From the mechanistic perspective, quantifi ed.36 In principle, the theoreti- pendence. As mentioned above, there a fundamental understanding of the cal calculation shows that the intrin- are numerous reports on the plastic- shear-banding behavior in metallic sic yield strength of a micropillar, as ity enhancement seen in the miniatur- glasses can be attained by considering dictated by the Mohr–Coulomb’s law, ized MG samples, which span a wide the material instability in rate-depen- can be accurately extracted if the MG’s spectrum of MG alloys available to dent plastic solids, which is manifested normal stress coeffi cient, the pillar’s date.29,30,32,33,35Ð37 For instance, the Fe- by the spectral growth of the strain taper angle, and the distance from the based MG appears extremely brittle in fi eld,48 as opposed to the instantaneous shear band to the pillar’s central axis a bulk form, shattered into many piec- instability in rate-independent plas- are known. (The yielding of BMGs es upon yielding (Figure 3a,b). How- tic solids (such as soils and clays) for obeys the Mohr–Coulomb’s law, ever, signifi cant plastic fl ows were which the necessary condition of strain τ τ μ σ which can be written as: 0 = n + n, observed in the Fe-based MG micro- localization corresponds to the loss of τ in which 0 denotes the intrinsic shear pillars due to the stable shear-banding ellipticity of the equations governing τ σ 49 strength; n and n denote the shear and process at the microscopic scale (Fig- the incremental equilibrium. When normal stress acting on the shear plane ure 3cÐe). Such sharp contrast in the materials display rate dependence, with the unit normal of n, respective- plastic deformation behavior across starting from initial imperfections or ly; and μ is the normal stress coeffi - a large length scale for a wide range small spatial fl uctuations, the strain cient. In uniaxial compression experi- of MG compositions implies that the fi eld will gradually grow into the local- ments, μ is related to the shear angle size effect should be a characteristic of ized mode, and this strain-localization θ through μ = 1/tan(2θ).) In practice, the MG plasticity, which needs to be process can be signifi cantly delayed the fi rst two factors can be determined considered in the design of the future by the material-rate sensitivity. The prior to the micro-compression testing, MG-based micro- /nano-devices. amount of the delay in the onset of lo- whereas the last factor can hardly be measured in a precise manner. Howev- er, it can be shown that, in a 3-tapered micropillar, it only introduces a maxi- mum measurement uncertainty of 5% by assuming that shear bands always nucleate at the pillar’s edge, which, on the other hand, simplifi es the measure- ment of the yield strengths of the MG micropillars. Figure 2a presents the yield σ strengths, y, obtained from the Cu-, Figure 2. (a) The measured yield Zr-, Mg-, and Fe-based MG micropil- strengths of the MG micropillars as lars as a function of the pillar’s top di- a function of the pillar’s top diameter for four types of amorphous alloy ameter. Evidently, the yield strengths systems and (b) the variation in of all the MG micropillars exhibit the the measured yield strengths as trend of a modest size effect. The yield normalized by the corresponding Young’s moduli in comparison with strengths increase slightly with the de- those of bulk samples. creasing pillar’s diameter. In the litera- ture, such a size effect in MGs is at- a tributed to the Weibull statistics,31Ð33,36 which indicates that the size shrinkage in MGs contributes to the strength el- evation mainly by reducing the con- tents of the defects of lowest strengths (also called the weakest links). In do- ing so, the pillar’s Young’s modulus is also elevated through the reduction of the weakest-link content.34,36 As a re- σ sult, the ratios of y/E extracted from the micro-compression experiments largely overlap with those reported for 11 bulk samples, as shown in Figure 2b. b Although the normalized yield

Vol. 62 No. 2 • JOM www.tms.org/jom.html 95 calization depends on the kinetics of strain fi eld evolves into narrow bands Consequently, the stable shear-banding strain softening, structural relaxation, inside which the strain fi elds localize process manifests itself as the plastic and strain-rate hardening, as well as on and outside of which the materials ex- fl ow serrations on a load-displacement the imperfections. perience elastic unloading, the equilib- curve as shown in Figure 3e. In macroscopic tests, the plastic strain rium deformation pathway corresponds On the other hand, our understand- in unconstrained shear bands can be ar- to an elastic snap-back instability. In ing of the shear-banding mechanism bitrarily large so that the catastrophic micromechanical tests, the release rate in MGs has advanced greatly over the failure will quickly follow the initia- of the elastic energy in micro-pillars are past decades.6,7,11,19,23,23,52,53 The recent tion of shear bands. With geometrical far less than those in the bulk samples, rate-jump experiments on MGs at dif- constraints, enhanced plasticity can be and the magnitude of the elastic spring ferent temperatures clearly indicate that achieved by blocking the propagation back can be decreased.29,35,50,51 Conse- structural relaxation plays an important of shear bands so that multiple shear quently, the resulting shear-banding role during the propagation of a shear bands take place to accommodate the behavior in samples of reduced dimen- band.53 The effect of structural relax- applied strain fi eld. Therefore, the plas- sions will be in a stable manner even ation is to mitigate or even completely tic strain will be distributed over many under unconstrained conditions, and cancel out the softening effect brought shear bands, and the onset of fracture various kinetic processes involved in about by the shear-induced dilation. As can be delayed and noticeable macro- the evolution of strain fl uctuation can a rate-dependent diffusion process, the scopic plastic strains may be realized. be observed in this extended delay. effect of the shear-induced structural In these cases, rate effects are usually The above argument does not need relaxation depends on the waiting time weak unless the rate approaches the ki- a detailed picture of the atomistic de- between two successive atom migra- netic limit of shear-band initiation. formation mechanisms. If the elastic tion events and, thus, the kinetic energy In contrast to the observations in unloading outside of shear bands re- acquired by atoms at the onset of shear large specimens where unstable shear- sults in a negligible spring back (as banding, which scales with the released band propagation usually leads to cata- in small specimens), and the material elastic energy.35 In quasi-static me- strophic fracture, in small specimens has a rate-dependent strain-softening chanical experiments, because of the the shear bands propagate in a slug- mechanism, be it the stress-assisted dimensional misfi t between the elastic gish, and sometimes stable, manner so evolution of free volumes or STZs, the energy release (from a volume) and that extensive plastic deformation can strain localization process will become plastic energy dissipation (on a plane), be achieved. When a smoothly varying non-catastrophic and sluggish in time. a size effect naturally arises. In small samples, the elastic energy release rate becomes inadequate to drive a shear band into a runaway defect as in large samples, and the shear band is arrested before a complete load drop occurs, which, then, gives rise to the serrated plastic fl ows as seen in microcompres- sion experiments.35 acd In the literature, a variety of consti- tutive models have been proposed by different researchers targeting at the size effect.9,37,42,50,54 Shimizu et al. con- jectured that, once atoms are displaced from their original positions in a shear- banding region, they can still form a new stable structure with their new neighbors for a small shear displace- ment; complete shear fracture occurs only after a large shear displacement be occurs.9 Likewise, Wu et al. proposed a similar concept, termed as the criti- cal shear offset, to rationalize the size 54 Figure 3. The brittle-to-ductile transition in the Fe-based MG as the sample size decreases effect reported in the literature. They from the macroscopic to microscopic scale: (a) the fragmentation fracture of the Fe-based argued that each type of MG possesses BMG bulk sample indicating the RT brittleness at the macroscopic scale (by courtesy of a critical shear offset and stable shear Prof. Z.P. Lu of the University of Science and Technology, Beijing); (b) the sketch of the typical load-displacement curve for brittle fracture in a Fe-based BMG bulk sample; (c) the banding takes place when the actual stable single shear-banding in the Fe-based BMG micropillar with the diameter of ~8 μm; shear offset is less than the critical val- (d) the stable multiple shear-banding in the Fe-based BMG micropillar with the diameter of ue. However, based on the above-men- ~1 μm; and (e) the sketch of the typical load-displacement curve for stable shear-banding 35,52,53 in a Fe-based BMG micropillar. tioned shear-banding kinetics, the so-called critical shear offset, as left by

96 www.tms.org/jom.html JOM • February 2010 length21,22 or the diffusional length.55 Starting from randomly distributed free volumes or STZs, the strain fi elds will evolve from these initial imperfec- tions into mature shear bands, and the spatial spectrum of the growth rate will be determined by the kinetic processes and these material-dependent length scales. Here, these material-dependent length scales correspond to the nucle- ation lengths of shear bands. Speci- mens with sizes comparable to these scales will exhibit a scale dependence of the deformation modes. However, most up-to-date micro-compressions show stable, but still inhomogeneous, Figure 4. The schematic illustration of shear banding through an irreversible energy deformation mode. To probe the homo- barrier crossing event on the confi gurational potential energy landscape of a metallic geneous plastic deformation mode in glass (the dashed black arrow indicates a stable shear-banding process whilst the dashed gray arrow indicates an unstable one). MGs, a sample size much smaller than those of the current micropillars is de- manded. a shear band on a sample surface before by Johnson et al.11 As the strain in MGs CONCLUSIONS shear-banding instability, is just a re- reaches a critical value, the shear trans- fl ection of the competition between the formation occurring on the individual Based on the above discussions, it is shear-induced materials softening and STZs percolates, which then leads to evident that the transition of plasticity subsequent structural relaxation, which the nucleation of a shear band. From to brittleness is intrinsic to the shear- is expected to be rate- and tempera- the perspective of CSM, this strain- band mediated plastic deformation ture dependent rather than a materials controlled shear-band nucleation pro- mode in metallic glasses. With the re- constant. On the other hand, through cess corresponds to the elevation of the duction in the sample size, a metallic energy balance, it can be shown that system’s internal energy from a basin glass tends to exhibit plasticity rather the instantaneous shear offset, Δs, of a point to its nearest saddle point on the than brittleness because of the time propagating shear band scales with the potential energy landscape (Figure 4). delay in the localization process, or, in dimension of a sample as following:34 After passing through the saddle point, other words, because of the recovery the whole system can step down to a effect of structural relaxation. Howev- Δs  H + αD (1) 0 new basin point by dissipating away the er, such a physical picture for the size where H and D0 denote the height and accumulated mechanical energy, which effect relies on the operation of shear diameter of an MG sample, respective- is the case for MG micropillars. Such bands. In this regard, a question that ly, and the pre-factor α accounts for the balanced energy dissipation is achieved naturally arises is whether a fundamen- elastic energy transfer into the sample through the irreversible structural re- tal change in the plastic deformation from its surroundings and depends on laxation process (α-relaxation) as in mode can be experimentally probed by the elastic and size mismatch between the metallic glass-forming liquids.12 In shrinking the sample size to the level the sample and its elastic supports. contrast, if unbalanced energy dissipa- where shear bands cannot nucleate, as When an MG sample is compressed tion occurs, the whole system is unable already witnessed in atomistic simula- between two ideally rigid supports, to fi nd any basin point to rest on the en- tions.56,57 such an elastic transfer is prohibited ergy potential landscape because of the Such a small sample size, which is and α = 0. In other cases, the pre-factor excessive elastic energy released dur- close to the shear-band nucleation (or α is of a fi nite value and the size effect ing the energy-barrier crossing event process zone) length, has not been manifests itself essentially as a geom- and, thus, a shear catastrophe occurs. accessed in the micropillar tests. In etry effect on the plasticity of MGs. Mechanistically, the material rate the literature, it was argued that shear A rather intriguing point made by dependence introduces a length scale bands could not nucleate as the sample Harmon et al. through their study on through the initial imperfections. size was reduced below the shear-band the anelastic to plastic transition in Smaller specimens will have fewer thickness at nucleation,15 which was es- metallic glass-forming liquids is that imperfections for strain localization to timated as ~10 nm for MGs.20,58 How- irreversible deformation occurs via in- initiate a plastic fl ow, which is simi- ever, the analysis of such a shear-band terbasin hopping events on a confi gura- lar to the explanations based on the thickness was based on the scenario of tional potential energy landscape.12 The Weibull distribution.32,33 On the other shear-band nucleation in a bulk mate- similar yielding mechanism is expected hand, length scales can be introduced rial and it can be anticipated that the for metallic glasses according to the co- into the constitutive relationship of the boundary conditions, such as free sur- operative shear model (CSM) proposed material, such as the heat conduction faces, could affect the nucleation of

Vol. 62 No. 2 • JOM www.tms.org/jom.html 97 shear bands at the very small scale.57 To 24. A. Furukawa and H. Tanaka, Nature Mater., 8 45. R. Maab et al., Appl. Phys. Lett., 91 (2007), (2009), pp. 601–609. 131909. fully understand the size effect in MGs, 25. H.W. Sheng et al., Nature Mater., 6 (2007), pp. 46. R. Maab et al., Appl. Phys. Lett., 89 (2006), the nanomechanical testing technique, 192–197. 151905. which targets the MG-based nanostruc- 26. Y.Q. Cheng, A.J. Cao, and E. Ma, Acta Mater., 57 47. D. Kiener et al., Mater. Sci. Eng. A, 459 (2007), pp. (2009), pp. 3253–3267. 262–272. tures, such as the recently synthesized 27. A.L. Greer and E. Ma, MRS Bulletin, 32 (2007), pp. 48. A. Needleman, Comput. Method Appl. Mech. Eng., MG nanowires,59 is still needed. 611–615. 67 (1988), pp. 69–85. 28. A.L. Greer, Materials Today, 12 (1-2) (2009), pp. 49. J.R. Rice, Theoretical and Applied Mechanics, ed. ACKNOWLEDGEMENTS 14–22. W.T. Koiter (Amsterdam: North-Holland, 1977), pp. 29. C.A. Volkert, A. Donohue, and F. Spaepen, J. Appl. 207–220. Y.Y. acknowledges the internal re- Phys., 103 (2008), 083539. 50. S. Xie and E.P. George, Intermetallics, 16 (2008), search fund (Grant # 1-ZV4J), provided 30. Z.W. Shan et al., Physical Review B, 77 (2008), pp. 485–489. 155419. 51. Y.F. Gao and A.F. Bower, Modell. Simul. Mater. Sci. for newly recruited staff, by the Hong 31. B.E. Schuster et al., Acta Mater., 56 (2008), pp. Eng., 12 (2004), pp. 453–463. Kong Polytechnic University for con- 5091–5100. 52. F.H. Dalla Torre et al., Acta Mater., 56 (2008), pp. ducting the current research. J. Lu is 32. C.J. Lee, J.C. Huang, and T.G. Nieh, Appl. Phys. 4635–4646. Lett., 91 (2007), 161913. 53. A. Dubach, F.H. Dalla Torre, and J.F. Loffl er, Acta grateful for the fi nancial support from 33. Y.H. Lai et al., Scripta Mater., 58 (2008), pp. 890– Mater., 57 (2009), pp. 881–892. GRF, the Hong Kong Government, with 893. 54. F.F. Wu, Z.F. Zhang, and S.X. Mao, Acta Mater., 57 the grant number of PolyU 5203/08E. 34. Y. Yang et al., Acta Mater., 57 (2009), pp. 1613– (2009), pp. 257–266. 1623. 55. Y. Shi et al., Phys. Rev. Lett., 98 (2007), 185505. P.K.L. would like to greatly acknowl- 35. J.C. Ye et al., Acta Mater., 57 (2009), pp. 6037– 56. Y.Q. Cheng et al., Acta Mater., 56 (2008), pp. 5263– edge the support of the National Sci- 6046. 5275. ence Foundation International Mate- 36. J.C. Ye et al., Intermetallics (2009), doi:10.1016/ 57. F. Delogu, Phys. Rev. B, 79 (2009), 184109. j.intermet.2009.1008.1011. 58. M.Q. Jiang, W.H. Wang, and L.H. Dai, Scripta Ma- rials Institutes (IMI) Program (DMR- 37. H. Guo et al., Nature Mater., 6 (2007), pp. 735–739. ter., 60 (2009), pp. 1004–1007. 0231320). Y.F.G. acknowledges support 38. U. Ramamurty et al., Acta Mater., 53 (2005), pp. 59. K.S. Nakayama et al., Adv. Mater., 21 (2009), pp. from the National Science Foundation 705–717. 1–4. 39. J. Schroers, Q. Pham, and A. Desai, J. Microelectro- (DMR 0909037). The authors appreci- mechanical Sys., 16 (2) (2007), pp. 240–247. Yong Yang, Jianchao Ye, and Jian Lu are with the ate the provision of testing materials by 40. G. Kumar, H.X. Tang, and J. Schroers, Nature, 457 Department of Mechanical Engineering, The Hong Prof. C.T. Liu, Prof. J.S.C. Jang, Prof. (2009), pp. 868–872. Kong Polytechnic University, Hong Kong, China; 41. J.S Jang et al., Adv. Eng. Mater., 10 (11) (2008), Yanfei Gao and Peter K. Liaw are with the De- Y. Yokoyama, and Dr. G. Wang. pp. 1048–1052. partment of Materials Science and Engineering, 42. M.D. Uchic et al., Science, 305 (2004), pp. 986– The University of Tennessee, Knoxville, TN. Yanfei References 989. Gao also holds a joint position in the Computer 43. R. Dou and B. Derby, Scripta Mater., 61 (2009), pp. Science and Mathematics Division, Oak Ridge Na- 1. K. Klement, R.H. Willens, and P. Duwez, Nature, 187 524–527. tional Laboratory, Oak Ridge, TN. Dr. Yang can be (1960), pp. 869–870. 44. H. Bei et al., Appl. Phys. Lett., 91 (2007), 111915. reached at [email protected]. 2. A.L. Greer, Science, 267 (1995), pp. 1947–1953. 3. M.F. Ashby and A.L. Greer, Scripta Mater., 54 (2006), pp. 321–326. 4. W.H. Peter et al., Intermetallics, 10 (2002), pp. 1157–1162. 5. A. Inoue, Acta Mater., 48 (2000), pp. 279–306. 6. F. Spaepen, Acta Metall., 23 (1977), pp. 407–415. 7. A.S. Argon, Acta Metall., 27 (1979), pp. 47–58. 8. D.B. Miracle, Nature, 3 (2004), pp. 697–712. These proceedings are from The International 9. F. Shimizu, S. Ogata, and J. Li, Acta Mater., 54 (2006), pp. 4293–4298. Symposium on Liquid Metal Processing and 10. C.E. Packard and C. Schuh, Acta Mater., 55 (2007), Casting, held on September 20-23 at the El Dorado pp. 5348–5358. 11. W.L. Johnson and K. Samwer, Phys. Rev. Lett., 95 Hotel in Santa Fe, New Mexico, USA. (2005), 195501. The International Symposium on Liquid Metal 12. J.S. Harmon et al., Phys. Rev. Lett., 99 (2007), 135502. Processing proceedings cover the following areas: 13. C.A. Schuh, T.C. Hufnagel, and U. Ramamurty, Acta Mater., 55 (2007), pp. 4067–4109. • Primary and secondary melt processing including VIM, 14. C.A. Schuh and A.C. Lund, Nature Mater., 2 (2003), VAR, ESR, EBCHR and PAM pp. 449–452. 15. C.A. Schuh, A.C. Lund, and T.G. Nieh, Acta Mater., • Physical property measurements of liquid metals 52 (2004), pp. 5879–5891. • Refi ning, evaporation and gas/metal reactions 16. D. Pan et al., Proc. Natl. Acad. Sci., 105 (39) (2008), pp. 14769–14772. • Casting and solidifi cation of ingots, and mechanisms for 17. H. Zhang, S. Maiti, and G. Subhash, J. Mech. Phys. defect formations during the casting and solidifi cation Solids, 56 (6) (2008), pp. 2171–2187. process 18. B. Yang, C.T. Liu, and T.G. Nieh, Appl. Phys. Lett., 88 (2006), 221911. • Modeling of metallurgical processes, including heat/ 19. J.J. Lewandowski and A.L. Greer, Nature Mater., 5 mass fl ow modeling of liquid metal and solidifi cations TMS (2006), pp. 15–18. Member price $109 20. Y. Zhang and A.L. Greer, Appl. Phys. Lett., 89 • Ceramic, slag and refractory reactions with liquid metals (2006), 071907. • Fundamentals of reactions involving liquid metals in TMS Student 21. M.Q. Jiang and L.H. Dai, J. Mech. Phys. Solids, 57 production processes Member price $84 (2009), pp. 1267–1292. 22. Y.F. Gao, B. Yang, and T.G. Nieh, Acta Mater., 55 • Direct forming of parts from liquid metals (2007), pp. 2319–2327. 23. M.W. Chen, Annu. Rev. Mater. Res., 38 (2008), pp. To order these or related publications, contact TMS: 445–469. E-mail [email protected] • Phone (724) 776-9000, ext. 256 • Fax (724) 776-3770 Meetings Calendar Presenting upcoming meetings and calls for papers

Symposia Proposals and Abstract Submissions: For TMS-sponsored meetings, symposia proposals and abstracts must be submitted through ProgramMaster, the on-line TMS conference and proceedings management system. The system can be accessed at www.tms.org. Using the Meetings & Events pulldown menu, select the Upcoming TMS Meeting for which you wish to submit a symposia proposal or an abstract, and follow the on-line instructions. The society especially encourages the submission of “hot-topic” symposia or special-session proposals on timely or developing subjects. To advance an idea, use the symposium creation form in ProgramMaster. Additional information can be acquired from the TMS Technical Support Services Department, 184 Thorn Hill Road, Warrendale, PA 15086; (724) 776-9000, ext. 212; fax (724) 776-3770.

Offshore Technology II-VI materials and devices; oxides (TCO, Contact: Mickael Kopke, Norddeutsche Conference 2010 dielectrics); novel materials; heteroepitaxy; Affinerie AG; Hoverstrasse 50, 20539 nanostructures; patterned growth and selective Hamburg, Germany; +49-10-78-83-38-01; May 3–6, 2010 area; epitaxy; devices; atomic layer deposition e-mail: [email protected]; www.cu2010 Houston, Texas and epitaxy; virtual substrates and epitaxial .gdmb.de/. The Offshore Technology Conference lift-off; in-situ monitoring and process con- (OTC) is the world’s foremost event for the trol; in-situ etching; characterization; growth Electronic Materials Conference development of offshore resources in the of devices; energy technology (solid state 2010 (EMC 2010) fi elds of drilling, exploration, production, lighting, PV, thermoelectrics, etc.); equip- June 23–26, 2010 and environmental protection. OTC also ment, safety, environmental and production Notre Dame, Indiana offers case studies and updates on the latest issues. Contact: TMS Meeting Services, deepwater projects from majors, indepen- 184 Thorn Hill Road, Warrendale, PA 15086; Electronic materials are defi ned as relating dents, national oil companies, and service (724) 776-9000, ext. 243; email:mtgserv to, produced, or operated by the controlled companies. Technical topics to be addressed @tms.org; www.tms.org/Meetings/Specialty fl ow of electrons through a semiconduc- at OTC .10 include alternative energy (such /icmovpe-xv/home.aspx/. tor, gas, or free space along with those as offshore wind energy, ocean wave energy, relating to devices, systems, or circuits ocean thermal energy conversion, and has Copper 2010 that employ components such as vacuum hydrates), drilling technology, facilities tubes, integrated circuits, or transistors in June 6–10, 2010 and production operations, enhanced oil their design. The TMS Electronic Materials Hamburg, Germany recovery, facilities and production opera- Conference is the premier annual forum tions, fi nancial aspects, fi eld development The theme of Copper 2010 is CopperÐIndica- on the preparation and characterization of concepts, marine geoscience and geohazards, tor of the Progress of Civilization. Copper electronic materials. Held in conjunction materials technology, ocean engineering 2010 is an important technical event for with the Device Research Conference, this resources, offshore pipelines, public policy, engineers, scientists, fabricators, and users conference presents both invited and contrib- regulatory issues, and more. Contact: Off- of copper. Recently copper has become an uted oral presentations, an exhibition, and shore Technology Conference, 222 Palisades important factor in the developing economies related activities. Technical symposia topics Creek Drive, Richardson, Texas 75080; (972) of East Asian countries and can be seen include: chemical and biological sensors: 952-9494; e-mail [email protected]; www as an indicator of the progress of civiliza- materials, interfaces, and integration; dilute .otcnet.org/2010/index.html. tion in each country. All relevant areas of nitride semiconductors; epitaxy; fl exible and copper production and application and their printed thin fi lm electronics; nanotubes and 15th International Conference on surroundings like energy savings, process nanowires; organic-inorganic hybrid pho- Metal Organic Vapor Phase optimization, health and safety issues, cost tovoltaics; solar cell materials and devices; Epitaxy (ICMOVPE 2010) and commerce, recycling, and new devel- thermoelectrics and thermionics; and several opments will be presented in a series of others. Contact: TMS Meeting Services, May 23–28, 2010 parallel sessions. The program will include 184 Thorn Hill Road, Warrendale, PA 15086; Lake Tahoe, Nevada such topics as economics; downstream (724) 776-9000, ext. 243; email:mtgserv The objectives are to provide a forum for fabrication, applications and new prod- @tms.org; www.tms.org/Meetings/Specialty the latest advances in science, technology, ucts; mineral processing; pyrometallurgy; /EMC10/home.aspx. and applications of MOVPE and to bring hydrometallurgy; electrowinning, electro- together researchers from around the world to metallurgy and refi ning; process control, 1st TMS-ABM International discuss and share experiences in the growth, automatization and optimization; recycling, Materials Congress characterization, and device applications of and sustainable development/health, safety July 26–30, 2010 metal organic vapor phase epitaxy. Techni- and environmental control; as well as plenary Rio de Janeiro, Brazil cal topics will include: basic growth studies; sessions of general interest. The conference nitrides; arsenides, phosphides, antimonides will be followed by post-conference tours Held in conjunction with 65th Annual and dilute nitrides; low bandgap materials; and fi eld trips to local industry productions. Congress of ABM (Brazilian Metallurgical,

Vol. 62 No. 2 • JOM www.tms.org/jom.html 99 Materials and Mining Association) and the Third International Conference on in conjunction with the 2010 Conference 18th International Federation for Heat Treat- Uranium —U2010 of Metallurgists, the premier annual event ment and Surface Engineering Congress, of the Metallurgical Society of the Cana- August 15–18, 2010 this inaugural congress will feature seven dian Institute of Mining, Metallurgy and Saskatoon, Saskatchewan, Canada proposed symposia covering important Petroleum (MetSoc) from October 3Ð6, contemporary issues in materials science The plenary session will discuss: The Next 2010 in Vancouver, British Columbia. The and engineering. This congress builds on Generation of Nuclear Power; Uranium Lead-Zinc 2010 symposium will provide the TMS Alliance of the Americas initia- Resources Ð Exploration and New Mines; an international forum for the lead and zinc tive to work together with Society partners Uranium Processing – Milling, Refi ning, processing industries to exchange informa- in South America and Canada. Technical Conversion and Enrichment; and Regulatory tion about current processing technologies symposia themes include: Characterization and Public Issues. The technical sessions for primary and secondary lead and zinc, and Application of Biomaterials; Dynamic will include papers on mining, uranium as well as emerging technologies for both Behavior of Materials; Composite Materi- processing, refi ning/conversion/fuel fab- metals. The symposium scope extends als; Light Weight Materials for Transporta- rication, reactor designs and decommis- from process fundamentals to operational tion: Processing and Properties; Materials sioning, radiation safety and advances, practices, and also includes the important and Society; Mechanical Properties of and various other topics. Contact: Brigitte aspect of environmental issues. At the Materials with Emphasis on Grain-size Farah, Metallurgical Society of CIM, 3400 operations level, comprehensive reviews of Effects; and Computational Modeling de Maisonneuve West, Suite 855, Montreal, the major applications of both metals will and Advanced Characterization. There Quebec, Canada H3Z 3B8; (514) 939-2710, be outlined. Emphasis will be placed on will be a CD-only proceedings published. ext. 1317; e-mail: [email protected]; www recent commercial developments with less Contact: Meeting Services, TMS, 184 .metsoc.org/u2010. energy-intensive technologies; the better Thorn Hill Road, Warrendale, PA 15086; understanding of existing technologies (724) 776-9000, ext. 243; e-mail: mtgserv 10th International Conference on and the development of new processing @tms.org; www.tms.org/meetings/specialty Steel Rolling—ICSR concepts; environmental concerns; a series /ABM-TMS/home.aspx. of plenary lectures by industry leaders. A September 15–17, 2010 short course on lead and zinc processing and Beijing, China 7th Pacifi c Rim International industrial tours are also planned. Contact: Conference on Advanced Materials The theme of the 10th ICSR is “Rolling the Brigitte Farah; Metallurgical Society of and Processing —PRICM 7 Future: Process, Products, and Environ- CIM; 3400 de Maisonneuve West, Suite 855, ment,” with the aim of showing the progress Montreal, Quebec, Canada H3Z 3B8; (514) August 1–5, 2010 and development in the fi eld of steel rolling 939-2710, ext. 1317; e-mail: metsoc@cim Cairns, Australia (technology, process, and products) in the .org; www.metsoc.org/com2010/tp_lead This international conference is the 7th past four years. The sessions will examine _zinc.asp. in a series devoted to advanced materials new process, technology, and facilities for fl at and processing. The conference, which is products, long products, and pipes; special 7th International Symposium on held every three years, is jointly sponsored rolling technology and products; equipment Alloy 718 & Derivatives by the Chinese Society for Metals (CSM), and maintenance; clean rolling; mathemati- October 10–13, 2010 The Japan Institute of Metals (JIM), The cal modeling, simulation, automatization Pittsburgh, Pennsylvania Korean Institute of Metals and Materials and intelligentization; advanced products (KIM), Materials Australia (MA), and The development and application; measure- This symposium will address recent inno- Minerals, Metals & Materials Society (TMS) ments and control technology for rolling and vations in material processing and under- and is organized in rotation. The purpose of products; etc. A plant tour of a rolling mill standing of superalloys, including alloy PRICM is to provide an attractive forum for in China is being organized as well. Con- 718 and related compositions. Specific the exchange of scientifi c and technological tact: Bin Gao, Chinese Society for Metals, topics will include: raw material devel- information on materials and processing. 46 Dongsixi Dajie, Beijing 100711, China; opments and trends; ingot metallurgy The symposia will cover such topics as +86-10-65133925; e-mail: [email protected]. and melt process advances; supply chain advanced steels and processing, advanced cn; hy.csm.org.cn/. expansion and new capability; casting high temperature structural materials, light technology and developments; fabrica- metals and alloys, bulk metallic glasses Lead-Zinc 2010 [in conjunction tion technology and developments; novel and nanomaterials, bulk metallic glasses with Conference of Metallurgists product forms technology and applications; and nanomaterials, advanced ceramics, (COM 2010)] fi eld applications and trends (energy, aero, energy generation harvesting and storage chemical, etc.); microstructure, properties, October 3–6, 2010 materials, and others. Contact: Ms. Helen and advanced characterization; and corro- Vancouver, BC, Canada Woodall, Materials Australia, Suite 205, sion, coatings, and environmental effects. 21 Bedford Street, North Melbourne, VIC Lead-Zinc 2010 is the fi fth decennial sym- Contact: TMS Meeting Services; 184 3111, Australia; +61-9326-7266; e-mail: posium by the Minerals, Metals & Materials Thorn Hill Road, Warrendale, PA 15086; [email protected]; www Society (TMS) that is devoted to the theory (724) 776-9000, ext. 243; e-mail: mtgserv .materialsaustralia.com.au/scripts/cgiip.exe and practice of the extractive metallurgy @tms.org; www.tms.org/Meetings/Specialty /WService=MA/ccms.r?PageID=19070. of Lead and Zinc. This meeting is held /superalloys2010/home.aspx.

100 www.tms.org/jom.html JOM • February 2010 MATERIALS RESOURCE CENTER: Consultants Directory ISO-9001 and ISO-17025 Certifi ed Analytical Services & Reference Standards SEM/X-ray, Electron Microprobe, Surface TRACE Analysis (Auger) Metallography, Particle ELEMENT ANALYSIS Size Counting & Surface Roughness Quantitive analysis of small amounts of material in concentrations to 10 ppm, surface oxides, stains & 5 Day Turnaround cleanliness, corrosion, diffusion, reverse engineer- ing, elemental mapping, diffusion gradients (carbon, • High Purity Metals & Alloys • Thick Films nitrogen, etc.) Get your data by e-mail. Digital • Ceramics • Carbon, Graphite micrographs and elemental maps. We manufacture • Glasses • High Temperature traceable standards for magnifi cation (SEM, OM, • Semiconductors Alloys etc.) and over 240 pure elements, alloys, glasses and compounds for micro-analysis. Utilizing State of the Art Techniques Put our years of experience to work on your specimens! Glow Discharge Mass Spectrometry (GDMS) Inductively Coupled Plasma Mass Spectrometry (ICPMS) Inductively Coupled Plasma Optical Emission Spectrometry (ICP-OES) 426e Boston St. (Rt. 1), Topsfi eld, MA 01983-1216 Spark Source Mass Spectrometry (SSMS) 978 887-7000 fax: 887-6671 www.gellermicro.com Combustion and Inert Gas Fusion Methods (LECO)

Northern Analytical Laboratory Worldwide experience in 13 Delta Drive Unit #4 technology development Londonderry, NH 03053 micron inc. FLOGEN Technologies______Tel: 603-434-8400 • Fax: 603-434-8500 ANALYTICAL SERVICES θ Non-Ferrous and Ferrous Metallurgy, Inorganic and E-Mail: [email protected] Organic Chemistry, Environment and InfoTechnology etc θ Website: www.northernanalytical.com Thermodynamic and kinetic quantifi cation of slags, 3815 LANCASTER PIKE mattes, metals, etc., smelting, converting and refi ning WILMINGTON, DE 19805 process development and optimization, modeling and A C T Aluminum Casting Technology, LLC simulation, Atomic Absorption Spectrometry, minor Professional Consulting Services; Project, Metallurgy, Field Service, 302-998-1184, FAX 302-998-1836 and trace element analysis, electroanalytical techniques Engineering of Casting Equipment & upstream/downstream Operations (biosensors), heavy metal pollution analysis, analytical quality assurance, mercury speciation, etc. Contact: JOHN V. GRIFFIN, P.E. [email protected] ______www.fl ogen.com______7 Lone Cedar Way Offi ce: (1) 201-767-9044 WWW.MICRONANALYTICAL.COM Old Tappan, New Jersey 07675 Fax : (1)201- 750-2279 5757 Decelles Ave., Suite 511 Tel: (514) 344-8786 United States E-mail : [email protected] Montreal, Quebec Fax: (514) 344-0361 Canada H3S 2C3 Email: [email protected] James E. Hoffmann and Associates Co. P.O. Box 420545 Houston, Texas USA 77242-0545 JAMES E. HOFFMANN, FIMM. P.E. CONSULTING EXTRACTIVE METALLURGICAL ENGINEER Your Professional Partner NON FERROUS, RARE, AND PRECIOUS METALS for Career Advancement BUS (281) 493-9441 E-MAIL [email protected] FAX (713) 780-0761 www.tms.org Read by thousands of the fi eld’s leading professionals and decision-makers, JOM can deliver an impressive international audience, both in print and online. The new TMS Media Kit is now available. To request your copy, contact [email protected] Check out the TMS media kit at http://www.tms.org/pubs/journals/JOM/ads.html

Vol. 62 No. 2 • JOM www.tms.org/jom.html 101 MATERIALS RESOURCE CENTER: Positions Available

Multiscale Modeling POSTDOCTORAL RESEARCH ASSOCIATE The University of North Texas IN NEUTRON SCATTERING The University of North Texas invites applications for a tenured Full Professor in multiscale modeling. This hire in the Materials Neutron Scattering Sciences Division Modeling Research Cluster (MMRC) is part of a multi-year, $25 Oak Ridge National Laboratory M investment by UNT in multi-disciplinary research collabora- Oak Ridge, Tennessee - ORNL10-34-NSSD tions. The successful candidate will have an established inter- national reputation with an active, externally funded research The Powder Diffraction Group the Neutron Scattering Sciences program. An earned doctorate in Materials Science, Engineering, Division (NSSD) of Oak Ridge National Laboratory (ORNL) Chemistry, Mechanical Engineering, Physics or a related fi eld (http://www.ornl.gov) has several opportunities for postdoctoral is also required. The area of specialization is broadly defi ned, research. Specific areas of interest include the following: but preference will be given to candidates who complement existing MMRC strengths (http://mmrc.unt.edu). Please visit: • Phase transformation and deformation behavior in http://facultyjobs.unt.edu/applicants/Central?quickFind=50513 high-strength steel for additional information, qualifi cations and to apply. Review of • Structure, thermal stability, and deformation of nano applicants will begin immediately and continue until the search structured ferritic alloys is closed. The University of North Texas is undergoing signifi cant • Texture and dynamic recrystallization of Mg alloys expansion in research personnel and capabilities: the new hire • In-situ neutron scattering measurement of transient will join an exceptionally productive and collaborative cluster phenomena of researchers Ð 13 faculty, >50 researchers Ð encompass- ing computational chemistry, material science, physics and Each candidate will be part of an interdisciplinary research engineering. UNT is a Class I Ð Doctorate Granting Institution team involving scientists in several divisions at ORNL, prima- located in Dallas-Fort Worth (DFW), 30 minutes from the DFW rily Materials Science and Technology Division and Computer International Airport. DFW is an area of more than six million Science and Mathematics Division. The research will utilize the VULCAN diffractometer at the Spallation Neutron Source people, with signifi cant economic growth, low cost of living, as a primary research tool, although there will be comple- numerous industrial establishments, and excellent school dis- mentary experiments involving synchrotron scattering and tricts. This area and the university provide excellent cultural and microscopy. In commissioning as of summer 2009, VULCAN educational opportunities as well as exceptional employment is a world-class engineering diffractometer designed to opportunities for spouses. UNT is the fourth largest university tackle a broad range of problems in materials science and in Texas with over 36,000 students. UNT is an EOE/ADA/AA engineering, including stress mapping in structural institution. components, in-situ deformation studies, transient behaviors during synthesis and processing, and the kinetics of multi- length scale phase transformations. Another TMS Member Benefi t:

QUALIFICATIONS: A Ph.D. in materials science, physics, Materials Technology@TMS mechanical engineering, or related fields is required. www.materialstechnology.tms.org/TECpage.asp Preference will be given to candidates with experience in neu- tron or synchrotron scattering techniques (e.g., diffraction or Your resource for small angle scattering). Strong written and oral communica- tions skills are desirable. The candidate must be willing to technology overviews, work in a team environment on technically and scientifically news stories, downloadable challenging problems. Applicants cannot have received the most recent degree more than five years prior to the date of resources, and links. application appointment and must complete all degree requirements before starting their appointment. FREE WEB PLACEMENT OF EVERY CONSULTANTS DIRECTORY AD JOM’s Consultants Directory ads reach experienced professionals who look to JOM for technical and professional guidance. It is also the fi rst place to which the staff directs readers with technical questions. HOW TO APPLY: Qualified applicants must apply online at https://www2.orau.gov/ORNL_POST/. Technical ques- For as long as the consultant's ad appears in the print tions regarding the position can be directed to Dr. Xun-Li version of the journal, the ad also appears in the on-line Wang at [email protected]. Please include the requisition version of the directory at no additional cost. If a web number and title when corresponding. This appointment is address is available, the ad will link to that site as well. offered through the ORNL Postgraduate Research Participation Program and is administered by the Oak Ridge To check it out, visit Institute for Science and Education (ORISE). http://www.tms.org/consultants.html

102 www.tms.org/jom.html JOM • February 2010 MATERIALS RESOURCE CENTER: Positions Available

I’VE SPECIALIZED FOR 30 YEARS

in the placement of Metallurgical, Materials, and Welding Engineers in the areas of R&D, Q.C. Production, Sales & Marketing, nationwide. My background as a Met. Eng. can help you! Salaries to $190K. Fees paid by Co. Call/Send/E-mail Resume: Michael Heineman, Meta-Find, Inc.; P.O. Box 610525, Bayside, NY, 11361; Phone (212) 867- 8100; E-mail mikeh@meta-fi ndny.com; Web: www.meta-fi ndny.com

If you have a key position to fi ll, JOM Classifi ed Advertising can help you track down that special individual with the right qualifi cations and background to match your needs.

One Source, One Partner for all of your Mineral Processing Needs www.FLSmidthMinerals.com

www.magnesium-elektron.com

Meetings and Workshops

FRAY INTERNATIONAL SYMPOSIUM December 4-7, 2011 * Cancun, Mexico Incorporating three international symposia: • Advanced Sustainable Iron and Steel Making • Sustainable Non-Ferrous Smelting in 21st Century • Molten Salts and Ionic Liquids 2011 Visit www.flogen.com/FraySymposium for additional details.

Vol. 62 No. 2 • JOM www.tms.org/jom.html 103 End Notes Materials science and engineering in our community Global Innovations Symposium Flies High at TMS 2010 Lynne Robinson

Editor’s Note: This is excerpted from an article Materials Applications Engineering tegration for Boeing. that fi rst appeared on MaterialsTechnology@TMS for GE Aviation and symposium orga- As a member of both the MPMD’s and is now archived in the Emerging Materials Community at http://materialstechnology.tms.org nizer, “The symposium has been split Shaping and Forming Committee and /EMT/home.aspx. into several sessions to highlight dif- the Structural Materials Division’s ferent trends across various manufac- (SMD) High Temperature Alloys Com- Floating on muslin wings and bound turing processes such as casting, form- mittee, Whitis said she was pleased to together with horsehide glue, the Wright ing/forging, and machining/joining, as have been able to integrate the interests Flyer launched powered air travel in well as alloy systems, including tita- and involvement of both groups in de- four shuddering fl ights of less than a nium, superalloys, and intermetallics. veloping the symposium. Exchanging minute each in 1903. Its fragile wooden One exciting trend to be examined will knowledge and ideas across disciplines frame, ultimately reduced to splinters be additive manufacturing, which has to provide a more comprehensive view by a heavy gust of wind, stood in stark recently seen a resurgence due to ad- of trends and advancements has been contrast to the powerful aircraft that vances in powder manufacturing meth- the cornerstone of the Global Innova- now routinely soar above the clouds. ods.” tions Symposium since it was intro- What these two eras of fl ight have very Computational materials engineering duced a decade ago as a new concept much in common, however, is a drive will also be highlighted, because, said in programming at the TMS Annual for discovery and innovation. This pur- Whitis, “An overarching challenge and Meeting. In addition to organizing their suit of new technology to enable faster, opportunity for our community is to own symposia, the MPMD’s technical safer, and more cost-effi cient fl ight will better incorporate modeling and simu- committees were asked to contrib- be explored and celebrated during the lation into metals processing in order to ute to a division “super-symposium” TMS 2010 Annual Meeting at the 11th develop new alloys and processes more that refl ected the breadth of materials Materials Processing & Manufacturing quickly and with fewer iterations.” processing and manufacturing within Division (MPMD) Global Innovations According to Whitis, this year’s sym- an emerging technology or industry. Symposium: Global Innovations in posium provides a balanced perspec- “It was our desire to engage as many Manufacturing of Aerospace Materials. tive with a a number of invited speakers production organizations as possible Chosen, in part, for the meeting’s from industry, academia, and govern- to strengthen the technical breadth of location in the aerospace center of ment. Kicking off the symposium on our membership—both within the divi- Seattle, the symposium’s topic will February 15 will be “Future Materials sion and within the society,” said John highlight key advances in the manu- and Process Needs for Commercial Smugeresky, technical staff member at facturing of aerospace materials across Jet Transports: the 21st Century Chal- Sandia National Labs and MPMD chair a wide spectrum of technologies. Said lenge,” a keynote presentation by Alan at the time that the symposium was es- Deborah D. Whitis, section manager, Miller, director of 787 Technology In- tablished. “The Global Innovations Sympo- sium has historically featured high pro- fi le speakers to provide a foundation for focused topics where new innova- tion has been occurring,” said Thomas R. Bieler, professor at Michigan State University and symposium co-organiz- Figure 1. Featuring state-of-the-art ad- er. “This year’s event is no exception vanced materials tech- and will provide participants the op- nology, a Boeing 787 portunity to gain a high-level view of Dreamliner completes engine runs necessary where the industry is heading in using for its fl ight test at Boe- new technologies to improve produc- ing’s facility in Everett, tion of aerospace materials and parts.” Washington. (Photo courtesy of Boeing.) Lynne Robinson is a news and feature writer for TMS.

104 www.tms.org/jom.html JOM • February 2010

ABB’s solutions help you increase your plant productivity and product quality Molten Metal Quality Measurement LiMCA CM Liquid Metal Cleanliness Analyzer ■ Available to all Aluminum Plants ■ Integrated to the Process and Plant Network ■ In-Line Continuous Monitoring ■ Automated Operation and Minimum Maintenance

Electromagnetic Stirring for State-of-the-Art Aluminum Treatment ABB, the world’s leading supplier of electromagnetic stirring systems for aluminum melting and holding furnaces of all types and sizes, announces that AL-EMS for ladles and crucible stirring is as equally reliable when it comes to temperature control and total melt homogenization. As with furnace stirring, investment for this powerful new application is normally recovered in less than one year.

Stop by the ABB booth #513.

ABB ABB Inc. Analytical Business Unit JME Division Quebec City, Qc, Canada Whitby, Ontario, Canada Phone: +1 418-877-2944 Phone: +1 905-666-2251 [email protected] [email protected] www.abb.com/analytical [email protected] www.abb.com/metals