Materials and Processes Research and the Information Highway: Summary Record of a Workshop (1996)

Discussion Group Members:

W. Wade Adams, Wright-Patterson Air Force Base

Jane Alexander, Advanced Research Projects Agency

James Davis, Xerox

Alan Hurd, Sandia National Laboratories

Ruth Pachter, Wright-Patterson Air Force Base

Alton Romig, Sandia National Laboratories

David Williams, Lehigh University

Williams presented a summary of the findings of the group. The dependence of materials and processes research on high-quality images is natural. Structure, especially microstructure as illuminated by micrographs, is the link to processing variations and materials properties. It is natural that researchers favor visual data because it allows rapid assimilation of vast amounts of data via the human visual cortex. No machinery can yet match the ability of the human eye to discern subtle variations in pattern, shading, coloring, or periodicity. In fact, visualization is favored so completely that researchers have a strong tendency to render even nonstructural data, such as composition and temperature, in image fields.

The desire for high-resolution images incompletely expresses the needs of materials researchers. Many other qualities in an image can be important, depending on its application, such as the dynamics of videoconferencing or security for proprietary images. In some niches of materials research it is a decisive advantage to add depth visualization, such as stereo microscopy in fractography and stereograms of protein structure calculations. Tactile and haptic (haptic meaning research dealing with touch perception and feedback, including the development of devices and software to effect haptic simulation) visualization feedback may be increasingly important as biomaterials research expands. Hence, the focus of Williams' presentation expanded to images of high quality, with the understanding that various qualities are considered in addition to resolution.

The imaging requirements of the materials community fall into three categories—publications, teleconferencing, and remote experimentation. Several types of publications were discussed, including archival journals, conference proceedings, image databases, and shared designs and images among small working groups.

The high cost of instrumentation argues strongly for centralization of major facilities. Electron microscopes, while improving rapidly in quality, have increased in price from one to five dollars per eV in the last 20 years, thus limiting the number of top-flight microscopes that the country will provide. Similarly, scattering facilities at accelerators and nuclear reactors necessarily serve large multinational regions due to their complexity and cost. Although not all such

facilities deal in real space images, most require visualization of data in some sort of manipulable images, including “run records ” from facilities that process and fabricate materials. Many small, relatively affordable facilities are likewise unique, and their frequency of use by a given customer is not great enough to justify individual researchers establishing their own facilities.

Williams argued that remote access to these facilities would make it possible to maximize the productivity of both human and machine resources. When a complex instrument breaks down, visiting users can be without work for days at a time. Scheduling remote users makes it possible to optimize the use of high-demand facilities. The ensuing cost savings from remote experimentation are expected to be increasingly important because future funding appears to be uncertain. Hence, high-quality online imaging capabilities are crucial to the feasibility of remote experimentation.

Williams addressed several issues that affect most, if not all, of the products and services which need to be addressed in order to make the use of the Internet for materials research more compelling. Data security (protection for proprietary and classified images) and validation and authentication of information (images) will continue to be important considerations. For both remote experiments and teleconferencing, bandwidth is a critical issue. Although current equipment is sufficient for present needs (e.g., T3 lines, which are high-speed data line connections which transfer data at a rate of 45 Mb/sec), it is not widely utilized by the materials community due to high cost or lack of awareness. Storage media formats and volatility also concern users. The average materials scientist struggles with everyday e-mail transmission of large image files, due in part to a lack of a standardized file compression protocol. In many cases it is more practical to mail a floppy disk using an overnight service than it is to attempt to e-mail a large file.

Printing (hard copy of high resolution images) and scanning technology is not yet compelling enough to convince more than a small fraction of the current generation of microscopists to convert from conventional silver halide film-based imaging to digital imaging. In the case of interactive white boards, better projection and image-capture equipment (more vivid, wider viewing angle) is needed. In virtual conferencing, a decisive factor is whether the participant will feel sufficiently engaged to participate in discussions with speakers and colleagues.

There are several advantages inherent in a centralized image database. Although indexes of sample type and processing conditions are clear-cut, materials researchers may not want to make searches based on those characteristics. Instead, they may want to search the images themselves for matching features. Some materials researchers are likely to be interested in correlating and tracking image features, such as defects, as a part of quality assessment (e.g., the microelectronics and aviation industries). The pedagogical possibilities of such a database are enormous. X-ray diffraction patterns and Nuclear Magnetic Resonance spectra offer similar opportunities.

The materials and processes community has the opportunity to emulate technology now being used in the entertainment sector, such as full-color dynamic

and static images. The panel agreed with Cargill that commercial enterprises will make increasing use of the Internet, as evidenced by such advertising-supported services as Yahoo and Netscape. Commercially driven improvements in the bandwidth, quality of service, multimedia capabilities, and storage capacity can be used now for research images and remote control operations. Williams stated that existing technology offers the opportunity to coordinate the remote-control activities of the three world-class electron microscope facilities now operating independently in the United States.

Williams asserted that the highest priority is to make it possible for the materials and processes research community to acquire what it needs from the information highway in spite of the fact that it is a relatively small group. Materials researchers and engineers must fully capitalize on the commercial drivers of the Internet— especially the image-intensive entertainment and health-care sectors —and take lessons from other scientific groups that have found it possible to obtain benefits from the information highway.

Finally, the inherent appeal of images should be used to engage the interest of others in materials research. Nowhere is this appeal more evident than in the enthusiasm of pre-college students given the opportunity to view objects as simple as a fruit fly under a scanning electron microscope. Eye-catching images can also be used to attract greater interest in materials research among government policy makers.

Williams concluded by saying that there are few major obstacles to use of the Internet to realize the basic imaging needs of materials and processes research. The biggest obstacle appears to be the materials and processes research community itself, which in general has been unwilling to embrace change.

LeClair asked about visual reasoning as a powerful way to analyze complex two- and three-dimensional patterns. He described it as seeking simpler patterns in higher dimensions that, when projected to lower dimensions, resemble the attending structure.

One attendee suggested the possibility of creating a “bulletin board” or “image consultation site” on the Internet where micrographs could be posted. Another participant said that a critical adjunct to an images database is a user-friendly program for simulating high-resolution electron microscope images. It was pointed out that the WWW site at Lausanne offers micrographs.

Manzione asked how the materials research community, which is small relative to such disciplines as electrical engineering and chemistry, could marshal support for funding research involving the application of the Internet to M|andsymbol|P research. He pointed out that although there are only about 1,000 astronomers in the United States, such endeavors as the billion-dollar Hubble Telescope have attracted broad political and financial support.

Knowles and Baskes noted that astronomy benefited from having NASA as its sole funding champion and that medical researchers were equally fortunate to have the National Institutes of Health (NIH). Materials funding comes roughly equally from three federal sources (NSF, DOE, and DOD), making it difficult to mount a coordinated effort. One attendee noted this lack of coordination with respect to High-Performance Computing, even within defense circles where both DOD and DOE have initiatives.

Discussion Group Members:

Frank Crossman Lockheed-Martin

Lance Davis, U.S. Department of Defense

Larry Davis, User Technology Associates

Joseph Friday, Alcoa

Nirav Kapadia, Purdue University

Larry Kerschberg, George Mason University

Steve LeClair Wright-Patterson Air Force Base

David Maluf, Stanford University

John Mitchiner, Sandia National Laboratoaries

Mark Munson, Minerals, Metals, and Materials Society

Leslie Schneider, Tufts University

In his presentation of this group's findings, LeClair began by describing the compelling advantages that the Internet offers to materials and processes researchers. They included:

• Competitiveness. The Internet's ability to give researchers rapid access to information and rapid communications can improve their productivity. However, the specific Internet application for this purpose still needs to be defined.

• Handbook Information. The Internet can provide rapid access to handbook data, and the information can be updated much more easily than with a printed handbook.

• Access to Advice. The Internet facilitates finding experts who can provide needed information.

• Collaboration. With the technology provided through the Internet, “virtual teams” composed of researchers working at widely separated sites can be formed to carry out research projects that might not be feasible otherwise.

• Virtual Design. The Internet could give materials researchers better access to information and computational resources that would facilitate the work of making virtual designs of new materials.

• Quick Decisions. By speeding up interchanges of information between distant parts of an organization as well as with vendors, the Internet could enable materials developers to make more rapid decisions on what directions to take in development.

• Remote Access to Research Tools. The Internet can give researchers the ability to use research tools that are not available at their own facilities and thus avoid the expenditures that would be necessary to create an in-house capability.

LeClair described the barriers found by the discussion group that inhibit materials researchers from taking advantage of the Internet 's capabilities.

• Data and Quality. Only a limited mount of materials and processes research data can be accessed through the Internet, and it is often difficult or impossible to assess the data that is accessible.

• No Standard Taxonomy. The absence of a consistent method of classification of materials properties often makes it difficult or impossible to determine the credibility of data on the Internet and to merge and analyze information from different sources.

• Computer Use. The materials and processes community is generally not inclined to take full advantage of online resources. Training will be required to promote widespread use. This is not true for the majority of recent graduates, however.

• No Commercial Market. No commercial firms have attempted to set up a commercial market that would provide Internet users with information on materials and the opportunity to procure them.

• Cost Recovery. Because of the multiple variables involved, materials databases are quite extensive. No clear strategy has evolved on how those who spend the time and effort necessary to put such information on the Internet can recover the costs they incur. Transaction charges, while easy to implement, have a dampening effect on use of an information source.

• Web Junkies. Ineffective use of the Web (e.g., indiscriminate searching), can distract researchers from taking practical advantage of the Internet

Future directions defined by the discussion group as pertinent to Internet use by the materials and processes community are:

• virtual materials design and control

• development of taxonomy for materials and processes databases

• continuous learning methods over the Internet

• retraining researchers

• establishing an environment for brokering and sharing tools

• methods of conceptualizing data

In closing, LeClair stated that is difficult for materials researchers to recognize dramatic benefits from using the Internet. Adoption of widespread use by the materials and processes community will involve showing explicitly how its use can provide specific advantages.

A participant asked a general question about sharing research instruments and knowledge that prompted LeClair to recall that Russian researchers had amassed a database on the molecular structure of pharmaceutical compounds.

Hurd asked how continuous learning could be implemented. Crossman said that one method is to share best practices. LeClair said that working for a long time as a manager had left him out of touch with new technology, and that education should go beyond the practicing engineer to include technical management. Baglin said Constant's presentation showed that the Internet could be used to teach classes at many locations and get widespread participation. Sonwalkar commented that there are no formal retraining requirements for computer scientists despite rapid advances in the discipline, whereas medical doctors are required to periodically attend classes to update their skills.

Mahin said that the biggest obstacle to use of the Internet by materials researchers is that they are not adept at accessing information on the Internet or making their research results available on it.

Discussion Group Members:

Michael Baskes, Sandia National Laboratories

James Callan, University of Massachusetts

James Davis, Xerox

Robert Dutton, Stanford University

Calton Pu, Oregon Graduate Institute

Thomas Rohmiller, Wright-Patterson Air Force Base

Michael Schwartz, @Home Network

Nishikant Sonwalkar, Massachusetts Institute of Technology

Gio Wiederhold, Stanford University

Baskes presented this group's findings. He began by discussing the limitations on materials information that impede the work of materials end-users such as integrated circuit designers. Typically, he said, the product designer pre-specifies one or a few materials that seem like good possibilities, and no consideration is given to other kinds of materials unless major problems occur.

This approach, she said, is hardly conducive to exploiting the capabilities of modern materials. If designers are to make the best use of their talents, they need

“machine-friendly” access to the information now found in handbooks and the background knowledge needed to understand handbook tables and the effects of changes in design parameters which are not shown by the tables. Often, designers are limited by incomplete materials property tables which do not have values under the conditions needed. Interpolation and extrapolation of existing data may be a way to complete these sparse data tables.

The group identified three types of information that could usually be obtained through the Internet: identification of materials suppliers, identification of candidate materials for products, and assessments of the quality of the data. Baskes stated that although the Internet in general is overloaded with information (i.e., much more information is available than the user can process in a reasonable time), the discussion group concluded that this is not yet true for materials and processes information. Nonetheless, a future overload of data on materials is a real possibility. Thus, as more and more information becomes available online for M&P researchers, there will be an increasing need to assess the value of the data and discard what is unimportant. Thus, there will be a need for intermediaries to perform this function. There will also be more data online which has not been verified because of the existence of easy tools for self-publishing, which bypasses standard publishers, as well as reviewers. Another factor that may contribute to overload is the ever-increasing number of links at WWW sites. Automatic search engines indiscriminately collect information by way of these links and thus provide mountains of unneeded information.

Baskes then discussed the possibility of creating intermediate workstations on the Internet that would be owned by domain knowledge experts and that would function as agents to convert, validate, integrate, and abstract source data. Potential mediators include information brokers who would be employed by commercial firms that would be supported by fees, nonprofit professional societies, and university scientists willing to provide mediation without pay because it enhances their professional stature.

In the meantime, however, there is little evidence that M|andsymbol|P researchers are switching from paper to electronic sources of information. They have yet to see the potential benefits. The professional societies continue to publish the books, journals, and newsletters, that have traditionally been one of their main sources of income. The American Vacuum Society, which is small and makes considerable use of volunteers, has been more aggressive than any of the other materials societies in moving to electronic publication. The other societies sense the conservative attitude of their members to the electronic media and therefore have limited their investments in electronic media. Representatives of professional societies who were in attendance at the workshop said that an aggressive approach to electronic publishing would require a large investment and reduce revenues from printed publications, leading to pressures to increase membership fees. They said they doubted that their members would look favorably on increases in annual dues. The societies also have doubts as to whether they would be able to collect sufficient revenues from electronic publication to make up for revenue lost from

sales of printed publications because of the precedent of free use of information sources on the WWW.

Although the design community would benefit greatly from having this data available, the designers have little voice in the M|andsymbol|P research community. In the meantime, handbooks are felt to be adequate, particularly in light of the fact that the technology needed for multi-parametric retrieval of materials and processes data is not yet available in the electronic medium. A further barrier to sharing is concern about proprietary information. Materials properties obtained by testing for specific applications at end-user sites are very valuable but are usually not made available because they are considered proprietary. Although the properties of the materials offered by suppliers are not proprietary information, the processes employed to make the materials often are. Few suppliers have made their materials catalogs available on the Internet.

The discussion group felt that the underload of online M|andsymbol|P data could be remedied by using automatic interpolation and extrapolation techniques to fill in and extend the databases. These techniques can be built into retrieval mechanisms.

Baskes's final comments pertained to security considerations. The materials community has an interest in how authentication, cryptography, publication security (read only), and firewalls (to prevent unauthorized access to information) technology can best be used to enhance Internet application. These concerns, however, are not peculiar to materials.

Losleben stated that Manzione gave a talk earlier in the workshop on eight-month product development cycles indicating the need for rapid access to new materials data and materials development. The M|andsymbol|P research community needs to address this need.

Baskes suggested that the inability to meet this need is partially due to complacency by materials researchers in making early developmental results available.

Mahin added that another problem in making materials data available is that great value comes from important experiments conducted by end-users who consider the results proprietary. The tests are usually extensive and costly because they involve design qualification.

Baglin commented that the Materials Research Society (MRS) was not reluctant to publish electronically because of the negative impact it might have on MRS's paper publications. In fact, Baglin said, MRS made no profits from publishing its Journal of Materials Research and its Bulletin. Baglin went on to say that the society had organized a forum in San Francisco to inform its membership about electronic alternatives to print publishing and to sound out member opinion. Given that MRS is a small society, Baglin said, it had to be conservative with its financial resources. One big loss arising from electronic publishing could cause the society to go out of business.

Baskes then asked about the prospects for collaboration among the societies in electronic publishing. Baglin said that at present it was minimal.

Knowles said that electronic publishing was expensive and often failed to turn a profit. “People who got in with CD-ROM, perhaps half lost money.” Knowles said that although ASM's paper products were selling well, it had also created a WWW site containing a directory of producers. He added that while ASM realizes the future importance of electronic publishing, it wants to make progress gradually, “without losing our shirts.” Income from print publications, he noted, helped to subsidize ASM's membership.

Goldie said that most of his customers lose money, and business is difficult. Many societies offer both CD-ROM and traditional paper products containing the same information, and the electronic products may therefore be considered redundant and not needed.

Rohmiller said that he buys both paper products and CD-ROMs. However, upon further discussion with Goldie it was clarified that he generally does not buy journals on CD-ROM.

Bederson said that the American Physical Society (APS) had agonized over electronic activities for many years, since funding for them came directly from the society's treasury. The society, he said, had demonstrated a good version of Physics Review Letters online. The APS is also working with another society to develop an electronic version of one of APS's journals. “We have responsibility to our 40,000 members,” Bederson said. They don't want to spend all the cash reserves on putting publications online if there is the possibility of a large financial loss. Although APS, ASM, MRS, and AVS all see electronic publishing as a future option, all of them except for AVS think that putting journals online now is too expensive.

Weaver noted that the American Vacuum Society put its journal online with minimal financial risk by taking a streamlined approach which resulted in a low cost for the project.

Knowles presented the findings of this discussion group. He began by talking about inadequacies in the exchange of information between materials and processes researchers and end-users. As a result, new materials for which there is a strong need are not being developed. Further, new materials that are being

discovered by researchers are not finding their way into products because end-users are unaware of them. This is detrimental to both researchers and end-users, and is a result of the lack of exchanges of requirements and capabilities.

Knowles stated that the group felt that there were very strong drivers for promoting communications through the Internet. The length of the product development cycle is decreasing, thus providing a further impetus for extensive and rapid exchanges of information. There is an increasing need for the development of new high-performance materials which can provide end-users with a product performance advantage. Costs for developing new materials are high, so knowledge of market need is required before making substantive investments. The globalization of both the product and materials markets has put great pressure on industry to remain competitive.

The challenges which must be addressed to meet these needs were summarized as follows:

• Persuade industrial competitors to collaborate on posting their materials needs even though they may consider the information proprietary and likely to give competitors a clue to future plans.

• Assure confidentiality of company proprietary data.

• Correct the lack of intelligent user interfaces that makes information exchange difficult because irrelevant information is not filtered out.

• Acknowledge that the materials community is too small to drive Internet tool development for this application and will have to depend on commercially available solutions.

• Find ways to encourage the materials community to make greater use of the Internet and online services.

• Improve information pedigrees.

• Work toward standardizing materials information.

• Answer questions about protecting intellectual property rights.

Knowles then described the advantageous functions that an Internet-based system could provide:

• Linking technical, marketing, financial, and manufacturing personnel together in concurrent engineering environments.

• Providing for mutual exchange of technology roadmaps between end-users and materials researchers.

• Establishing a needs catalog and a materials solutions catalog that would be continually updated and that would provide pedigrees for the information.

• Providing a service to match items in the needs catalog with entries in the solutions catalog.

• Assuring anonymity in information exchange if desired.

• Making possible the formation of virtual teams.

• Providing the ability to send information to a multitude of users (multicast).

Knowles concluded with a description of the potential solutions defined by the discussion group:

• Develop search engines that are driven jointly by functional requirements of the end-users and materials and process solutions.

• Create technology roadmapping sites where end-users can collaborate on conveying the key trends in their industry to the materials research community and other interested communities (general management, investment, marketing). Similarly, the materials research community should collaborate on the roadmap on materials innovations of interest to the end-user community.

• Use the internet to develop virtual partnerships or teams that are drawn together to work on a particular problem. These partnerships would include collaborations among end-users to identify key needs, collaborations among materials suppliers (including smaller companies) to share the risks of developing solutions, and collaborations between end-users and suppliers to drive these solutions to value in products.

• Anonymous brokers, such as the professional societies, could moderate these interactions and supply quality and pedigree of the information.

• Develop “Materials Problem Solutions” discussion sites which would function as “chat rooms” where technical needs and solutions could be exchanged.

• Develop libraries for new materials, materials processes, and end-users needs.

Crossman asked if trade journals would be able to meet the growing need for rapid interchange of information. Knowles and Manzione responded that the issues of timeliness and specific needs are what can make the Internet a superior solution to this issue. Trade journals are general in nature and do not offer much information that is new to the experts.

Mahin said thet there is one issue we didn't discuss. If databases are put online, how do you let users become quickly aware of new options and new links, a broader perception of what people need. Baglin asked what incentives are provided for people to do that. Mahin said that a monetary incentive could be set up with an information link that requires a fee. Knowles noted if the market requires this information, users would be willing to pay. Baglin stated that Researchers could provide this service gratis, but would want some form of recognition for their service. Knowles added that technical societies perhaps could take this role.

Discussion Group Members:

John Baglin, IBM

Ben Bederson, American Physical Society

Peter Goldie, Lightbinders, Inc.

Paul Losleben, Stanford University

Mark Lundstrom, Purdue University

John Rumble, National Institute of Standards and Technology

John Smugeresky, Sandia National Laboratories

John Weaver, University of Minnesota

Jerry Woodall, Purdue University

Baglin presented the group's findings. The first was that the materials and processing (M&P) research community has a wide spectrum of validation and authentication (V&A) needs. There are many different types of information, including peer-reviewed articles, textbooks, handbooks, and conference and meeting proceedings. In addition, the M&P community utilizes a wide variety of data-lists and tables. These lists and tables also have many classes or categories which present special problems for V& A, depending on whether the data are raw, commercially generated, reviewed, critically evaluated, or certified. This great variety of information, which is unique to the M&P community, has come about for several reasons. First, there are strong links between basic R&D and industry as a whole, with different industries having different requirements. Second, R&D within the M&P community is highly interdisciplinary. Finally, the need for reliable and validated information versus “cutting edge” breakthrough data varies from one application to another. Commercial and military aircraft designers, for example, have a much stronger need for validated materials data than for “cutting edge” information. Conversely, researchers investigating compound semiconductors, for instance, need rapid access to breakthrough data and are less interested in acquiring validated data.

Special challenges emerge for the V&A process for M&P information on the Internet, and the group provided several suggestions for dealing with these challenges. First, it was felt that a “roadmap ” for electronic publishing technology was needed to set standards and goals to compel publishers, including professional and technical societies, to use the information highway comprehensively, efficiently, and economically. The current peer review process is fundamentally sound, but the Internet offers improved mechanisms and options for effective V&A. The Internet, for example, offers an opportunity to improve the accuracy and scope of data pedigrees, especially if tags are used to provide more effective search criteria. Unfortunately, the capabilities of the information highway also create an environment which may allow for abuse, including the rapid and widespread dissemination of unsubstantiated data, an issue which also needs to be addressed.

There is also a need for a second “roadmap” that would address the development of standardized V|andsymbol|A procedures and thus complement the electronic publishing technology roadmap.

Of the many potential applications of the information highway for M|andsymbol|P research, electronic publishing seems to offer several important advantages, such as appending almost unlimited amounts of supplemental information to a research paper published electronically, hyperlinks to reference documents and other sources, online citation indices, interactive reader participation, and the identification of previously overlooked but relevant research.

To implement some of the above suggestions, there are several issues which need special attention. One is the issue of attribution of authorship for varying levels of publication. Another is the problem of defining in legal terms what constitutes a “publication.” Finally, perhaps the most important issue is determining how to handle intellectual property issues, such as copyright, patents, etc. These issues must be resolved before the M|andsymbol|P community can make significant use of the information highway.

It was important, Goldie said, that access be limited to reviewers for online documents reviewed prior to publication. Baskes asked if there should be a specific electronic publishing roadmap for materials science. Baglin responded that although some issues are specific to materials, electronic publishing is a broader issue than materials publishing, and therefore an electronic publishing roadmap should reflect that. Collaboration between materials scientists and publishers could foster the development of an electronic publishing roadmap that encompassed the needs of the materials community.