Morphing Matter for Teens: Research Processes as a Template for Cross-Disciplinary Activities

“Morphing matter” refers to a class of active materials that can be controlled to change shape, properties, or functionalities due to external stimuli [61]. Morphing matter can be engineered to respond to stimuli including temperature [1], pH [41], moisture [50], pressure [82], and magnetic fields [87]; responses can include changes in shape, texture [47], permeability [58], color [17], or stiffness [24] toward a wide range of functional and experiential goals, including suitability for conventional engineering [13] and bioengineering [44] scenarios as well as day-to-day contexts such as food [77] and clothing [43].

Morphing matter has been studied by researchers in a wide variety of disciplines. Materials Scientists develop morphing materials for challenging contexts [36] and in response to theoretical models [16], and computer scientists apply heuristic and statistical methods to model and predict morphing reactions [85], supporting interactive predictive and inverse design with morphing matter [30]. Architects and designers have used morphing matter for form-finding [5], deployability [27], and adaptive building skins [68]. Within HCI, morphing matter research often focuses on improving accessibility to the field with inexpensive [22, 84] or easy to fabricate materials [78], integrating morphing behavior into specific contexts like fine dining [79], and streamlining the design process with software toolkits [31].

In preparing interdisciplinary activities for a student audience with the goal of providing a meaningful and educational experience, we join a large and diverse set of initiatives and research projects. Related educational tactics have a long history, from an early emphasis on student self-determination [54] and hands-on learning [83] to the constructivist theory of learning as a inherently participatory process of building knowledge structures [33] and constructionist ideas of learning as physical and material process [60]. These approaches have been particularly manifest in contemporary STEM (Science, Technology, Engineering, and Math) or STEAM (the preceding, with the addition of Art) education, which combines an emphasis on “technological” tools and materials like microcontrollers and digital fabrication with curricular science, math, and engineering education [34] and typically equity-centered values of personalized individual learning (including broadened participation and cultural relevance) [21, 63, 71]. Notably, STE[A]M education often takes place alongside, and often outside, of students’ main education context and can introduce students to concepts that are not yet integrated into curricular standards, such as early work on defining computational thinking and how it could be taught [9]. It also frequently takes its cues from the broader “Maker Movement” of adult hobby “makerspaces” and “fab[rication] labs,” showing how broader subcultural movements can serve as templates for learning experiences [62]. These values informed our overall goal of supporting an authentic experience with the “real stuff” of research, and they influenced our activity planning, in which we prioritized a blend of hands-on making and group-based discussion and reflection.

With our out-of-school context and de-emphasis on specific quantified learning goals, we are particularly aligned with “tinkering” approaches which center context- and participant-specific trajectories, multi-faceted and open-ended projects, and iterative prototyping as a core mode of engagement [6]. When compared to pure tinkering, the “research as a template” approach documented in this paper involves a relatively high degree of structure: we guided our participants through a planned sequence of activities demonstrating different ways of viewing the overall problem. However, the tinkering paradigm, with its support of fluidity and iteration, was a fundamental aspect of our activities–as indeed it is a foundation of research itself. As such, we base our method of assessment (section 5) on the insights of researchers working in tinkering environments, who observe a variety of indicators of learning more suited to such contexts than quantitative testing assessments would be [64].

However, unlike the researchers behind these works, we are not primarily a makerspace or tinkering studio. Our collaboration between a university research laboratory and an educational nonprofit might be more closely compared to a “visiting artist” [20], albeit for a cross-disciplinary topic. As such, we present our approach to bridging between the research and educational contexts, without necessarily having the expertise or infrastructure of dedicated maker-instructional contexts.

We chose to frame our workshop around a fashion-based project. In terms of learning outcomes, fashion design incorporates many kinds of challenge, including visual composition, understanding social context, handcraft skills, and soft engineering. Additionally, we hoped that basing our workshop exploration on fashion would support the participants in working on a project that could be meaningful to them.

We chose to center our workshop on inflatable actuators in the style of Ou et al’s Aeromorph paper [57]. We show the production of one such actuator in Figure 3: these are airbags made by heat-bonding thin polyethylene plastic sheet, such as plastic sandwich bags, into flat, airtight pouches. Each pouch has a lightweight hose attached via a barb, for inflation, as well as additional sealed inner edges. When inflated, the inner edges constrain the three-dimensional geometry of the overall airbag, causing to bend. In the style of airbag actuator that we initially showed the participants, the sealed inner edges form a diamond shape, and the bending angle of the actuator is determined by the height and width of the diamond relative to the overall airbag. Variant geometries alter the overall shape of the pouch and/or the arrangement of inner edges for a variety of bending, creasing, curling, or twisting effects.

The airbag can be inflated very simply by using a large syringe fitted with a Luer Lock tip to interface with the tubing, or it can be controlled with a microcontroller driving an electric air pump. The tools needed to produce these actuators are an impulse heat sealer, a hot glue gun, and scissors or a hobby knife. The materials are the polyethylene sheet, plastic hose barbs, rubber tubing compatible with the hose barbs, printer paper for selecting masking during the heat bonding process, and some means of pumping air. Overall, this type of actuator is inexpensive to produce, is lightweight and suitable for on-body design, and has a wide range of possible variations.

Our overall goal was to lead the participants through processes patterned after our own research. To make the processes legible in such a short timeframe, we scaffolded each to a greater or lesser extent. In particular, since we had pre-selected the airbag actuator as our focus, our identifying phase included some explanation of related technical terms like “pneumatics” and “actuator,” but was mainly a group discussion of how and why pneumatics, and pneumatic actuators, are used in the world. We additionally started participants off with step-by-step instructions on constructing the airbags based on our own accumulated tips and tricks, though we encouraged them to iterate and improve the fabrication workflow as they saw fit. We lead an initial round of characterizing as a full-group guided activity, and we structured the early brainstorming and peer feedback portions of integrating to keep up the tight pace of the workshop. While the second day of the workshop was primarily about the integrating and communicating processes, we were confident that self-directed returns to each other process would occur as well, due to the cyclic and iterative nature of research. Evaluation within each phase was typically guided by questions from us. For example, in the constructing phase, we asked questions like “what problems could make your actuator not work? How could you fix it?”; in the characterizing phase, our questions included “which of the measurement systems we discussed should we use, and why?”

Within this overall flow, we wanted to insure that we touched on a healthy assortment of the skills that tacitly comprise our research processes, as established and practiced in the component disciplines of morphing matter research. For example, empirical material characterization, derived from materials science and mechanical engineering, requires knowing how to use physical measuring implements and algorithmic models, as well as understanding how to assess reasonable data quantity and accuracy. We generated a list of medium-level skills relevant to carrying out each research process, tuned specifically for our pneumatic fashion focus, as well as evaluating the success of that process. These skills are shown as the left column in Table 1.

Within each day of the two-day workshop, we aimed for a range of activity types, from full-group discussions to facilitated breakout activities to hands-on self-directed making time. A summary of our activity plan is in Table 1. In the rest of this subsection, we explain specific activities.

As primarily morphing matter researchers, not educators, we did not have a firm understanding of current curricular expectations. To check that our lesson plan would likely be at a level of challenge that was exciting but not overly frustrating (the “zone of proximal development” [52], and to understand how morphing matter processes might be relevant to our participants’ overall learning, we decided to look to learning standards. Learning standards are educational guidelines that outline the expected knowledge and skills students should acquire at various stages of their education. These standards serve as foundational elements for schools in formulating curricula and assessment frameworks across various grades and disciplines. They are typically established and published by either governmental or non-governmental organizations, such as professional teaching associations, within a particular geographic region. Under the guidance of our educator collaborator, we examined a variety of locally-relevant standards:

Most of the learning standards in our region focus either on a particular subject (e.g. NSTA NGSS, National Core Arts Standards) or context (e.g. Common Core is meant for classroom based learning), so we opted for a combination of standards from varying disciplines to appropriately characterize the learning goals of the workshop. In Table 2, we show how our morphing matter process skills align to learning standards from these sources.

We recruited via email to local schools and after-school programs. We had twelve participants, of which ten stayed for the entire workshop. (One left early on the second day and the other was unable to attend on the second day due to a family emergency.) Demographic and prior experience information is shown for the ten full participants in Table 3.

Of note, participants P6, P7, and P10 had a prior friendship. Participants P8 and P9 were home-schooled sisters, and the other participants came from six different local high schools. As mentioned in section 4, because our workshop was held as part of a larger effort toward STEM outreach for girls, all of our participants were female (self-identified).

Building on the curricular relevance demonstrated in section 4.3, we wanted to understand the morphing matter research skills’ salience to our actual participants. To do so, we analyzed participant narratives of their project and experience in their own words through two sources of self-narrative data: their oral project presentations and our one-on-one post-workshop interviews. (The post-workshop survey was largely multiple-choice/Likert-scale questions, not the participants’ own words.)

As described in section 4.2.7, the participants gave presentations thrice: once as an initial practice attempt in small groups, a “first round” for the whole workshop group, and “second round” in front of a small audience of invited guests. The presentations were generally more succinct for the final audience, averaging about two minutes in the first round, with an additional two minutes of questions per project, a minute and a half for the second round. Because some participants worked in groups for the final project, not every participant spoke in all presentations. In the three-person team, P6 and P7 co-presented for the first full-group round, and P10 presented for the guests round. Additionally, P5 presented her joint project with P11, who had left just before presentations. Two participants (P11 and P12) left before presenting or being interviewed, and are therefore not included in this analysis.

Our semi-structured post-workshop interviews were limited by their brevity, lasting an average of five minutes. This length was constrained by the relatively small number of our team trained to give interviews and the available time at the end of the two-day workshop. The post-survey given just before the interviews asked for feedback on individual activities and thus likely reminded the participants about what they had done, but the oral interview questions did not reference specific activities or learning goals. Instead, we asked these high-level questions:

We don’t assume that participant self-reports of skills learned are necessarily accurate; however, their responses could point toward what was new, exciting, or relevant to the participants.

To structure our analysis, we selected a subset of evaluation criteria from our list of relevant learning standards (Table 2). We prioritized criteria that were particularly relevant to morphing matter research, that cut across disciplines, and that were not explicitly enforced by us (e.g. not "Document the process of developing ideas from early stages to fully elaborated ideas" because we specifically told them to do so).

For each selected goal, we looked for instances of each participant either mentioning a related activity directly, or describing their experience in a way that shows evidence of learning in that area. For example, for the goal “learning to use tools, materials, and artistic conventions,” P1, P3, P8, P9 all specifically recollected “learning to use the heat sealer” in their interviews (direct mention), and P7 discussed her team’s incremental understanding of how to work with heat-sealed plastic, noting that “the material was hard to fold once it was already stapled down. So I might cut it into different pieces and then press those all together” (evidence of learning).

We present an overview of this data in Table 4.

Unsurprisingly because it was the focal activity of the workshop and the topic of the presentation, “Designing/refining a solution based on student-generated evidence, scientific knowledge, and prioritized criteria/tradeoffs” is well-represented: most participants mentioned how they designed and refined their pneumatic wearable with an eye toward tradeoffs and what they understood about diamond-inflatable actuator design. Especially within the presentations, participants mentioned factors influencing their results including time constraints, successful or failed prototypes, and how actuator motions would ideally cause their desired results. Participants particularly described their workarounds for assembling their project quickly (e.g. P2 used safety pins instead of sewing, and P9 made a deliberately partial prototype) and their processes for location and fixing holes in their air bags, indicating that they were particularly involved in troubleshooting and problem-solving processes.

Every participant – even P5, who was quite taciturn – touched on “making works that convey personal meaning and interpreting meaning in the works of others” at least once, whether in their presentation or their interview. We hypothesize that this is partially influenced by how we structured the initial brainstorming of the projects, which emphasized user and context. However, some of the “purposes” were deliberately exaggerated to the point of comedy for personal joy (e.g. an octopus-shaped sun hat, because P9 wanted to make something big), and even projects that weren’t explained as having a particular “purpose” were connected to something relevant to that participant (e.g. P4’s morphing skirt, because she is inspired by beetle exoskeletons).

Very few participants mentioned or invoked “using algorithmic representations to describe or support explanations” within the presentations or interviews. The strongest recollection was P2, whose commentary was still fairly lukewarm, noting that “it was a little bit difficult for me, but I was still able to do it.” This is in contrast to the reality that every participant did, in fact, participate in both the Telephone Game and main characterizing activities, with many spirited discussions especially during the Telephone Game, and participants engaged enough to call out errors in a proposed methodology (P9) and question/double-check specific data points (P6 and P7). Participants also clearly used their understandings of Ou et al’s identified parameters in refining their projects; for example, by adding more diamonds to produce more cumulative bending. We hypothesize that these activities may have been the least novel, or most school-like, and therefore least salient or interesting to most participants. Additionally, in the presentations, participants tended not need need algorithmic representations because they could simply use visual or gestural descriptions (e.g. “it bends like this”). Lastly, while we were focusing on the physical/geometric representation in this analysis, we observe that participants did tacitly incorporate the inherent time dimension of morphing matter into their presentations; for example, all three posters of our case study participants (Figure 4) include “before and after” sequential imagery.

We were curious how students’ own cross-disciplinary interests and aspirations influenced their perceptions of, and interactions with, morphing matter research processes because STEAM education, particularly in e-textiles, is often positioned as a basis for positively shaping student aspirations and improving transfer across knowledge domains, particularly for minoritized students such as girls in electrical engineering [11] However, we noticed that many of our participants already self-defined as multidisciplinary learners with a wide variety of interests and aspirations. Many of them had done similar blended projects before, though not necessarily framed as research.

To better understand how participant backgrounds and aspirations influenced their experiences with the various research processes, and to understand how learning or inspiration could be supported, we document three case studies [86]: one group of three friends who chose to collaborate on their final project, as well as two individuals. We chose these three cases because, together, they provide good representative coverage of the overall participant group: different ways of working, different priorities, and different backgrounds.

Our analysis in this section is based on participant survey data (pre- and post-workshop), the post-workshop interviews, our audio/video and photographic recordings of the workshop and the artifacts produced, and our instructing team’s observations of the participants both during the session and in reviewing the recordings. To help spur and structure the instructor observations, we considered the four types of indicators of learning in "tinkering" environments proposed by Petrich et al. – engagement, intentionality, innovation, and solidarity [64] – as a prompt. To summarize Petrich et al., indications of engagement include “expressions of joy, wonder, frustration, and curiosity,” and “work inspired by prior examples”; intentionality is evident in variation, personalization, and self-direction toward projects and efforts; innovation includes “repurposing ideas/tools,” deliberately “redirecting efforts,” and “complexification of processes and products”; solidarity can be shown in borrowing and sharing “ideas, tools, approaches” and “helping others to achieve their goals.” For each, we asked each member of the instructing team for examples of each of our focal participants showing that indicator. These examples included some recollected anecdotes, as well as corroborating evidence from our recorded data. We used Google Sheets [28] to sort and analyze the survey data and Dovetail [35] to transcript, tag, and cross-reference the recording and interview data. All authors of this paper contributed to tagging the data, which was done iteratively, with each data source processed by at least two and typically three members of the team. Initially, we tagged based on learning goals, participant reactions (such as “wants to learn more” or “isn’t sure why this would be useful”), and each author’s discretionary highlighting of incidents worth discussion. As we periodically discussed emergent patterns, we re-visited the data to synthesize the tags into thematic codes to develop our “ways of engaging research processes” thematic framing for understanding the participant experiences.

Instead of choosing just one sub-process to focus on, we designed a multi-activity plan touching, at least briefly, on each mode of the whole morphing matter research process. While this approach had drawbacks – “not enough time!” was a frequent piece of feedback from the participants – the somewhat chaotic and multivalent atmosphere of the workshop was both authentic to our own methods and made space for participants to engage with aspects of the process that were relevant to their own interests and aspirations. We chose teenagers as our participants because they are capable of complex and self-guided work and are at a crucial moment with respect to their future aspirations and goals.

In the workshop itself, we did not have any higher-level conversations about the nature of research or how it is practiced in human-computer interaction, nor did we present specific morphing matter research other than the immediately relevant Aeromorph paper. Nonetheless, as discussed in section 5.2, participants did encounter various facets of research, as we experience it, in their own ways – as a collaborative building process, as a set of methodologies, and as a path for acquiring knowledge. These participant trajectories point to various roles that research activities could play in education over a longer timeframe. For example, P4 might be encouraged to articulate exactly what methodologies she was tacitly applying to her project to clarify how her existing knowledge relates to her future aspirations. P9 might similarly be encouraged to take a principled look at how her varied knowledge fits into higher taxonomies, sparking novel connections for her which could feed cyclically into new paths of learning.

A discipline is “a branch of learning or knowledge; a field of study or expertise” [59]. Any discipline has practices that have been honed into conceptual toolkits for its practitioners bring to bear on complex problems, including methods for conducting inquiry and values for assessing results. For example, the Scientific Method is a framework for rigorously constructing new scientific knowledge; scientific practice is ideally precise in its aims and reproducably accurate in its data and conclusions. Design methods like sketching, contextual inquiry, and critique are scaffolds for developing specific proposals and formalizing reflection. Design practice is ideally multivalent, context specific, and responsive to complex problems.

Our three core research processes often (though not always) align to three primary disciplinary areas:

In progressing through the processes, morphing matter research can be summarized as a shifting interplay between three archetypal disciplines, serially drawing on deep procedural knowledge to solve problems that may be intractable otherwise. We believe that approaches like this are necessary to solve the complex problems at the frontier of contemporary research.

However, we observe that shifting disciplinary foci are atypical in current K-12 education. STE[A]M approaches overlap to some extent with ours. However, STEAM is typically implemented in one of two modes. In the first, described in Mejias et al [53], disciplines are typically “unidirectionally instrumentalized”: one discipline is used as a frame or a prop for the other; for example, students may be encouraged to draw pictures on a microcontroller project, allowing the activity to claim a veneer of “art” over what is more fundamentally a programming lesson. Indeed, the specific meaning of “art” is often vague, and may refer to superficial aesthetics or simply any element of learner creativity [14] (an attitude trivializing both Art and STEM practices). Design as a discipline might be conflated with Art, considered only in the context of Engineering Design [19], or ignored entirely. In the second mode, described by Bevan et al [4], the disciplines are hybridized: the intersection of “STEM” and “Art” gives rise to a new, third discipline with its own methods and values. These methods may not fully represent the component disciplines–for example, the skills needed to successfully complete an e-textiles project, while valuable in their own right, do not necessarily translate to success with either electronics or textiles taken individually [40]–and Bevan points out that the arts practices in particular are often under-represented in the hybrid. Additionally, it may be difficult for students to understand how a specific, hybridized STEAM project relates to their other learning or their own interests, even if its component practices would.

We see morphing matter research as a key exemplar for how cross-disciplinarity can be practiced, not as an unbalanced or blended practice, but as one in which the disciplines are interleaved. By giving serious consideration to each base discipline, deeper insights and connections can be achieved. For example, we specifically set aside times for context-based brainstorming, peer critique, and developing presentation skills; these key Design methods are not frequently described in STEAM teaching reports. While it would be premature to make specific claims about learning outcomes, given our small participant pool, we observe that student self-reports did use language from all of design (e.g. P3 discussing fabric choices for aesthetic effect, Team Puff Muscle situating their project in a very specific scenario), science (P4’s sealing pattern experiments), and engineering (P9’s discussion of a single actuator as a prototype). Topic shifts (e.g. when we wrapped up the Characterizing activity and began the discussion of fashion design as a context) additionally provided opportunities to re-invigorate the room’s energy levels and re-integrate participants who were less engaged.)

We believe that morphing matter, with its deep blend of engineering, science, and design, is a particularly apt research domain to use as a template for cross-disciplinary education. However, we do not think that morphing matter is the only such domain. Because we have found that collaborating on the design of educational workshops can be highly worthwhile – it can help researchers reflect on their own processes (potentially leading to new lines of inquiry), it can improve public discourse about a topic area, and of course it might inspire the next generation of colleagues – we hope that this work inspires other researchers in other domains as well. In this section, we offer guidelines based on our experiences for such researchers.