We presume that the link between intention in design and the materials we choose is deterministic: The right collection of materials will yield the desired effects both aesthetically and performatively. For example, a glazed curtain wall may be designed to visually dematerialize the exterior of the building while providing a predictable and measurable quantity of daylight to the interior. In this mode, the material is the effect, and as such, it is often foregrounded. Indeed, many of us are more familiar with the facade materials of Herzog & de Meuron’s projects than we are with the projects’ spatial organizations. This de facto coupling of material with effect is reflected in our classification systems, whether organized by material—metal, wood, concrete—or by end use—tile, roofing, cladding. Conventional materials are thus understood and applied as artifacts: physical, tangible, and static. Our classification systems, such as the Construction Standards Institute categories (CSI), group these artifacts into a familiar and finite set of extensively documented materials with which there is collective experience.
 
It is through this lexicon that “smart” materials have entered the profession of architecture. While definitions as to smartness abound, the most generalized one is that smart materials are transformative: The transformation may be within the material itself, as in one of its properties or its physical state, or the material could be the vehicle to transform other things, such as energy forms or the surrounding environment. These are materials, then, whose most salient descriptor is motion. Even though this is starkly counterintuitive to the inherent stasis of conventional materials, we have tried to shoehorn smart materials into our normative categories. Any smart material with a transparent or translucent state is placed in the glazing category, and any light-emitting material is dispatched to the electric-lighting category. We do this to increase our comfort level with their use. When something like an electrochromic is placed into the glazing category, we understand it and thus use it as a substitute for glass, albeit a very expensive substitute. We specify it in the same sizes and shapes as glass. We justify its expense relative to standard glazing by comparing their performance. A smart material becomes but one of many choices for a given end use.
 
This subordination of the material to how we usually do things often prevents us from exploiting the truly remarkable characteristics of these materials. Instead, much of the attention has been placed on demonstration. For example, installations with color-changing materials didactically demonstrate that the materials change color. In the desire to provide a visual marker of interaction with these materials, we are instead continuing to articulate them as building-scaled artifacts. The unrealized potential of smart materials resides in their instrumentality, which is a bit of a paradox: The spectacle of their transformation, the aspect that has so captivated architects and designers, is incidental.
 
As materials of motion, all smart materials involve an energy transfer in some form or another for transformation to take place. The type of energy that is transferred determines how the material state—temperature, pressure, density, or internal energy—will change. The quantity of energy that is transferred to produce this change is determined by the properties of the material. This relationship governs the behavior of all materials, including smart ones. In conventional materials, the properties scale the relationship between state change and energy transfer. For example, a material’s specific heat (property) will determine how much heat (energy) is needed in order to change its temperature by a specified amount. From the deflection of a material under a structural load to the color that a material appears to be, the relevant material properties will produce repeatable results. Smart materials add a wrinkle to this predictability in that the relationship is no longer scalar or linear. A material’s property might shift to another value, an energy input might turn into a different form of energy.
 
Depending upon what is changing and what motivates the change, we can group smart materials into four categories:
  1. A change in state produces a change in the material property. A thermochromic material reflects a different color when the temperature changes. The material state of temperature determines the material’s property of spectral reflectivity. In this category, we find such things as liquid crystals (driven either by a change in temperature or chemical concentration).


  2. An energy input produces a change in the material property. A photochromic material changes its spectral transmissivity (the ability for light to pass through it) when radiation (light) transfers into the material. This category also describes electrochromics (an input of electrical energy) and many of the viscosity-changing materials.


  3. An energy input is transformed into a different energy output. A photovoltaic converts an input of radiation into an output of electricity. Most of the semiconductor-based smart materials are in this category, such as light-emitting diodes (electrical energy into radiation).


  4. A change in state produces another change in state (internal energy) that transforms properties and the energy output. Shape memory alloys are in this category: When the temperature of the alloy changes, the material undergoes a crystalline phase change (internal energy) that produces an output of mechanical energy (strain).

We begin to recognize that smart materials provide the ability to precisely design behaviors. The material becomes secondary to the effect, and if we properly map how a transformation could and should proceed, then we will discover that numerous materials are capable of producing the behaviors to yield a desired effect. Decoupling material from effect is but one challenge in working with these materials—a more difficult one is determining how to directly design for effect. An effect should be more than the production of a different color, it should result from the instrumentalization of the different color. Does the change in color reduce the ambient light levels? Does it obscure visibility? Should it activate an interior system? Could it visually change an object’s apparent location in the field of view? Looking at just one behavior—the transmission of light through a transparent medium—we find that several smart materials are capable of controlling transmission, but there are significant differences in how they do so, as well as in the specific results. The spectral composition of the light might be altered, the light may be diffused or redirected, view may be diminished, and ultraviolet or infrared radiation may be absorbed or reflected. Smart glazing is often proposed as a seamless method for stabilizing interior light transmission as daylight levels change, but no material actually does this.

The smart material that has penetrated furthest into the field of architecture is the light-emitting diode, or LED. Less than five years ago, LEDs were found only in novelty applications, such as disco dance floors. White light from LEDs has recently proliferated into general, or ambient, room lighting in competition with fluorescent and High Intensity Discharge (HID) systems. Although less efficient than conventional general lighting, and much more expensive, LED lighting has nevertheless been designated as a “green” technology, which has helped encourage its use. This evolution illustrates the path taken by many smart materials as they enter into the architecture field—first demonstrative, then performative as we fold them into our normative practice.

If we could step away from standard practice and conventional applications, we would find that smart materials present unprecedented opportunities to challenge the accepted, and unquestioned, beliefs about how building systems should perform. A closer examination of LEDs would reveal their specific characteristics: narrow spectral bands, precise angles and directionality, tiny size, discrete addressability, and reliance on DC power. None of these characteristics are desirable for a general lighting system based on a wide spectral bandwidth, diffuse spreading of the light over large surface areas, and an AC power supply. The characteristics of LEDs, however, are a remarkable match for how the receptors in the eye actually perform. Minute shifts in luminance in the field of view determine our reading of depth and space, while differences in discrete wavelengths determine how we recognize figures. Lighting design for the eye rather than for the building would result in radically different locations and distribution of LEDs, away from the ceiling and discretely placed within certain angles of view. Architects would be able to “script” the view field, and in so doing, meet needs-enhanced vision with dramatic reductions in energy use.

The Portable Light Project, by Sheila Kennedy, with Boston-based Kennedy & Violich Architecture and MATx, demonstrates the instrumental mating of material behavior to phenomena. Flexible photovoltaics, woven into cloth bags, produce DC power that then directly powers LEDs embedded into the same bag. The bag can be unfolded into several configurations depending on the quality of light needed and where it is needed. The project leverages the inherent behaviors of the individual smart materials to yield results that are direct, discrete, transient, and local. Lighting that is designed for the use and the user, rather than for the building, challenges the generic building systems we have tried to force-fit smart materials to engage. To truly exploit these materials, we must not only understand their specific behaviors but also the phenomena they act upon.

Developed as engineering materials, smart materials are constantly emerging, evolving, and rapidly becoming obsolete. When we background the material, using it instrumentally rather than performatively, we decouple material from effect. Any number of materials can manipulate any number of behaviors to produce a desired effect. Designing to manage transient phenomena breaks the hegemony of conventional building materials and systems, allowing many more materials to enter our field. It does, however, ask that architects be willing to step into unknown territory and let these materials teach us lessons about how things work.