We present the results from a workshop on interdisciplinary research on design of engineering material systems, sponsored by the National Science Foundation. The workshop was prompted by the need to foster a culture of interdisciplinary collaboration between the engineering design and materials communities. The workshop addressed the following: (i) conceptual barriers between materials and engineering design research communities; (ii) research questions that the interdisciplinary field of materials design should focus on; (iii) processes and metrics to be used to validate research activities and outcomes on materials design; and (iv) strategies to sustain and grow the interdisciplinary field. This contribution presents a summary of the state of the field—elicited through extensive guided discussions between representatives of both communities—and a snapshot of research activities that have emerged since the workshop. Based on the increasing level of sophistication of interdisciplinary research programs on design of materials it is apparent that the field is growing and has great potential to play a key role in a vibrant interdisciplinary materials innovation ecosystem. Sustaining such efforts will contribute significantly to the advancement of technologies that will impact many industries and will enhance society-wide health, security, and economic well-being.

## Introduction

### Emerging Importance of Materials Design.

Recent advances across several fields have afforded the opportunity to create advanced engineering material systems designed to have properties and functionality needed for specific applications. Researchers have articulated a vision for computational materials design over the past two decades (e.g., see Refs. [16]). However, it is only through recent improvements in several key areas—such as computational modeling of materials and structures, materials processing, advanced manufacturing, informatics, computing power, and computational design techniques—that it now is possible to realize this concept practically across a wide variety of engineering material systems.

The potential impact of designing materials specifically for their engineering applications is significant. In effect, materials are the ultimate enablers of technologies that impact our health, economic well-being, environment, and national security [7]. An acceleration of the materials discovery, development, and deployment cycle has potentially dramatic economic impact. A recent report sponsored by NIST [8] presents estimates of the potential economic benefit of an improved Materials Innovation Infrastructure of between $123 billion and$270 billion per year. Clearly, it is necessary to develop the culture, frameworks, and expertise ecosystem necessary to bring this enormous potential to fruition.

The traditional paradigm in engineering design is based on the selection of predetermined materials from a fixed set of alternatives, which means that the design activity would go no further than specification of already existing materials and their configuration necessary to perform the intended function. Although this paradigm always will have a place in engineering, many applications—particularly those under tight performance or economic requirements and those operating under extreme conditions—can be improved significantly by continuing the design activity into the material structure and processing levels. This is the case because, often times, existing materials cannot fulfill the functionality, performance, or economic requirements to enable key technologies. Materials design, thus, provides more freedom to the engineering designer as the materials themselves and their configurations become entities that are subject to design in order to satisfy (potentially system-level) functional and performance objectives [4].

### Materials Design Versus Materials Discovery.

Engineering design is a process by which products, processes, systems, services, and materials are created to achieve specific functional and performance objectives. Superficially, it is the inverse of engineering analysis—whereas engineering analysis seeks to determine the properties and performance of a specific design alternative, engineering design seeks to identify a design alternative that meets specific objectives. However, it is better viewed as a decision-making process involving the formulation of design objectives and search for alternatives that best meets them. In principle, one can formulate and execute this search process in a multitude of ways, and choices made at this level have a profound impact on design outcomes.

For several decades, the design research community has considered questions of how to structure and execute this search process for various types of problems and contexts. A key takeaway is that design methods are most successful when general design principles are specialized to a particular context of interest. For example, a method that works particularly well for designing the layout of a wind farm will not necessarily be appropriate for designing the blades of a wind turbine. Progress in design research requires perspectives on both general design principles and the engineering context, thereby placing a premium on interdisciplinary research.

One can view the field of materials design as a continuously evolving collection of design concepts and frameworks specialized to specific materials and material systems to support the search for hierarchical (multiscale) material configurations that best meet application-based functional and performance requirements [4]. This is in contrast to the endeavor of materials discovery, which refers mainly to the process of synthesizing, identifying, characterizing, and modeling different materials formulations (and processing/synthesis routes) in order to develop a better understanding of the materials performance space. In a sense, the key difference between materials discovery and materials design is the shift from the outcome of the exploration of the materials space to the goal-driven process through which such space is explored in order to identify materials solutions that best meet specific functional, performance, and economic objectives.

Major initiatives in materials discovery over the past two decades have been instrumental in providing a foundation for materials design. The major strategic goal of the materials genome initiative (MGI) is to accelerate the materials discovery and development cycle [7]. Prior to the MGI, the integrated computational materials engineering (ICME) paradigm emerged as a way to accelerate the development and insertion/deployment of materials into technology through the integration of experiments and computations across the entire materials workflow, from synthesis/processing to manufacturing and integration [9]. Although both MGI and ICME (and other related programs) state as their main goal the faster design and deployment of superior technical artifacts and systems, they do not address directly the process by which such a goal would be achieved. Despite this, the net effect is that MGI, ICME, and related activities have significantly advanced knowledge in areas critical to supporting engineering materials design, but much research remains to be done in realizing a comprehensive vision of materials design.

### A Workshop on Interdisciplinary Frontiers of Designing Engineering Material Systems.

To date, there have been two main barriers to widespread progress on materials design research. One has been the level of maturity of core knowledge and technology areas needed to support materials design activities. For example, until recently there were still major scientific and technological challenges associated with establishing (computational/experimental) linkages across the process–structure–property/performance paradigm [1] that underpins much of the research activities of the materials science community. Despite the many remaining challenges, this barrier has reduced considerably over the last decade through research programs and initiatives such as ICME [10,11] and MGI [1214]. The design research community also has made great strides in areas such as design representation, design informatics, and design automation. A second barrier has been the interdisciplinary nature of the problem. Whereas many knowledge areas associated with MGI and ICME can be advanced from a (materials-centric) single-discipline perspective, the nature of design methods requires an understanding of general principles of design frameworks and the specific material and engineering contexts.

An opportunity exists to accelerate research on materials design by furthering the development of a coherent interdisciplinary community dedicated to understanding materials design problems. With a main goal of fostering such a community, a 1.5-day workshop was held at Texas A&M University in July 2016. This workshop, supported by the National Science Foundation,2 brought together over 50 representatives from the materials and design communities across academia, government, and industry with the common goal of developing and articulating a vision for an interdisciplinary field of research focused on the engineering design of material systems.

The remainder of this article is a summary of workshop findings and review of select literature in the area of designing engineering material systems. Section 2 is a description of the workshop, including its goals and overall methodology. Section 3 includes the main findings and recommendations from the workshop, with an emphasis on steps to foster a vibrant interdisciplinary field of research on designing engineering material systems.

## Workshop Goals and Methodology

### Developing a Vision for Interdisciplinary Design of Engineering Material Systems.

The premise of the workshop was framed around the importance of interdisciplinarity in the context of design of materials systems. An interdisciplinary field is one that combines and adapts approaches from two or more fields, so they are better suited for a problem of interest [15]. This is in contrast to multidisciplinary research in which investigators pursue a single problem using different disciplinary perspectives, without much cross-communication among the disciplines involved. In fact, to date, the most successful and impactful prior research on designing engineering material systems has been from an interdisciplinary perspective [3,4,16,17]. It is recognized that such interdisciplinary efforts in which engineering design experts communicate and collaborate with materials experts requires the development of design frameworks specifically tailored to the materials design problem, rather than the use of design methods that have been developed without considering the complexities associated with the multiscale nature of materials behavior. An opportunity exists to accelerate research in this area by furthering the development of a coherent interdisciplinary community dedicated to understanding materials design problems.

To address the challenges associated with the creation of a coherent interdisciplinary field of design of engineering material systems, the workshop revolved around addressing four major questions:

1. (1)

What is the conceptual scope of an interdisciplinary field of research on the engineering design of materials and materials systems? (i.e., what is in the interdisciplinary field and what remains with the disciplines?)

2. (2)

What are some of the key open research questions in the context of designing engineering materials and materials systems? How should materials and design experts work together to address them?

3. (3)

How should researchers in the new interdisciplinary field evaluate and validate their claims? What are the types of evidence we expect to see in good research in this area?

4. (4)

How can we build a sustainable interdisciplinary research community focused on the design of engineering materials and materials systems when starting from separate disciplines?

### Workshop Format and Mechanics.

The workshop format was designed to allow maximum input and to bring new insights into the scope of research challenges associated with the design of engineering material systems. Design and materials researchers were invited from academia, government, and industry to share their experiences and views on new directions and challenges in materials design. The emphasis of discussions was on interdisciplinary activities and challenges that neither community is likely to solve alone.

Workshop participants were asked to complete a preworkshop survey so that organizers could assess participant background knowledge and interests, including any prevalent misconceptions. For example, participants were asked to define several core concepts related to workshop goals (materials science and engineering, engineering design, interdisciplinarity, etc.). Other questions pertained to the participants' research expertise and what expertise they were missing personally but might be contributed through a suitable collaboration. They also were asked whether they had any prior experience in interdisciplinary collaborations. This information informed discussions and was used to organize breakout groups (e.g., to ensure a diversity of expertise in each group).

The workshop included a mixture of plenary and breakout sessions. Survey results were used to ensure proper distribution of expertise and interests in the moderated breakout sessions (participants were subdivided in groups of approximately 6–8). Each breakout group was constructed with a mix of expertise based on these responses. Breakout sessions focused on issues critical to the development of the interdisciplinary field of Design of Engineering Material Systems. The workshop began by first trying to bridge the cultural/conceptual gaps between fields. Further development of the (interdisciplinary) field requires the definition of central/foundational concepts. The discussion was then directed toward the identification of challenges and opportunities associated with the design of materials systems. Properly addressing such challenges requires the understanding of the similarities and differences associated with the types of evidence and evidence-gathering techniques considered as appropriate in engineering design and materials. Finally, it was recognized that an interdisciplinary field cannot grow and become self-sustaining without the proper avenues for creating and propagating the relevant interdisciplinary knowledge. Survey responses and notes from each breakout group were analyzed by organizers and are the basis for findings reported in Sec. 3.

## Workshop Results: Establishing an Interdisciplinary Field of Materials Design

### Preworkshop Survey Responses.

Workshop participants represented a spectrum of design and materials research expertise. The preworkshop survey revealed that a majority of participants had expertise that one could characterize as primarily related to either design or materials research. However, expertise was very distributed within each of these two broad categories. Those with primarily a materials research background included experts on computational materials modeling (ranging from atomistic to structural mechanics), synthesis and characterization. This expertise spanned all commonly investigated materials classes, including metals, ceramics, polymers, architected materials, and composite structures. Those with primarily a design research background included experts on design theory and methodology, design automation, design for manufacturing, bio-inspired design, and topology optimization. Approximately 20% of participants reported being already engaged in interdisciplinary research on topics such as materials informatics or materials design, though many were in the early stages of such work.

Participants also were asked about the gaps in their expertise in terms of what types of collaborations they would need to advance the field of materials design. The purpose was to gain insight into the collaborative potential of the community. Figure 1 is a visualization of the results of matching stated expertise and knowledge gaps as reported by participants. Blue nodes indicate a research asset; yellow nodes indicate a collaborative opportunity. The complete labels of each node can be accessed.3 The larger nodes correspond to key terms that appeared frequently; specifically, materials, materials science, design methods, and topology optimization were identified as highly needed or available within the participants. A minority of individuals did not appear to have a strong collaborative match with others in attendance, but the majority of participants had complementary expertise matches with others. This is captured in network 16, in which there are multiple examples of a need of one participant being a capability of another. This analysis has limitations, such as being confined to workshop participants and relying on text entry, which can lead to responses at vastly different levels of abstraction (e.g., atomistic simulation versus materials modeling). However, the snapshot it provides is useful that many potential expertise matches were identified which suggests that community-building activities can lead to greater collaboration and overall impact.

In response to a survey query to define materials science and engineering, most survey respondents provided similar answers. Many identified the significance of process–structure–property relationships to materials science and engineering. Discovery, characterization, and modeling were three commonly cited processes involved in materials science and engineering. A few respondents distinguished between materials science and materials engineering, noting that materials science was more about basic understanding and materials engineering was more about the translation of basic knowledge into engineering practice.

Definitions offered for engineering design were more varied, but largely consistent with one another. Key themes were that engineering design is an iterative process intended to result in an engineered product in response to desired functionality, performance, and economic objectives. Several respondents noted that product should be construed broadly to include complex systems, individual components, processes, and materials. Other respondents identified the importance of customers or stakeholders in informing design objectives and that engineering design can be viewed as a decision-making process. A few respondents indicated a less-complete understanding of engineering design. For example, some explicitly equated engineering design with optimization. It is the general consensus that although computational optimization is a commonly used tool in engineering design, it is neither necessary nor sufficient to solve design problems—there is a wide range of design problems that cannot be framed as an optimization problem. This points to a need for increased communication and education about the field of engineering design.

### Conceptual Scope of the Interdisciplinary Field.

To be recognizable as a field of study, members of any (interdisciplinary) field must utilize a core set of concepts by which they describe their work. For an emerging interdisciplinary field, these concepts initially are drawn from several fields and then evolved in a manner most useful to the new field. Other concepts remain rooted in their original fields even though they may be used within the interdisciplinary field. Working in breakout teams, workshop participants were asked to identify concepts and knowledge that are essential to an interdisciplinary field of research on designing engineering material systems. This process is not expected to yield a fully comprehensive set of concepts but can be counted on to provide an important and substantial starting point.

Workshop participants identified the following as core concepts essential to materials design.

#### Basic Principles of Design Methodology.

This includes the major steps of a comprehensive design process (e.g., problem clarification, conceptual design, embodiment design, detail design), design problem formulation (including the role of customers and stakeholders, functional modeling, etc.), design exploration, and the need for iteration.

#### Optimization in Context.

It was viewed as highly important to have a working knowledge of optimization, including various problem formulations and solution techniques. However, it was not viewed as necessary for everyone to be deep experts in optimization algorithms. It was considered important that community members understand the proper role and application of optimization in a design process. Several participants emphasized the importance of not equating optimization and design.

#### Process-Structure-Property-Performance Relationships.

The process–structure–property–performance (PSPP) relationships were identified as a key abstraction for discussing materials design. Specific expertise in modeling the relationships can remain a disciplinary consideration, but anyone working in the materials design community should understand the PSPP relationships in concept.

#### Uncertainty.

Participants widely agreed that dealing with uncertainty is essential in materials design. Materials data, PSPP models, and material processing steps all involve uncertainties. To be an expert at materials design is to be proficient at understanding, modeling, reasoning about, and making decisions under these uncertainties. Basic concepts such as probability theory and risk are important as are techniques for uncertainty quantification and propagation.

#### Validation.

This is a subtle but critically important subject in the context of design. In materials science, particularly in the creation and advancement of PSPP models, there is a quest for accuracy of models. Moreover, researchers generally seek models that best reflect the empirical evidence. However, a highly accurate model can be a waste of resources in design. Instead, the standard is to use a model that is sufficiently accurate to make design decisions in the specific engineering context of the decision.

#### Informatics.

Data-driven techniques are increasingly common in both the design and materials communities. A working knowledge of informatics concepts, such as data cleansing and preparation, estimation, and machine learning, was considered important.

#### Types of Engineering Materials and Material Systems.

Although it almost goes without saying, some breakout groups made a point to emphasize the importance of having a working knowledge of various types of engineering materials and material systems. This includes broad material categories—metals, ceramics, polymers—and material systems—architected materials, composite structures, functional materials—of importance in engineering applications. An individual surely would have deep knowledge about a material they study, but it also is important to have a broader understanding of the material landscape. Importantly, this helps promote cross-pollination of research and overall community building.

### Guidelines for Evaluation of Research Contributions.

Critical to the operation of any scientific field is agreement on how research claims are to be evaluated. This discussion is important in the context of interdisciplinary research because different disciplines may have different norms for evidence-gathering and validating studies. In fact, workshop participants noted both similarities and some key differences between how materials-focused and design-focused research is conducted and what is considered best practices.

Design and materials research both are scientific endeavors and have considerable internal variation in terms of the types of claims made and the means by which they are validated. Broadly, materials research investigates hypotheses pertaining to materials, models of material, and processes by which materials are realized. Materials researchers typically can validate their hypotheses against empirical evidence gained through designed and reproducible experiments conducted in laboratory settings. In contrast, design broadly asks questions about processes and methods. These are logical constructs and research on them does not always fit into a paradigm of controlled laboratory experiments. Perhaps counter-intuitively, the fact that using a particular design method results in a good design is not sufficient to conclude that the method is good. To take an extreme example, a random number generator configured as a design method will occasionally produce a good design. Thus, logical consistency often is one of the first lines of evidence in design research. Other evidence, such as evaluation of method performance metrics or practical usefulness, also is important and can be gained through mathematical analysis, computational experiment, and engineering demonstration.

Although design and materials research tend to follow different epistemological processes—materials research being deductive and design typically being more inductive in reasoning—both rest on sound scientific foundations. The consensus among workshop participants was that these differences are complementary. In a sense, materials design can be considered to be an application of design research to materials problems and as such it is to be expected that evidence gathering will be primarily dependent on the evidence gathering approach most appropriate to the underlying design framework used to tackle the materials design problem.

Workshop participants identified the following recommendations for an interdisciplinary community on materials design. Although these could be said about any research field, it is particularly important to adhere to these directives in the area of materials design. Where appropriate, we highlight recent efforts by workshop participants that exemplify the consideration of evaluation in this field.

#### State Claims Explicitly.

Researchers in materials design must be clear about the intellectual claims they are making in their work. It will be common for projects to feature both a material system and a design methodology, but it may not be the case that both are novel. Which aspects of the overall work are novel or innovative? What specific claims of new knowledge are being made? Since readers may not have deep knowledge in all aspects of the work, authors must take greater care to be clear.

#### Gather Evidence that Reflects the Claims Being Made.

Rather than have a standard prototype for all research, the community should accept and encourage research formulated to reflect the specific claims being made. Not all claims will require the production and evaluation of materials samples. For example, if the claim is that an optimization algorithm is superior to another when applied to a particular class of materials problems, then “superior” should be defined clearly and evidence should include quantitative measures related to this definition in comparison with established alternative algorithms. The physical realization of a design output by this algorithm would lend credibility to the claim that it is practically useful to materials designers but would not address the claim of superiority relative to other algorithms.

#### Evaluate Internal Consistency.

There was general consensus that establishing internal consistency of materials design methods was of fundamental importance. Internal consistency refers to the logical validity of a method. An internally consistent method is consistent with stated assumptions and will not lead to self-contradictory results. The importance of internal consistency is widely acknowledged in the design community and approaches to establishing internal consistency range from qualitative (but logical) argumentation to formal mathematical proof [1820]. Although some materials researchers had not considered this issue prior to the workshop, they acknowledged that this dimension of design research must become an underpinning of materials design research.

#### Evaluate Practical Usefulness.

Another point of consensus was that materials design methods must be evaluated for practical usefulness. Many material design methods hinge on very narrow/specific assumptions about the material system, materials processing, or engineering application context. For example, a method for optimizing the microstructure of an alloy on a fine scale may not be useful if there is no reliable processing technique to produce the microstructures specified by the method—in other words, an optimal microstructure may not necessarily be feasible. In some cases, practical usefulness may be argued logically. In others, empirical demonstration may be the only evidence that would be sufficiently compelling. Method evaluation has been addressed previously in the design literature. For example, Binder and Paredis [21] address the challenge of evaluating the performance of different design methods by introducing a method for comparing computational design approaches (see Fig. 2). They focus on key considerations for such evaluations, including the creation of fair and unbiased comparisons that take into account modeling abstraction, accuracy, and uncertainty representations of a given method. In the specific context of materials design, researchers have evaluated practical usefulness using research approaches that include benchmark problems and demonstration on an array of practical applications (e.g., as done in works such as [2224]). The main conclusion from the workshop is that it is necessary for researchers to provide evidence about the practical usefulness of newly proposed methods, but that there is not a single approach by which everyone should do so.

#### Articulate Context and Assumptions Clearly.

A disconnect can exist between the level of abstraction at which a design problem is most easily formulated and the level of abstraction at which models are most readily available. For example, it is easy to say that the main design objective is to maximize profitability of the product being designed, but less easy to relate this concern to objectives for material properties. It is common to invoke context-specific assumptions to formulate a design problem at lower levels of abstraction, such as to declare that material strength should be maximized. Although workshop participants noted this issue and stressed its importance, there was less clarity on how the materials design research community should address it. One recommendation was that researchers in the field must be clear to make explicit such assumptions, their rationale and their implications. This would reduce the risk that readers of a work misapply reported materials design methods. Another point made by participants is that techniques for spanning levels of abstraction and for formalizing contextual assumptions are areas open research challenges in the field.

### Research Challenges and Opportunities in Materials Design.

An important aspect of establishing the research needs of the field is the identification of outstanding challenges and opportunities, as well as strategies for addressing them. To this end, workshop participants were asked to identify open research questions in the field as well as interdisciplinary strategies for investigating these questions. The participants identified the following outstanding research challenges and opportunities. For each opportunity, we also present preliminary efforts by workshop participants aimed at addressing the points raised at the workshop. We note that most of the referenced works address multiple challenges raised at the workshop and we have categorized them for illustrative purposes.

#### Lack of Predictive Models.

There was general agreement that a major limitation is the lack of predictive models that would enable the establishment of quantitative process–structure–property relationships. This problem is associated with the considerable complexity of the materials space as materials information is multidimensional, multiscale and multiphysics. Further investment in models of various materials is needed, but this alone is insufficient. As has been demonstrated in recent work, advanced materials design methods can integrate data from experiments and simulations efficiently to enable efficient performance predictions of materials design alternatives. The work of Xue et al. [25] approach this general problem by coupling experiments and calculations in a sequential process for exploring and exploiting the materials design space. Their strategy focuses on balancing exploration–exploitation tradeoffs through an adaptive design loop involving prior knowledge, machine learning, designed experiments, new theory potential, and the augmenting of what has been learned thus far in the loop (see Fig. 3). Another predictive framework is described by Bessa et al. [26]. It is a data-driven framework that considers experiment design, numerical simulation for populating a material response database, and the application of machine learning techniques to arrive at new designs or modeling capabilities. The field will benefit from simultaneous advances in fundamental models and advanced frameworks such as these.

#### Computational Expense and Design Space Complexity.

The computational cost of simulation is often significant in materials science research activities and the creation of less expensive simulation capabilities is not typically addressed in the community. This creates significant design challenges owing to the need to evaluate design alternatives as efficiently as possible. To further complicate the issue, the materials design space is vast, involving multiple scales and multiphysics processes and potentially requiring an extremely large number of variables to describe. This topic was universally recognized at the workshop as being important. The design research community has recognized the general challenge previously, with much prior work on response surface methods [27,28], information economics [2931], and dimension reduction [32,33], among other related topics. However, there was broad agreement that materials design problems are particularly demanding and can benefit significantly from new representations and computational techniques.

#### Creating the Right Level of Abstraction.

It is not clear whether materials design problems can be stated with a sufficient level of abstraction necessary to develop generalizable design frameworks/policies. Without the proper level of abstraction, it is not possible to deploy particular design strategies, such as idea/concept generation. Engineering design problems are formulated at a device or system level, which is at a higher level of abstraction than material properties and features. This is particularly of issue in the concurrent design of materials and the device or system in which it is used. Here, expensive materials modeling capabilities would overburden the computational requirements for designing the larger system and require abstraction. For example, Sivapuram et al. [34] perform the simultaneous design of an engineering structure at the macroscale and the design of the material at the microscale via topology optimization of the engineering structure and the material microstructure. Their results demonstrated that a multiscale optimization approach can exploit the expanded design space that results when considerations of material tailoring are included in the formulation. Figure 4 shows notionally the concept of multiscale topology optimization of Ref. [34], in which the application of topology optimization methodologies is well suited to both the macro and microscale. Other approaches include the work of Hasan et al. [35], who have developed a multiscale computational framework for the concurrent simulated moving bed process optimization and the identification of novel zeolites. Their approach combines discovery or selection of materials process systems to screen large databases of candidate materials with certain properties. Lee et al. [36] contributed a surrogate-based multiobjective optimization methodology for enabling the simultaneous design of the shape of a textured surface and a non-Newtonian lubricant for enhancement of friction reduction. These and other contributions are important steps toward dealing with multiple levels of abstraction in materials design, but much room remains for further research. Representations and approaches that are suitable for one class of materials may be ineffective or incompatible with others. Consequently there is value in future work on techniques for specific types of materials as well as general theories for navigating different levels of problem abstraction.

#### Materials Modeling for Design.

Despite the many challenges, there are opportunities. For example, while the predictive (computational) models used by materials scientists to establish quantitative PSPP relationships tend to be incomplete and are always computationally costly, when it comes to the evaluation of design choices, approximate solutions are often times sufficient. For example, work by Pfiefer et al. [37] focuses on the process–structure linkage where they develop an integrated process–structure exploration framework designed to systematically identify viable processing conditions that result in tailored microstructures. The need to craft an appropriate cost function for optimization is noted as being critical to producing effective results and is emphasized in their work. Other work aimed at creating materials science modeling capabilities suitable for design needs includes the work of Hartl et al. [38], who have created an efficient modeling capability aimed at capturing microscale single crystalline shape memory alloy responses needed for design and property optimization. This capability is expected to provide significant efficiency improvements for the analysis of these materials while maintaining robustness in the analysis results. From the perspective of control, Friedrich et al. [39] studied the ability of acoustic focusing and direct ink writing to control microstructures. Such methods are expected to enable additively manufactured multiphase materials with unprecedented complexity.

#### Strategically Employing Informatics Tools.

There are also considerable opportunities associated with employing informatics tools to discover underlying relationships across the process–structure–property causal chain. Strategies to explore the materials space via high throughput computational and experimental approaches can provide a wealth of data to be analyzed with advanced techniques. However, there remain significant bottlenecks in the synthesis, characterization and modeling of materials. It was noted during workshop discussion that when it comes to data availability in materials design, the situation seems to be one of feast or famine. In some cases—such as high-throughput experimentation, image-based metrology, etc.—the problem is how to synthesize large quantities of data into useful conclusions. In other cases, data is rare and the challenge becomes how to infer anything useful from such limited data. The techniques required for each situation are not the same and both are important areas for future research. The importance of these topics is evident in the literature, where there are many examples of machine learning teachings being adapted to support materials design activities. For example, in addition to many of the previously referenced works, Cang et al. [40] created a multiscale methodology based on a convolutional deep belief network for feature extraction and reconstruction of complex microstructures. The technique may lead to more efficient representations of complex microstructures, which could lead to a better definition of the design space over existing ICME methods. Xiao et al. [41] study the quantification of uncertainty in the use of machine learning based interpolation methods for estimating interatomic potential energy surfaces. These models are often employed to enable the use of high fidelity information from atomic simulations in materials design and discovery processes. Cecen et al. [42] have created convolutional neural networks for 3D microstructure analytics. The goal is the extraction of highly reliable and robust structure-property linkages. Figure 5 shows three microstructures from this work and their architectural features as well as their spatial statistics. Going forward, there are many opportunities to further adapt informatics tools in a materials design context.

### Building a Sustainable Interdisciplinary Research Community.

Workshop discussions raised several questions and considerations associated with the formation of a vibrant and sustainable interdisciplinary research community on the design of engineering material systems. However, this particular topic resulted in fewer concrete recommendations than other issues discussed. The main outcomes were a list of considerations along with some possible action pathways for going forward.

#### Improve Communication Modalities.

The general consensus was that (a) it is important to establish an interdisciplinary field of materials design research and (b) students and researchers currently are not prepared well to conduct the research needed. The lack of preparation stems from traditional disciplinary training models. Design focused students often know little about materials and materials students often know little about design research. Further, and perhaps most importantly, individuals coming from one discipline often lack understanding about the objectives, communication style, and forums for dissemination of other disciplines. The need to overcome communication barriers for effective interdisciplinary collaboration was the most noted barrier to the success of the new field. Several ideas were put forth to achieve communication across disciplines and enable the learning and appreciation needed for interdisciplinary collaboration. These included the identification of venues, such as conferences and workshops, to share research and offer short courses. Also discussed was the identification of several journals amenable to the type of research expected from this field, as well as the potential for special issues to create interest from researchers in different communities.

#### Provide Education Pathways for a New Type of Student.

A key topic discussed was in regards to whether we are trying to educate researchers with expertise in all aspects of the interdisciplinary field or trying to create researchers capable of effective collaboration on an interdisciplinary team. The latter emerged as the more manageable task. Some mechanisms to deal with this were proposed, such as the development of certificate programs, project-based learning with interdisciplinary teams in traditional courses, and intensive short courses.

#### Foster Interdisciplinary Research Strategies.

A notable barrier to collaboration between materials and design experts is the fundamentally different nature of how research is conducted in these two fields. In materials science and engineering, research goals are often aligned with generating knowledge regarding fundamental scientific understanding of a material or material system. In design, research goals are often aligned with creating systematic approaches to achieve certain outcomes, and are thus, purpose driven. Features of a materials scientist that chooses to describe the effective behavior of a material are not necessarily the set of features that can be used for designing a material. A significant challenge that must be addressed for effective collaboration in this field is the creation of precise problem formulations that lead to the definition of observable or controllable parameterizations of materials and materials systems at different scales.

#### Support and Credit for Interdisciplinary Work.

A lack of proper credit for interdisciplinary work and interdisciplinary teaching was a significant concern of faculty members. Availability of funding was also noted as a key ingredient to the establishment of the field, without which progress is not expected to occur. Finally, the involvement of industry, both in terms of making it clear these skillsets are valued and in the creation of funding opportunities, was considered to be of great importance.

Given the value expected to be generated by research in this interdisciplinary field if it succeeds, a key task is ensuring that the field persists. Potentially, the best means of ensuring this is through building a sustainable research community. As noted from workshop discussions, key aspects for this involve funding mechanisms, industry involvement, interdisciplinary dissemination efforts, and training of current researchers and new graduate students. This current special issue is a testament to community efforts for ensuring interdisciplinary dissemination. Education is another way to build the community. There are several active education efforts addressing materials design.

For example, Fowler et al. [43] and Chang et al. [44] outline and provide rationale for the recently initiated interdisciplinary graduate program on data-enabled discovery and design of energy materials at Texas A&M University in the website link.4 The represented disciplines in this program are design, materials science, and informatics, as shown in Fig. 6. The overarching goal is to develop and institutionalize a new training model that produces scientists/engineers who are grounded in one discipline and have the professional and technical skills to effectively communicate within, collaborate, and lead in interdisciplinary teams, throughout their academic careers and afterward focused on materials development.

Other training efforts include the From Learning, Analytics, and Materials to Entrepreneurship and Leadership (FLAMEL) traineeship program at Georgia Tech, which aims to prepare Ph.D. students through interdisciplinary training in computing, mathematics, and material sciences to use and develop tools for improving the efficiency, design, and manufacturing of new materials in the website link.5 At Northwestern University, an ICME Masters of Science Certificate is offered in a program led by Greg Olson and Chris Wolverton in the website link.6 Another ICME training opportunity exists with the University of Michigan Summer School for Integrated Computational Materials Education led by Katsuyo Thornton and Mark Asta.7

These educational efforts are a significant step forward for the materials design community. In addition to educating a next generation of materials designers and materials design researchers, they can serve as models to be replicated or adapted at other institutions.

## Outlook

A vibrant materials innovation ecosystem is important to the prosperity and well-being of society. Materials often serve as catalysts for novel engineering technologies and can enable the solution of design problems that previously were insoluble. Over the last two decades, initiatives such as MGI and ICME have advanced the frontier of materials knowledge and contributed to the discovery of new engineering materials and material processing techniques. During this same time, the engineering design community has advanced the foundations of design methodology and created advanced techniques to support design representation, visualization, exploration, evaluation, and decision making.

Now is an opportune time for the emergence of an interdisciplinary field of engineering materials design. This workshop brought together experts in various areas of materials and design research to consider the potential for such as field. It is evident from work being done by workshop participants and others that this community already is on the rise. A strong consensus was reached about several factors that will promote sustainable high-quality research in the community. This includes core concepts and knowledge that must be shared by community members and norms for what constitutes high-quality research. In fact, a key conclusion of the workshop is that the prior concepts and research methodologies from the design and materials communities were highly compatible. Deliberate efforts must be made to sustain and grow the field, such as identifying new avenues for communication among community members and new educational and professional pathways for students with hybridized skill sets. Progress already is being made in all these fronts, but further investment is required for the community to flourish.

## Acknowledgment

The authors would like to acknowledge the support of the National Science Foundation through grants No. NSF-CMMI-1642648, NSF Workshop: Interdisciplinary Frontiers of Designing Engineering Materials Systems and grant No. NSF-DGE-1545403, NRT-DESE: Data-Enabled Discovery and Design of Energy Materials (D3EM).

## Funding Data

• National Science Foundation (Grant Nos. CMMI-1642648 and DGE-1545403).

2

NSF Award 1642648.

## References

References
1.
Olson
,
G. B.
,
1997
, “
Computational Design of Hierarchically Structured Materials
,”
Science
,
277
(
5330
), pp.
1237
1242
.
2.
Olson
,
G. B.
,
2000
, “
Designing a New Material World
,”
Science
,
288
(
5468
), pp.
993
998
.
3.
,
C.
,
Fernandez
,
M.
,
Panchal
,
J.
,
Choi
,
H.
,
Allen
,
J.
,
McDowell
,
D.
, and
Mistree
,
F.
,
2004
, “
Foundations for a Systems-Based Approach for Materials Design
,”
AIAA
Paper No. 2004-4300.
4.
McDowell
,
D. L.
,
Panchal
,
J.
,
Choi
,
H.-J.
,
,
C.
,
Allen
,
J.
, and
Mistree
,
F.
,
2009
,
Integrated Design of Multiscale, Multifunctional Materials and Products
,
Butterworth-Heinemann
, Burlington, MA.
5.
Horstemeyer
,
M. F.
,
2000
, “
,” Sandia National Laboratories, Albuquerque, NM, Technical Report No.
SAND2000-8662
6.
Horstemeyer
,
M. F.
,
McDowell
,
D. L.
, and
Fan
,
J.
,
2000
, “
From Atoms to Autos: A New Design Paradigm Using Microstructure-Property Modeling—Part 2: Cyclic Fatigue
,” Sandia National Laboratories, Albuquerque, NM, Technical Report No.
SAND2000-8661
.https://prod.sandia.gov/techlib-noauth/access-control.cgi/2000/008661.pdf
7.
Holdren
,
J. P.
, and
ad hoc Group of the National Science and Technology Council
,
2011
,
Materials Genome Initiative for Global Competitiveness
,
National Science and Technology Council OSTP
, National Science and Technology Council, Washington, DC.
8.
Scott
,
T.
,
Walsh
,
A.
, and
Anderson
,
B.
,
O'Connor
,
A.
,
2018
, “
Economic Analysis of National Needs for Technology Infrastructure to Support the Materials Genome Initiative
,” Technology Partnerships Office, Office of Innovation and Industry National Institute of Standards and Technology U.S. Department of Commerce Gaithersburg, MD, accessed June 19, 2018 https://mgi.nist.gov/sites/default/files/uploads/\\MGI\_Econ\_Analysis.pdf
9.
Allison
,
J.
,
2011
, “
Integrated Computational Materials Engineering: A Perspective on Progress and Future Steps
,”
JOM
,
63
(
4
), pp.
15
18
.
10.
Glamm
,
R. J.
,
,
D. M.
,
Pripstein
,
E. D.
, and
Cotton
,
J. D.
,
2015
, “
Recent Progress in Implementation of ICME for Metallic Materials in the Airframe Industry
,”
AIAA
Paper No. AIAA 2015-0199.
11.
Howe
,
D.
,
Goodlet
,
B.
,
Weaver
,
J.
, and
Spanos
,
G.
,
2016
, “
Insights From the 3rd World Congress on Integrated Computational Materials Engineering
,”
JOM
,
68
(
5
), pp.
1378
1384
.
12.
OMara
,
J.
,
Meredig
,
B.
, and
Michel
,
K.
,
2016
, “
Materials Data Infrastructure: A Case Study of the Citrination Platform to Examine Data Import, Storage, and Access
,”
JOM
,
68
(
8
), pp.
2031
2034
.
13.
Hill
,
J.
,
Mulholland
,
G.
,
,
K.
,
,
R.
,
Wolverton
,
C.
, and
Meredig
,
B.
,
2016
, “
Materials Science With Large-Scale Data and Informatics: Unlocking New Opportunities
,”
MRS Bull.
,
41
(
5
), pp.
399
409
.
14.
Kalidindi
,
S. R.
,
Brough
,
D. B.
,
Li
,
S.
,
Cecen
,
A.
,
Blekh
,
A. L.
,
Congo
,
F. Y. P.
, and
Campbell
,
C.
,
2016
, “
Role of Materials Data Science and Informatics in Accelerated Materials Innovation
,”
MRS Bull.
,
41
(
8
), pp.
596
602
.
15.
Borrego
,
M.
, and
Newswander
,
L. K.
,
2010
, “
Definitions of Interdisciplinary Research: Toward Graduate-Level Interdisciplinary Learning Outcomes
,”
Rev. Higher Educ.
,
34
(
1
), pp.
61
84
.
16.
Choi
,
H.
,
McDowell
,
D. L.
,
Allen
,
J. K.
,
Rosen
,
D.
, and
Mistree
,
F.
,
2008
, “
An Inductive Design Exploration Method for Robust Multiscale Materials Design
,”
ASME J. Mech. Des.
,
130
(
3
), p.
031402
.
17.
Xu
,
H.
,
Li
,
Y.
,
Brinson
,
C.
, and
Chen
,
W.
,
2014
, “
A Descriptor-Based Design Methodology for Developing Heterogeneous Microstructural Materials System
,”
ASME J. Mech. Des.
,
136
(
5
), p.
051007
.
18.
Hazelrigg
,
G. A.
,
2003
, “
Validation of Engineering Design Alternative Selection Methods
,”
Eng. Optim.
,
35
(
2
), pp.
103
210
.
19.
,
C. C.
,
Pedersen
,
K.
,
Emblemsvg
,
J.
,
Bailey
,
R.
,
Allen
,
J. K.
, and
Mistree
,
F.
,
2006
, “
The Validation Square: How Does One Verify and Validate a Design Method
?,”
Decision Making in Engineering Design
,
American Society of Mechanical Engineers
,
New York
, pp.
303
314
.
20.
Olewnik
,
A. T.
, and
Lewis
,
K.
,
2005
, “
On Validating Engineering Design Decision Support Tools
,”
Concurrent Eng.
,
13
(
2
), pp.
111
122
.
21.
Binder
,
W. R.
, and
Paredis
,
C. J.
,
2017
, “
Optimization Under Uncertainty Versus Algebraic Heuristics: A Research Method for Comparing Computational Design Methods
,”
ASME
Paper No. DETC2017-68371.
22.
Massoni
,
B.
, and
Campbell
,
M. I.
,
2018
, “
,”
ASME J. Manuf. Sci. Eng.
,
140
(
3
), p.
031001
.
23.
Galvan
,
E.
,
Malak
,
R. J.
,
Gibbons
,
S.
, and
Arroyave
,
R.
,
2017
, “
A Constraint Satisfaction Algorithm for the Generalized Inverse Phase Stability Problem
,”
ASME J. Mech. Des.
,
139
(
1
), p.
011401
.
24.
Nellippallil
,
A. B.
,
Shukla
,
R.
,
Ardham
,
S.
,
Goh
,
C.-H.
,
Allen
,
J. K.
, and
Mistree
,
F.
,
2017
, “
Exploration of Solution Space to Study Thermo-Mechanical Behavior of AA5083 Al-Alloy During Hot Rolling Process
,”
ASME
Paper No. DETC2017-68173.
25.
Xue
,
D.
,
Balachandran
,
P. V.
,
Hogden
,
J.
,
Theiler
,
J.
,
Xue
,
D.
, and
Lookman
,
T.
,
2016
, “
Accelerated Search for Materials With Targeted Properties by Adaptive Design
,”
Nat. Commun.
,
7
, p.
11241
.
26.
Bessa
,
M.
,
,
R.
,
Liu
,
Z.
,
Hu
,
A.
,
Apley
,
D. W.
,
Brinson
,
C.
,
Chen
,
W.
, and
Liu
,
W. K.
,
2017
, “
A Framework for Data-Driven Analysis of Materials Under Uncertainty: Countering the Curse of Dimensionality
,”
Comput. Methods Appl. Mech. Eng.
,
320
, pp.
633
667
.
27.
Jin
,
R.
,
Chen
,
W.
, and
Simpson
,
T. W.
,
2001
, “
Comparative Studies of Metamodelling Techniques Under Multiple Modelling Criteria
,”
Struct. Multidiscip. Optim.
,
23
(
1
), pp.
1
13
.
28.
Yeten
,
B.
,
Castellini
,
A.
,
Guyaguler
,
B.
, and
Chen
,
W.
,
2005
, “
A Comparison Study on Experimental Design and Response Surface Methodologies
,”
SPE Reservoir Simulation Symposium
, The Woodlands, TX, Jan. 31–Feb. 2.
29.
Galvan
,
E.
,
Hsiao
,
C.
,
Vermillion
,
S.
, and
Malak
,
R.
,
2015
, “
A Parallel Approach for Computing the Expected Value of Gathering Information
,”
SAE Int. J. Mater. Manuf.
,
8
(
2
), pp.
271
282
.https://www.jstor.org/stable/26268713
30.
Moore
,
R. A.
,
Romero
,
D. A.
, and
Paredis
,
C. J.
,
2014
, “
Value-Based Global Optimization
,”
ASME J. Mech. Des.
,
136
(
4
), p.
041003
.
31.
Messer
,
M.
,
Panchal
,
J. H.
,
Krishnamurthy
,
V.
,
Klein
,
B.
,
Yoder
,
P. D.
,
Allen
,
J. K.
, and
Mistree
,
F.
,
2010
, “
Model Selection Under Limited Information Using a Value-of-Information-Based Indicator
,”
ASME J. Mech. Des.
,
132
(
12
), p.
121008
.
32.
,
M. K.
,
Li
,
M.
,
Hu
,
C.
, and
Mackenzie
,
C. A.
,
2017
, “
High-Dimensional Reliability Analysis of Engineered Systems Involving Computationally Expensive Black-Box Simulations
,”
ASME
Paper No. DETC2017-68273.
33.
Hajikolaei
,
K. H.
, and
Wang
,
G. G.
,
2014
, “
High Dimensional Model Representation With Principal Component Analysis
,”
ASME J. Mech. Des.
,
136
(
1
), p.
011003
.
34.
Sivapuram
,
R.
,
Dunning
,
P. D.
, and
Kim
,
H. A.
,
2016
, “
Simultaneous Material and Structural Optimization by Multiscale Topology Optimization
,”
Struct. Multidiscip. Optim.
,
54
(
5
), pp.
1267
1281
.
35.
Hasan
,
M. F.
,
First
,
E. L.
, and
Floudas
,
C. A.
,
2017
, “
Discovery of Novel Zeolites and Multi-Zeolite Processes for p-Xylene Separation Using Simulated Moving Bed (Smb) Chromatography
,”
Chem. Eng. Sci.
,
159
, pp.
3
17
.
36.
Lee
,
Y. H.
,
Schuh
,
J. K.
,
Ewoldt
,
R. H.
, and
Allison
,
J. T.
,
2018
, “
Simultaneous Design of Non-Newtonian Lubricant and Surface Texture Using Surrogate-Based Optimization
,”
AIAA
Paper No. AIAA 2018-1906.
37.
Pfeifer
,
S.
,
Wodo
,
O.
, and
Ganapathysubramanian
,
B.
,
2018
, “
An Optimization Approach to Identify Processing Pathways for Achieving Tailored Thin Film Morphologies
,”
Comput. Mater. Sci.
,
143
, pp.
486
496
.
38.
Hartl
,
D. J.
,
Kiefer
,
B.
,
Schulte
,
R.
, and
Menzel
,
A.
,
2017
, “
Computationally-Efficient Modeling of Inelastic Single Crystal Responses Via Anisotropic Yield Surfaces: Applications to Shape Memory Alloys
,”
Int. J. Solids Struct.
,
136–137
, pp.
38
59
.
39.
Friedrich
,
L.
,
Collino
,
R.
,
Ray
,
T.
, and
Begley
,
M.
,
2017
, “
Acoustic Control of Microstructures During Direct Ink Writing of Two-Phase Materials
,”
Sens. Actuators A: Phys.
,
268
, pp.
213
221
.
40.
Cang
,
R.
,
Xu
,
Y.
,
Chen
,
S.
,
Liu
,
Y.
,
Jiao
,
Y.
, and
Ren
,
M. Y.
,
2017
, “
Microstructure Representation and Reconstruction of Heterogeneous Materials Via Deep Belief Network for Computational Material Design
,”
ASME J. Mech. Des.
,
139
(
7
), p.
071404
.
41.
Xiao
,
W.
,
Li
,
Y.
, and
Wang
,
P.
,
2018
, “
Uncertainty Quantification of Atomistic Materials Simulation With Machine Learning Potentials
,”
AIAA
Paper No. AIAA 2018-2166.
42.
Cecen
,
A.
,
Dai
,
H.
,
Yabansu
,
Y. C.
,
Kalidindi
,
S. R.
, and
Song
,
L.
,
2017
, “
Material Structure-Property Linkages Using Three-Dimensional Convolutional Neural Networks
,”
Acta Mater.
,
146
, pp.
76
84
.
43.
Fowler
,
D. A.
,
Arroyave
,
R.
,
Ross
,
J.
,
Malak
,
R.
, and
Banerjee
,
S.
,
2017
, “
Looking Outwards From the Central Science: An Interdisciplinary Perspective on Graduate Education in Materials Chemistry
,”
Educational and Outreach Projects From the Cottrell Scholars Collaborative Undergraduate and Graduate Education
, Vol.
1
,
ACS Publications
, Washington, DC, pp.
65
89
.
44.
Chang
,
C.-N.
,
Semma
,
B.
,
Pardo
,
M. L.
,
Fowler
,
D.
,
Shamberger
,
P.
, and
Arroyave
,
R.
,
2017
, “
Data-Enabled Discovery and Design of Energy Materials (D3EM): Structure of an Interdisciplinary Materials Design Graduate Program
,”